Kevin Systrom: Instagram - Lex Fridman 知识图谱

Lex Fridman (01:25.560)

but check in apps were all the rage.

Kevin Systrom (01:29.440)

So.

Lex Fridman (01:30.280)

What's a check in app?

Kevin Systrom (01:31.120)

Gowalla, Foursquare, Hotpotato.

Lex Fridman (01:34.320)

So I'm at a place, I'm gonna tell the world

Kevin Systrom (01:36.760)

that I'm at this place.

Lex Fridman (01:37.720)

That's right.

Lex Fridman (01:38.680)

What's the idea behind this kind of app, by the way?

Lex Fridman (01:41.240)

You know what, I'm gonna answer that,

Lex Fridman (01:42.760)

but through what Instagram became

Lex Fridman (01:45.320)

and why I believe Instagram replaced them.

Lex Fridman (01:48.040)

So the whole idea was to share with the world

Lex Fridman (01:49.800)

what you were doing, specifically with your friends, right?

Lex Fridman (01:53.880)

But there were all the rage,

Lex Fridman (01:54.800)

and Foursquare was getting all the press.

Lex Fridman (01:56.360)

And I remember sitting around saying,

Lex Fridman (01:57.800)

hey, I wanna build something,

Lex Fridman (01:58.920)

but I don't know what I wanna build.

Lex Fridman (02:00.580)

What if I built a better version of Foursquare?

Lex Fridman (02:04.320)

And I asked myself, well, why don't I like Foursquare

Lex Fridman (02:07.920)

or how could it be improved?

Lex Fridman (02:10.760)

And basically I sat down and I said,

Lex Fridman (02:13.040)

I think that if you have a few extra features,

Kevin Systrom (02:16.560)

it might be enough.

Lex Fridman (02:17.480)

One of which happened to be posting a photo

Kevin Systrom (02:19.080)

of where you were.

Lex Fridman (02:20.320)

There were some others.

Kevin Systrom (02:21.520)

It turns out that wasn't enough.

Lex Fridman (02:23.120)

My co founder joined, we were going to attack Foursquare

Lex Fridman (02:27.000)

and the likes and try to build something interesting.

Lex Fridman (02:30.400)

And no one used it.

Kevin Systrom (02:31.240)

No one cared because it wasn't enough.

Lex Fridman (02:32.600)

It wasn't different enough, right?

Lex Fridman (02:35.720)

So one day we were sitting down and we asked ourselves,

Lex Fridman (02:38.040)

okay, it's come to Jesus moment.

Lex Fridman (02:40.880)

Are we gonna do this startup?

Lex Fridman (02:43.120)

And if we're going to,

Kevin Systrom (02:44.400)

we can't do what we're currently doing.

Lex Fridman (02:46.160)

We have to switch it up.

Lex Fridman (02:47.020)

So what do people love the most?

Lex Fridman (02:48.360)

So we sat down and we wrote out three things

Kevin Systrom (02:51.760)

that we thought people uniquely loved about our product

Lex Fridman (02:54.640)

that weren't in other products.

Kevin Systrom (02:57.160)

Photos happened to be the top one.

Lex Fridman (02:59.200)

So sharing a photo of what you were doing,

Kevin Systrom (03:01.380)

where you were at the moment

Lex Fridman (03:03.160)

was not something products let you do really.

Kevin Systrom (03:06.640)

Facebook was like, post an album of your vacation

Lex Fridman (03:09.480)

from two weeks ago, right?

Kevin Systrom (03:11.440)

Twitter allowed you to post a photo,

Lex Fridman (03:13.400)

but their feed was primarily text

Lex Fridman (03:15.560)

and they didn't show the photo in line

Lex Fridman (03:17.400)

or at least I don't think they did at the time.

Lex Fridman (03:19.720)

So even though it seems totally stupid

Lex Fridman (03:23.520)

and obvious to us now, at the moment then,

Kevin Systrom (03:27.280)

posting a photo of what you were doing at the moment

Lex Fridman (03:29.360)

was like not a thing.

Lex Fridman (03:32.020)

So we decided to go after that

Lex Fridman (03:34.160)

because we noticed that people who used our service,

Kevin Systrom (03:36.480)

the one thing they happened to like the most

Lex Fridman (03:38.640)

was posting a photo.

Lex Fridman (03:40.240)

So that was the beginning of Instagram.

Lex Fridman (03:41.760)

And yes, like we went through and we added filters

Lex Fridman (03:44.460)

and there's a bunch of stories around that.

Lex Fridman (03:46.460)

But the origin of this was that

Kevin Systrom (03:48.160)

we were trying to be a check in app,

Lex Fridman (03:49.580)

realized that no one wanted another check in app.

Kevin Systrom (03:52.360)

It became a photo sharing app,

Lex Fridman (03:54.000)

but one that was much more about what you're doing

Lex Fridman (03:56.320)

and where you are.

Lex Fridman (03:57.480)

And that's why when I say,

Kevin Systrom (03:58.600)

I think we've replaced check in apps,

Lex Fridman (04:01.000)

it became a check in via a photo

Kevin Systrom (04:03.680)

rather than saying your location

Lex Fridman (04:06.580)

and then optionally adding a photo.

Kevin Systrom (04:08.760)

When you were thinking about what people like,

Lex Fridman (04:11.420)

from where did you get a sense

Lex Fridman (04:13.240)

that this is what people like?

Lex Fridman (04:14.520)

You said, we sat down, we wrote some stuff down on paper.

Kevin Systrom (04:18.920)

Where is that intuition that seems fundamental

Lex Fridman (04:21.920)

to the success of an app like Instagram?

Kevin Systrom (04:26.820)

Where does that idea,

Lex Fridman (04:27.880)

where does that list of three things come from exactly?

Kevin Systrom (04:31.500)

Only after having studied machine learning now

Lex Fridman (04:33.640)

for a couple of years, I like, I have a...

Lex Fridman (04:36.880)

You have understood yourself?

Lex Fridman (04:39.400)

I've started to make connections,

Kevin Systrom (04:40.800)

like we can go into this later,

Lex Fridman (04:43.360)

but obviously the connections between machine learning

Lex Fridman (04:50.960)

and the human brain, I think are stretched sometimes.

Lex Fridman (04:53.800)

At the same time, being able to back prop

Lex Fridman (04:57.400)

and being able to look at the world, try something,

Lex Fridman (05:01.000)

figure out how you're wrong, how wrong you are,

Lex Fridman (05:04.240)

and then nudge your company in the right direction

Lex Fridman (05:08.040)

based on how wrong you are,

Kevin Systrom (05:10.100)

is a fascinating concept.

Lex Fridman (05:12.280)

And I don't, we didn't know we were doing it at the time,

Lex Fridman (05:14.720)

but that's basically what we were doing, right?

Lex Fridman (05:17.120)

We put it out to, call it a hundred people,

Lex Fridman (05:20.480)

and you would look at their data.

Lex Fridman (05:22.780)

You would say, what are they sharing?

Lex Fridman (05:25.200)

Like what resonates, what doesn't resonate?

Lex Fridman (05:26.960)

We think they're gonna resonate with X,

Lex Fridman (05:29.040)

but it turns out they resonate with Y.

Lex Fridman (05:31.100)

Okay, shift the company towards Y.

Lex Fridman (05:33.660)

And it turns out if you do that enough quickly enough,

Lex Fridman (05:36.320)

you can get to a solution that has product market fit.

Kevin Systrom (05:39.800)

Most companies fail because they sit there

Lex Fridman (05:42.640)

and they don't, either they're learning rates too slow,

Kevin Systrom (05:45.520)

they sit there and they're just,

Lex Fridman (05:46.640)

they're adamant that they're right,

Kevin Systrom (05:47.980)

even though the data is telling them they're not right,

Lex Fridman (05:50.720)

or they're learning rates too high

Lex Fridman (05:53.240)

and they wildly chase different ideas

Lex Fridman (05:55.220)

and they never actually settle on one

Lex Fridman (05:57.120)

where they don't groove, right?

Lex Fridman (05:59.520)

And I think when we sat down

Lex Fridman (06:00.600)

and we wrote out those three ideas,

Lex Fridman (06:01.960)

what we were saying is, what are the three possible,

Lex Fridman (06:05.720)

whether they're local or global maxima in our world, right?

Lex Fridman (06:10.440)

That users are telling us they like

Kevin Systrom (06:12.800)

because they're using the product that way.

Lex Fridman (06:14.960)

It was clear people liked the photos

Kevin Systrom (06:16.480)

because that was the thing they were doing.

Lex Fridman (06:18.680)

And we just said, okay, like,

Lex Fridman (06:20.320)

what if we just cut out most of the other stuff

Lex Fridman (06:22.540)

and focus on that thing?

Lex Fridman (06:25.440)

And then it happened to be a multi billion dollar business

Lex Fridman (06:27.480)

and it's that easy by the way.

Kevin Systrom (06:30.240)

Yeah, I guess so.

Lex Fridman (06:32.320)

Well, nobody ever writes about neural networks

Kevin Systrom (06:34.320)

that miserably failed.

Lex Fridman (06:36.160)

So this particular neural network succeeded.

Lex Fridman (06:38.480)

Oh, they sell all the time, right?

Lex Fridman (06:40.240)

Yeah.

Lex Fridman (06:41.080)

But nobody writes about it.

Lex Fridman (06:41.900)

The default state is failing.

Kevin Systrom (06:42.920)

Yes.

Lex Fridman (06:44.660)

When you said the way people are using the app,

Kevin Systrom (06:48.320)

is that the loss function for this neural network

Lex Fridman (06:51.160)

or is it also self report?

Kevin Systrom (06:52.960)

Like, do you ever ask people what they like

Lex Fridman (06:54.880)

or do you have to track exactly what they're doing,

Lex Fridman (06:58.540)

not what they're saying?

Lex Fridman (07:00.020)

I once made a Thanksgiving dinner, okay?

Lex Fridman (07:03.120)

And it was for relatives and I like to cook a lot.

Lex Fridman (07:08.080)

Okay.

Lex Fridman (07:09.080)

And I worked really hard on picking the specific dishes

Lex Fridman (07:13.680)

and I was really proud because I had planned it out

Kevin Systrom (07:17.060)

using a Gantt chart and like it was ready on time

Lex Fridman (07:19.880)

and everything was hot.

Kevin Systrom (07:21.240)

Nice.

Lex Fridman (07:22.080)

Like, I don't know if you're a big Thanksgiving guy,

Lex Fridman (07:23.280)

but like the worst thing about Thanksgiving

Lex Fridman (07:25.640)

is when the turkey is cold and some things are hot

Lex Fridman (07:28.120)

and some things, anyway.

Lex Fridman (07:29.220)

You had a Gantt chart.

Lex Fridman (07:30.060)

Did you actually have a chart?

Lex Fridman (07:31.340)

Oh yeah.

Kevin Systrom (07:32.180)

OmniPlan, fairly expensive, like Gantt chart thing

Lex Fridman (07:36.040)

that I think maybe 10 people have purchased in the world,

Lex Fridman (07:39.380)

but I'm one of them and I use it for recipe planning

Lex Fridman (07:42.260)

only around big holidays.

Kevin Systrom (07:44.380)

That's brilliant, by the way.

Lex Fridman (07:45.920)

Do people do this kind of...

Lex Fridman (07:47.840)

Overengineering?

Lex Fridman (07:48.920)

It's not overengineering, it's just engineering.

Kevin Systrom (07:50.820)

It's planning.

Lex Fridman (07:51.660)

Thanksgiving is a complicated set of events

Kevin Systrom (07:54.720)

with some uncertainty with a lot of things going on.

Lex Fridman (07:57.420)

You should be able, you should be planning it in this way.

Kevin Systrom (07:59.760)

There should be a chart.

Lex Fridman (08:00.840)

It's not overengineering.

Kevin Systrom (08:01.660)

I mean, so what's funny is, brief aside.

Lex Fridman (08:04.860)

Yes, it's brilliant.

Kevin Systrom (08:06.420)

I love cooking, I love food, I love coffee,

Lex Fridman (08:08.660)

and I've spent some time with some chefs

Kevin Systrom (08:10.340)

who like know their stuff.

Lex Fridman (08:12.180)

And they always just take out a piece of paper

Lex Fridman (08:15.500)

and just work backwards in rough order.

Lex Fridman (08:17.900)

Like it's never perfect, but rough order.

Kevin Systrom (08:20.540)

It's just like, oh, that makes sense.

Lex Fridman (08:21.900)

Why not just work backwards from the end goal, right?

Lex Fridman (08:24.740)

And put in some buffer time.

Lex Fridman (08:26.060)

And so I probably over specified it a bit using a Gantt chart,

Lex Fridman (08:29.860)

but the fact that you can do it,

Lex Fridman (08:32.720)

it's what professional kitchens roughly do.

Kevin Systrom (08:35.220)

They just don't call it a Gantt chart,

Lex Fridman (08:36.880)

or at least I don't think they do.

Kevin Systrom (08:38.700)

Anyway, I was telling a story about Thanksgiving.

Lex Fridman (08:40.280)

So here's the thing.

Kevin Systrom (08:42.940)

I'm sitting down, we have the meal,

Lex Fridman (08:44.700)

and then I got to know Ray Dalio fairly well

Kevin Systrom (08:49.220)

over maybe the last year of Instagram.

Lex Fridman (08:53.060)

And one thing that he kept saying was like,

Kevin Systrom (08:55.100)

feedback is really hard to get honestly from people.

Lex Fridman (08:59.900)

And I sat down after dinner, I said,

Kevin Systrom (09:02.420)

guys, I want feedback.

Lex Fridman (09:03.680)

What was good and what was bad?

Kevin Systrom (09:05.640)

Yes.

Lex Fridman (09:06.480)

And what's funny is like,

Kevin Systrom (09:07.820)

literally everyone just said everything was great.

Lex Fridman (09:11.140)

And I like personally knew I had screwed up

Kevin Systrom (09:13.380)

a handful of things, but no one would say it.

Lex Fridman (09:17.580)

And can you imagine now not something as high stakes

Lex Fridman (09:20.740)

as Thanksgiving dinner, okay?

Lex Fridman (09:22.020)

Thanksgiving dinner, it's not that high stakes.

Lex Fridman (09:25.100)

But you're trying to build a product

Lex Fridman (09:26.380)

and everyone knows you left your job for it

Lex Fridman (09:28.380)

and you're trying to build it out

Lex Fridman (09:29.540)

and you're trying to make something wonderful

Lex Fridman (09:32.000)

and it's yours, right?

Lex Fridman (09:33.440)

You designed it.

Kevin Systrom (09:35.580)

Now try asking for feedback

Lex Fridman (09:37.900)

and know that you're giving this to your friends

Lex Fridman (09:39.980)

and your family.

Lex Fridman (09:42.180)

People have trouble giving hard feedback.

Kevin Systrom (09:45.580)

People have trouble saying, I don't like this

Lex Fridman (09:48.860)

or this isn't great, or this is how it's failed me.

Kevin Systrom (09:51.980)

In fact, you usually have two classes of people.

Lex Fridman (09:56.600)

People who just won't say bad things,

Kevin Systrom (09:58.780)

you can literally say to them,

Lex Fridman (10:00.440)

please tell me what you hate most about this

Lex Fridman (10:02.380)

and they won't do it.

Lex Fridman (10:03.220)

They'll try, but they won't.

Lex Fridman (10:05.580)

And then the other class of people

Lex Fridman (10:06.900)

are just negative period about everything

Lex Fridman (10:09.700)

and it's hard to parse out like what is true and what isn't.

Lex Fridman (10:14.840)

So my rule of thumb with this is

Kevin Systrom (10:17.700)

you should always ask people,

Lex Fridman (10:20.100)

but at the end of the day, it's amazing what data

Kevin Systrom (10:22.300)

will tell you.

Lex Fridman (10:23.440)

And that's why with whatever project I work on, even now,

Kevin Systrom (10:27.280)

collecting data from the beginning on usage patterns,

Lex Fridman (10:30.220)

so engagement, how many days of the week do they use it?

Lex Fridman (10:34.340)

How many, I don't know if we were to go back to Instagram,

Lex Fridman (10:36.660)

how many impressions per day, right?

Lex Fridman (10:39.460)

Is that growing?

Lex Fridman (10:40.280)

Is that shrinking?

Lex Fridman (10:41.120)

And don't be like overly scientific about it, right?

Lex Fridman (10:44.740)

Cause maybe you have 50 beta users or something.

Lex Fridman (10:48.300)

But what's fascinating is that data doesn't lie.

Lex Fridman (10:52.740)

People are very defensive about their time.

Kevin Systrom (10:57.580)

They'll say, oh, I'm so busy,

Lex Fridman (10:59.260)

I'm sorry I didn't get to use the app.

Kevin Systrom (11:00.740)

Like I'm just, you know,

Lex Fridman (11:03.300)

but I don't know you were posting on Instagram

Kevin Systrom (11:05.220)

the whole time.

Lex Fridman (11:07.140)

So I don't know at the end of the day,

Kevin Systrom (11:09.260)

like at Facebook there was, you know,

Lex Fridman (11:11.500)

before time spent became kind of this loaded term there.

Kevin Systrom (11:15.820)

The idea that people's currency in their lives is time.

Lex Fridman (11:21.060)

And they only have a certain amount of time to give things,

Kevin Systrom (11:23.380)

whether it's friends or family or apps or TV shows

Lex Fridman (11:26.260)

or whatever, there's no way of inventing more of it,

Kevin Systrom (11:28.940)

at least not that I know of.

Lex Fridman (11:32.700)

If they don't use it, it's because it's not great.

Lex Fridman (11:35.540)

So the moral of the story is you can ask all you want,

Lex Fridman (11:39.720)

but you just have to look at the data.

Lex Fridman (11:41.860)

And data doesn't lie, right?

Lex Fridman (11:45.340)

I mean, there's metrics, there's data can obscure

Kevin Systrom (11:51.140)

the key insight if you're not careful.

Lex Fridman (11:54.020)

So time spent in the app, that's one.

Kevin Systrom (11:57.940)

There's so many metrics you can put at this

Lex Fridman (11:59.820)

and they will give you totally different insights,

Kevin Systrom (12:02.500)

especially when you're trying to create something

Lex Fridman (12:04.600)

that doesn't obviously exist yet.

Kevin Systrom (12:07.660)

So, you know, measuring maybe why you left the app

Lex Fridman (12:12.660)

or measuring special moments of happiness

Kevin Systrom (12:17.780)

that will make sure you return to the app

Lex Fridman (12:20.020)

or moments of happiness that are long lasting

Kevin Systrom (12:22.380)

versus like dopamine short term, all of those things.

Lex Fridman (12:26.440)

But I think I suppose in the beginning,

Kevin Systrom (12:29.660)

you can just get away with just asking the question,

Lex Fridman (12:33.180)

which features are used a lot?

Kevin Systrom (12:35.380)

Let's do more of that.

Lex Fridman (12:37.260)

And how hard was the decision?

Lex Fridman (12:40.720)

And I mean, maybe you can tell me what Instagram

Lex Fridman (12:43.180)

looked in the beginning,

Lex Fridman (12:44.460)

but how hard was it to make pictures

Lex Fridman (12:47.060)

of the first class citizen?

Kevin Systrom (12:48.500)

That's a revolutionary idea.

Lex Fridman (12:50.460)

Like at whatever point Instagram became this feed of photos,

Kevin Systrom (12:56.820)

that's quite brilliant.

Lex Fridman (12:58.460)

Plus, I also don't know when this happened,

Lex Fridman (13:02.020)

but they're all shaped the same.

Lex Fridman (13:05.660)

It's like a...

Kevin Systrom (13:06.500)

I have to tell you why, that's the interesting part.

Lex Fridman (13:09.100)

Why is that?

Lex Fridman (13:10.980)

So a couple of things.

Lex Fridman (13:11.900)

One is data, like you're right.

Kevin Systrom (13:16.820)

You can overinterpret data.

Lex Fridman (13:18.020)

Like imagine trying to fly a plane by staring at,

Kevin Systrom (13:22.780)

I don't know, a single metric like airspeed.

Lex Fridman (13:25.700)

You don't know if you're going up or down.

Kevin Systrom (13:27.060)

I mean, it correlates with up or down,

Lex Fridman (13:28.580)

but you don't actually know.

Kevin Systrom (13:29.980)

It will never help you land the plane.

Lex Fridman (13:32.240)

So don't stare at one metric.

Kevin Systrom (13:33.940)

Like it turns out you have to synthesize a bunch of metrics

Lex Fridman (13:36.840)

to know where to go.

Lex Fridman (13:37.900)

But it doesn't lie.

Lex Fridman (13:40.140)

Like if your airspeed is zero,

Lex Fridman (13:41.380)

unless it's not working, right?

Lex Fridman (13:43.100)

If it's zero, you're probably gonna fall out of the sky.

Lex Fridman (13:46.380)

So generally you look around and you have the scan going.

Lex Fridman (13:50.540)

Yes.

Lex Fridman (13:51.660)

And you're just asking yourself,

Lex Fridman (13:52.900)

is this working or is this not working?

Lex Fridman (13:56.340)

But people have trouble explaining how they actually feel.

Lex Fridman (14:02.700)

So just, it's about synthesizing both of them.

Lex Fridman (14:05.100)

So then Instagram, right?

Lex Fridman (14:07.780)

We were talking about revolutionary moment

Kevin Systrom (14:10.460)

where the feed became square photos basically.

Lex Fridman (14:14.180)

And photos first and then square photos.

Kevin Systrom (14:16.340)

Yeah.

Lex Fridman (14:19.020)

It was clear to me that the biggest,

Lex Fridman (14:21.500)

so I believe the biggest companies are founded

Lex Fridman (14:26.900)

when enormous technical shifts happen.

Lex Fridman (14:30.780)

And the biggest technical shift that happened

Lex Fridman (14:32.860)

right before Instagram was founded

Kevin Systrom (14:34.860)

was the advent of a phone that didn't suck.

Lex Fridman (14:38.020)

The iPhone, right?

Kevin Systrom (14:38.980)

Like in retrospect, we're like, oh my God,

Lex Fridman (14:40.740)

the first iPhone that almost had,

Kevin Systrom (14:43.180)

like it wasn't that good.

Lex Fridman (14:44.540)

But compared to everything else at the time,

Kevin Systrom (14:46.800)

it was amazing.

Lex Fridman (14:48.940)

And by the way,

Kevin Systrom (14:50.140)

the first phone that had an incredible camera

Lex Fridman (14:54.760)

that could like do as well as the point and shoot

Kevin Systrom (14:57.740)

you might carry around was the iPhone 4.

Lex Fridman (15:01.100)

And that was right when Instagram launched.

Lex Fridman (15:02.940)

And we looked around and we said,

Lex Fridman (15:04.660)

what will change?

Kevin Systrom (15:05.940)

Because everyone has a camera in their pocket.

Lex Fridman (15:08.540)

And it was so clear to me that the world

Kevin Systrom (15:11.800)

of social networks before it was based in the desktop

Lex Fridman (15:16.860)

and sitting there and having a link you could share, right?

Lex Fridman (15:20.780)

And that wasn't gonna be the case.

Lex Fridman (15:21.780)

So the question is what would you share

Lex Fridman (15:23.420)

if you were out and about in the world?

Lex Fridman (15:26.900)

If not only did you have a camera that fit in your pocket,

Lex Fridman (15:29.600)

but by the way, that camera had a network attached to it

Lex Fridman (15:31.980)

that allowed you to share instantly.

Kevin Systrom (15:34.780)

That seemed revolutionary.

Lex Fridman (15:36.100)

And a bunch of people saw it at the same time.

Kevin Systrom (15:37.700)

It wasn't just Instagram.

Lex Fridman (15:38.760)

There were a bunch of competitors.

Kevin Systrom (15:41.340)

The thing we did, I think was not only,

Lex Fridman (15:44.100)

well, we focused on two things.

Lex Fridman (15:45.320)

So we wrote down those things, we circled photos

Lex Fridman (15:47.700)

and we said, I think we should invest in this.

Lex Fridman (15:50.100)

But then we said, what sucks about photos?

Lex Fridman (15:52.760)

One, they look like crap, right?

Kevin Systrom (15:55.500)

They just, at least back then.

Lex Fridman (15:57.060)

Now my phone takes pretty great photos, right?

Lex Fridman (16:01.340)

Back then they were blurry, not so great, compressed, right?

Lex Fridman (16:06.140)

Two, it was really slow, like really slow to upload a photo.

Lex Fridman (16:12.200)

And I'll tell a fun story about that

Lex Fridman (16:13.700)

and explain to you why they're all the same size

Lex Fridman (16:15.900)

and square as well.

Lex Fridman (16:18.220)

And three, man, if you wanted to share a photo

Kevin Systrom (16:21.780)

on different networks, you had to go to each

Lex Fridman (16:24.020)

of the individual apps and select all of them

Lex Fridman (16:26.620)

and upload individually.

Lex Fridman (16:27.780)

And so we were like, all right, those are the pain points.

Kevin Systrom (16:30.580)

We're gonna focus on that.

Lex Fridman (16:31.940)

So one, instead of, because they weren't beautiful,

Kevin Systrom (16:36.020)

we were like, why don't we lean into the fact

Lex Fridman (16:37.540)

that they're not beautiful?

Lex Fridman (16:38.380)

And I remember studying in Florence,

Lex Fridman (16:40.620)

my photography teacher gave me this Holga camera

Lex Fridman (16:42.980)

and I'm not sure everyone knows what a Holga camera is,

Lex Fridman (16:44.940)

but they're these old school plastic cameras.

Kevin Systrom (16:48.180)

I think they're produced in China at the time.

Lex Fridman (16:52.100)

And I wanna say the original ones were like

Kevin Systrom (16:54.240)

from the 70s or the 80s or something.

Lex Fridman (16:55.980)

They're supposed to be like $3 cameras for the every person.

Kevin Systrom (16:58.940)

They took nice medium format films, large negatives,

Lex Fridman (17:05.200)

but they kind of blurred the light

Lex Fridman (17:07.880)

and they kind of like light leaked into the side.

Lex Fridman (17:10.240)

And there was this whole resurgence where people looked

Lex Fridman (17:13.200)

at that and said, oh my God, this is a style, right?

Lex Fridman (17:16.900)

And I remember using that in Florence and just saying,

Kevin Systrom (17:19.020)

well, why don't we just like lean into the fact

Lex Fridman (17:20.840)

that these photos suck and make them suck more,

Lex Fridman (17:24.320)

but in an artistic way.

Lex Fridman (17:26.220)

And it turns out that had product market fit.

Kevin Systrom (17:28.040)

People really liked that.

Lex Fridman (17:29.200)

They were willing to share their not so great photos

Lex Fridman (17:31.720)

if they looked not so great on purpose, okay?

Lex Fridman (17:35.820)

The second part.

Kevin Systrom (17:37.400)

That's where the filters come into picture.

Lex Fridman (17:40.320)

So computational modification of photos

Kevin Systrom (17:42.560)

to make them look extra crappy to where it becomes art.

Lex Fridman (17:46.200)

Yeah, yeah.

Lex Fridman (17:47.040)

And I mean, add light leaks, add like an overlay filter,

Lex Fridman (17:51.720)

make them more contrasty than they should be.

Kevin Systrom (17:54.120)

The first filter we ever produced was called X Pro 2.

Lex Fridman (17:58.080)

And I designed it while I was

Kevin Systrom (17:59.600)

in this small little bed and breakfast room

Lex Fridman (18:02.160)

in Todos Santos, Mexico.

Kevin Systrom (18:03.640)

I was trying to take a break from the bourbon days.

Lex Fridman (18:06.800)

And I remember saying to my co founder,

Kevin Systrom (18:08.840)

I just need like a week to reset.

Lex Fridman (18:11.080)

And that was on that trip worked on the first filter

Kevin Systrom (18:15.080)

because I said, you know, I think I can do this.

Lex Fridman (18:17.240)

And I literally iterated one by one over the RGB values

Kevin Systrom (18:22.080)

in the array that was the photo and just slightly shifted.

Lex Fridman (18:26.280)

Basically there was a function of R, function of G,

Kevin Systrom (18:29.400)

function of B that just shifted them slightly.

Lex Fridman (18:32.480)

It was in rocket science.

Lex Fridman (18:34.000)

And it turns out that actually

Lex Fridman (18:36.540)

made your photo look pretty cool.

Kevin Systrom (18:38.520)

It just mapped from one color space to another color space.

Lex Fridman (18:42.820)

It was simple, but it was really slow.

Kevin Systrom (18:44.440)

I mean, if you applied a filter,

Lex Fridman (18:47.320)

I think it used to take two or three seconds to render.

Kevin Systrom (18:50.500)

Only eventually would I figure out how to do it on the GPU.

Lex Fridman (18:53.200)

And I'm not even sure it was a GPU, but it was using OpenGL.

Lex Fridman (18:56.400)

But anyway, I would eventually figure that out

Lex Fridman (18:59.960)

and then it would be instant, but it used to be really slow.

Kevin Systrom (19:02.800)

By the way, anyone who's watching or listening,

Lex Fridman (19:06.000)

it's amazing what you can get away with in a startup

Kevin Systrom (19:10.080)

as long as the product outcome is right for the user.

Lex Fridman (19:14.080)

Like you can be slow.

Kevin Systrom (19:15.440)

You can be terrible.

Lex Fridman (19:16.480)

You can be, as long as you have product market fit,

Kevin Systrom (19:21.000)

people will put up with a lot.

Lex Fridman (19:22.440)

And then the question is just about compressing,

Kevin Systrom (19:25.840)

making it more performant over time

Lex Fridman (19:27.640)

so that they get that product market fit instantly.

Lex Fridman (19:30.440)

So fascinating because there's some things

Lex Fridman (19:32.800)

where those three seconds would make or break the app,

Lex Fridman (19:37.440)

but some things you're saying not.

Lex Fridman (19:39.480)

It's hard to know when, it's the problem Spotify solved

Kevin Systrom (19:43.480)

for making streaming like work.

Lex Fridman (19:48.500)

And like delays in listening to music is a huge negative,

Kevin Systrom (19:53.060)

even like slight delays.

Lex Fridman (19:55.480)

But here you're saying, I mean,

Lex Fridman (19:57.260)

how do you know when those three seconds are okay

Lex Fridman (1:00:00.820)

But I mean, but what's interesting is like,

Kevin Systrom (1:00:02.660)

it's one thing to train a game and like,

Lex Fridman (1:00:06.660)

I mean that paper where they just took Atari

Lex Fridman (1:00:09.220)

and they used a ConvNet to basically just like

Lex Fridman (1:00:12.020)

train simple actions, mind blowing, right?

Kevin Systrom (1:00:15.100)

Absolutely mind blowing, but it's a game, great.

Lex Fridman (1:00:18.780)

So now what if you're constructing a feed for a person,

Lex Fridman (1:00:23.620)

right?

Lex Fridman (1:00:24.460)

Like how can you construct that feed in such a way

Kevin Systrom (1:00:29.580)

that optimizes for a diversity of experience,

Lex Fridman (1:00:32.380)

a longterm happiness, right?

Lex Fridman (1:00:36.860)

But that reward function,

Lex Fridman (1:00:38.420)

it turns out in reinforcement learning again,

Kevin Systrom (1:00:41.180)

as I've learned, like reward design is really hard.

Lex Fridman (1:00:45.220)

And I don't know, like how do you design a scalar reward

Lex Fridman (1:00:49.740)

for someone's happiness over time?

Lex Fridman (1:00:51.340)

I mean, do you have to measure dopamine levels?

Lex Fridman (1:00:53.300)

Like, do you have to?

Lex Fridman (1:00:54.140)

Well, you have to have a lot more signals

Kevin Systrom (1:00:58.020)

from the human being.

Lex Fridman (1:00:59.500)

Currently it feels like there's not enough signals

Kevin Systrom (1:01:01.700)

coming from the human being users of this algorithm.

Lex Fridman (1:01:06.140)

So for reinforcement learning to work well,

Kevin Systrom (1:01:08.780)

you need to have a lot more data.

Lex Fridman (1:01:10.860)

Needs to have a lot of data.

Lex Fridman (1:01:11.980)

And that actually is a challenge for anyone

Lex Fridman (1:01:13.420)

who wants to start something,

Kevin Systrom (1:01:14.500)

which is you don't have a lot of data.

Lex Fridman (1:01:16.460)

So how do you compete?

Lex Fridman (1:01:17.500)

But I do think back to your original point,

Lex Fridman (1:01:20.020)

rethinking the algorithm, rethinking reward functions,

Kevin Systrom (1:01:23.420)

rethinking utility, that's fascinating.

Lex Fridman (1:01:28.060)

That's cool.

Lex Fridman (1:01:28.900)

And I think that's an open opportunity

Lex Fridman (1:01:31.060)

for a company that figures it out.

Kevin Systrom (1:01:34.540)

I have to ask about April, 2012,

Lex Fridman (1:01:37.420)

when Instagram, along with its massive employee base

Kevin Systrom (1:01:43.220)

of 13 people was sold to Facebook for $1 billion.

Lex Fridman (1:01:48.020)

What was the process like on a business level,

Lex Fridman (1:01:50.220)

engineering level, human level?

Lex Fridman (1:01:52.140)

What was that process of selling to Facebook like?

Lex Fridman (1:01:54.940)

What did it feel like?

Lex Fridman (1:01:56.460)

So I want to provide some context,

Kevin Systrom (1:01:58.100)

which is I worked in corporate development at Google,

Lex Fridman (1:02:00.900)

which not a lot of people know,

Lex Fridman (1:02:02.660)

but corporate development is effectively the group

Lex Fridman (1:02:04.500)

that buys companies, right?

Kevin Systrom (1:02:06.220)

You sit there and you acquire companies.

Lex Fridman (1:02:08.140)

And I had sat through so many of these meetings

Kevin Systrom (1:02:10.940)

with entrepreneurs.

Lex Fridman (1:02:12.100)

We actually, fun fact, we never acquired a single company

Kevin Systrom (1:02:14.540)

when I worked in corporate development.

Lex Fridman (1:02:15.740)

So I can't claim that I had like a lot of experience,

Lex Fridman (1:02:21.140)

but I had enough experience to understand,

Lex Fridman (1:02:23.420)

okay, like what prices are people getting

Lex Fridman (1:02:25.580)

and what's the process?

Lex Fridman (1:02:27.340)

And as we started to grow,

Kevin Systrom (1:02:32.260)

we were trying to keep this thing running

Lex Fridman (1:02:33.980)

and we were exhausted and we were 13 people.

Lex Fridman (1:02:36.220)

And I mean, we were trying to think back,

Lex Fridman (1:02:38.820)

it was probably 27, 37 now,

Lex Fridman (1:02:43.700)

so young on a relative basis, right?

Lex Fridman (1:02:48.380)

And we're trying to keep the thing running.

Lex Fridman (1:02:50.540)

And then we go out to raise money

Lex Fridman (1:02:53.020)

and we're kind of like the hot startup at the time.

Lex Fridman (1:02:57.420)

And I remember going into a specific VC and saying,

Lex Fridman (1:03:00.460)

our terms we're looking for are,

Kevin Systrom (1:03:02.340)

we're looking for a $500 million valuation.

Lex Fridman (1:03:05.460)

And I've never seen so many jaws drop all in unison, right?

Lex Fridman (1:03:10.260)

And I was like, thanked and walked out the door

Lex Fridman (1:03:12.340)

very kindly after.

Lex Fridman (1:03:14.300)

And then I got a call the next day

Lex Fridman (1:03:16.660)

from someone who was connected to them.

Lex Fridman (1:03:18.660)

And they said, we just wanna let you know

Lex Fridman (1:03:21.220)

that like it was pretty offensive

Kevin Systrom (1:03:22.900)

that you asked for a $500 million valuation.

Lex Fridman (1:03:25.060)

And I can't tell if that was like just negotiating or what,

Lex Fridman (1:03:30.060)

but it's true, like no one offered us more, right?

Lex Fridman (1:03:32.580)

So we were...

Lex Fridman (1:03:33.420)

So can you clarify the number again?

Lex Fridman (1:03:35.740)

You said how many million?

Kevin Systrom (1:03:37.100)

500.

Lex Fridman (1:03:37.940)

500 million.

Kevin Systrom (1:03:38.780)

500 million, yeah, half a billion.

Lex Fridman (1:03:42.420)

So in my mind, I'm anchored like, okay,

Kevin Systrom (1:03:44.140)

well, literally no one's biting at 500 million.

Lex Fridman (1:03:46.660)

And eventually we would get Sequoia and Greylock

Lex Fridman (1:03:50.100)

and others together at 500 million, basically, post.

Lex Fridman (1:03:54.100)

It was 450 pre, I think we raised $50 million.

Lex Fridman (1:03:56.940)

But just like no one was used to seeing

Lex Fridman (1:03:59.140)

a $500 million companies then.

Kevin Systrom (1:04:01.620)

Like, I don't know if it was because we were just coming

Lex Fridman (1:04:03.980)

out of the hangover of 2008

Lex Fridman (1:04:06.860)

and things were still on recovery mode.

Lex Fridman (1:04:10.500)

But then along comes Facebook.

Lex Fridman (1:04:13.300)

And after some negotiation, we've two X to the number

Lex Fridman (1:04:18.300)

from a half a billion to a billion.

Kevin Systrom (1:04:21.020)

Yeah, it seems pretty good.

Lex Fridman (1:04:22.860)

And I think Mark and I really saw eye to eye

Kevin Systrom (1:04:27.300)

that this thing could be big.

Lex Fridman (1:04:29.100)

We thought we could...

Kevin Systrom (1:04:30.220)

Their resources would help us scale it.

Lex Fridman (1:04:32.620)

And in a lot of ways it de risks.

Kevin Systrom (1:04:34.900)

I mean, it de risks a lot of the employees lives

Lex Fridman (1:04:37.100)

for the rest of their lives,

Lex Fridman (1:04:38.420)

including me, including Mike, right?

Lex Fridman (1:04:41.620)

I think I might've had like 10 grand

Lex Fridman (1:04:42.900)

in my bank account at the time, right?

Lex Fridman (1:04:44.740)

Like we're working hard, we had nothing.

Lex Fridman (1:04:48.380)

So on a relative basis, it seems very high.

Lex Fridman (1:04:51.340)

And then I think the last company to exit

Kevin Systrom (1:04:53.940)

for anywhere close to a billion was YouTube

Lex Fridman (1:04:55.900)

that I could think of.

Lex Fridman (1:04:57.900)

And thus began the giant long bull run of 2012

Lex Fridman (1:05:02.420)

to all the way to where we are now,

Kevin Systrom (1:05:04.540)

where I saw some stat yesterday

Lex Fridman (1:05:06.940)

about like how many unicorns exist and it's absurd.

Lex Fridman (1:05:11.060)

But then again, never underestimate technology

Lex Fridman (1:05:13.500)

and like the value it can provide.

Lex Fridman (1:05:15.740)

And man, costs have dropped and man scale has increased.

Lex Fridman (1:05:19.060)

And you can make businesses make a lot of money now.

Lex Fridman (1:05:23.180)

But on a fundamental level, I don't know,

Lex Fridman (1:05:26.300)

like how do you describe the decision

Lex Fridman (1:05:29.100)

to sell a company with 13 people for a billion dollars?

Lex Fridman (1:05:32.460)

So first of all, like how did it take a lot of guts

Kevin Systrom (1:05:36.420)

to sit at a table and say 500 million

Lex Fridman (1:05:38.860)

or 1 billion with Mark Zuckerberg?

Kevin Systrom (1:05:41.660)

It seems like a very large number with 13.

Lex Fridman (1:05:44.220)

Like, especially...

Kevin Systrom (1:05:45.580)

It doesn't seem, it is.

Lex Fridman (1:05:47.180)

It is.

Kevin Systrom (1:05:48.020)

They're all large numbers.

Lex Fridman (1:05:49.140)

Especially like you said before the unicorn parade.

Kevin Systrom (1:05:55.620)

I like that, I'm gonna use that.

Lex Fridman (1:05:56.860)

The unicorn parade?

Kevin Systrom (1:05:57.980)

Yeah.

Lex Fridman (1:05:59.700)

You were at the head of the unicorn parade.

Kevin Systrom (1:06:01.660)

It's the, yeah, it's a massive unicorn parade.

Lex Fridman (1:06:04.900)

Okay, so no, I mean, we knew we were worth

Kevin Systrom (1:06:10.140)

quote unquote a lot, but we didn't,

Lex Fridman (1:06:12.340)

I mean, there was no market for Instagram.

Kevin Systrom (1:06:14.580)

I mean, it's not, you couldn't mark to market this thing

Lex Fridman (1:06:17.140)

in the public markets.

Kevin Systrom (1:06:18.060)

You didn't quite understand what it would be worth

Lex Fridman (1:06:20.940)

or was worth at the time.

Lex Fridman (1:06:23.020)

So in a market, an illiquid market

Lex Fridman (1:06:24.700)

where you have one buyer and one seller

Lex Fridman (1:06:26.340)

and you're going back and forth,

Lex Fridman (1:06:27.620)

and well, I guess there were like VC firms

Kevin Systrom (1:06:31.340)

who were willing to invest at a certain valuation.

Lex Fridman (1:06:34.860)

So I don't know, you just go with your gut.

Lex Fridman (1:06:38.460)

And at the end of the day, I would say

Lex Fridman (1:06:41.540)

the hardest part of it was not realizing,

Kevin Systrom (1:06:50.060)

like when we sold, it was tough

Lex Fridman (1:06:52.140)

because like literally everywhere I go,

Kevin Systrom (1:06:54.900)

restaurants, whatever, like for a good six months after,

Lex Fridman (1:06:59.220)

there was a lot of attention on the deal,

Kevin Systrom (1:07:01.060)

a lot of attention on the product,

Lex Fridman (1:07:03.100)

a lot of attention, it was kind of miserable, right?

Lex Fridman (1:07:06.380)

And you're like, wait, like I made a lot of money,

Lex Fridman (1:07:08.140)

but like, why is this not great?

Lex Fridman (1:07:10.220)

And it's because it turns out,

Lex Fridman (1:07:14.060)

I don't know, like I don't really keep in touch with Mark,

Lex Fridman (1:07:16.260)

but I've got to assume his job right now

Lex Fridman (1:07:17.940)

is not exactly the most happy job in the world.

Kevin Systrom (1:07:19.980)

It's really tough when you're on top

Lex Fridman (1:07:21.740)

and it's really tough when you're in the limelight.

Lex Fridman (1:07:24.420)

So the decision itself was like, oh, cool, this is great.

Lex Fridman (1:07:27.500)

How lucky are we, right?

Kevin Systrom (1:07:29.260)

So, okay, there's a million question I want to ask.

Lex Fridman (1:07:30.620)

Yeah, go, go, go.

Lex Fridman (1:07:32.020)

First of all, why is it hard to be on top?

Lex Fridman (1:07:37.500)

Why did you not feel good?

Lex Fridman (1:07:39.340)

Like, can you dig into that?

Lex Fridman (1:07:40.700)

It always, I've heard like Olympic athletes say

Kevin Systrom (1:07:45.020)

after they win gold, they get depressed.

Lex Fridman (1:07:49.340)

Is it something like that where it feels like

Lex Fridman (1:07:53.860)

it was kind of like a thing you were working towards?

Lex Fridman (1:07:56.340)

Yeah, sure.

Kevin Systrom (1:07:57.180)

Some loose definition of success.

Lex Fridman (1:07:58.860)

And this sure as heck feels like

Kevin Systrom (1:08:01.740)

at least according to other startups,

Lex Fridman (1:08:03.460)

this is what success looks like.

Lex Fridman (1:08:04.860)

And now why don't I feel any better?

Lex Fridman (1:08:08.420)

I'm still human, I still have all the same problems.

Lex Fridman (1:08:10.540)

Is that the nature?

Lex Fridman (1:08:11.740)

Or is it just like negative attention of some kind?

Kevin Systrom (1:08:14.380)

I think it's all of the above.

Lex Fridman (1:08:16.020)

But to be clear, there was a lot of happiness

Kevin Systrom (1:08:18.540)

in terms of like, oh my God, this is great.

Lex Fridman (1:08:20.780)

Like we won the Super Bowl of startups, right?

Kevin Systrom (1:08:25.660)

Anyone who can get to a liquidity event

Lex Fridman (1:08:28.700)

of anything meaningful feels like,

Kevin Systrom (1:08:31.020)

wow, this is what we started out to do.

Lex Fridman (1:08:32.860)

Of course we want to create great things that people love,

Lex Fridman (1:08:35.100)

but like we won in a big way.

Lex Fridman (1:08:38.580)

But yeah, there's this big like,

Lex Fridman (1:08:39.980)

oh, if we won, what's next?

Lex Fridman (1:08:43.860)

So they call it the we have arrived syndrome,

Kevin Systrom (1:08:49.180)

which I need to go back and look where I can quote that from.

Lex Fridman (1:08:52.660)

But I remember reading about it at the time.

Kevin Systrom (1:08:54.140)

I was like, oh yeah, that's that.

Lex Fridman (1:08:56.060)

And I remember we had a product manager leave very early on

Kevin Systrom (1:08:59.220)

when we got to Facebook.

Lex Fridman (1:09:00.340)

And he said to me,

Kevin Systrom (1:09:01.940)

I just don't believe I can learn anything

Lex Fridman (1:09:03.740)

at this company anymore.

Kevin Systrom (1:09:05.100)

It's like, it's hit its apex.

Lex Fridman (1:09:08.180)

We sold it great.

Kevin Systrom (1:09:09.900)

I just don't have anything else to learn.

Lex Fridman (1:09:12.380)

So from 2012 all the way to the day I left in 2018,

Kevin Systrom (1:09:15.940)

like the amount I learned and the humility

Lex Fridman (1:09:18.940)

with which I realized, oh, we thought we won.

Kevin Systrom (1:09:22.420)

Billion dollars is cool,

Lex Fridman (1:09:23.500)

but like there are a hundred billion dollar companies.

Lex Fridman (1:09:26.140)

And by the way, on top of that, we had no revenue.

Lex Fridman (1:09:29.580)

We had, I mean, we had a cool product,

Lex Fridman (1:09:31.260)

but we didn't scale it yet.

Lex Fridman (1:09:32.660)

And there's so much to learn.

Lex Fridman (1:09:34.900)

And then competitors and how fun was it to fight Snapchat?

Lex Fridman (1:09:38.340)

Oh my God.

Kevin Systrom (1:09:39.180)

Like it was, it's like Yankees Red Sox.

Lex Fridman (1:09:42.020)

It's great.

Kevin Systrom (1:09:42.900)

Like that's what you live for.

Lex Fridman (1:09:46.060)

You know, you win some, you lose some,

Lex Fridman (1:09:47.900)

but the amount you can learn through that process,

Lex Fridman (1:09:52.700)

what I've realized in life is that there is no,

Lex Fridman (1:09:56.860)

and there's always someone who has more,

Lex Fridman (1:09:58.620)

there's always more challenge, just at different scales.

Lex Fridman (1:10:02.260)

And it sounds like a little Buddhist,

Lex Fridman (1:10:04.940)

but everything is super challenging,

Kevin Systrom (1:10:09.680)

whether you're like a small business

Lex Fridman (1:10:11.100)

or an enormous business.

Lex Fridman (1:10:13.100)

I say like choose the game you like to play, right?

Lex Fridman (1:10:17.100)

You've got to imagine that

Kevin Systrom (1:10:18.060)

if you're an amazing basketball player,

Lex Fridman (1:10:19.660)

you enjoy to some extent practicing basketball.

Kevin Systrom (1:10:22.900)

It's gotta be something you love.

Lex Fridman (1:10:24.020)

It's gonna suck.

Kevin Systrom (1:10:24.860)

It's gonna be hard.

Lex Fridman (1:10:26.300)

You're gonna have injuries, right?

Lex Fridman (1:10:27.460)

But you gotta love it.

Lex Fridman (1:10:28.460)

And the same thing with Instagram,

Kevin Systrom (1:10:30.460)

which is we might've sold, but it was like, great.

Lex Fridman (1:10:35.020)

There's one Super Bowl title.

Lex Fridman (1:10:37.220)

Can we win five?

Lex Fridman (1:10:38.660)

What else can we do?

Kevin Systrom (1:10:40.180)

Now I imagine you didn't ask this, but okay, so I left.

Lex Fridman (1:10:43.500)

There's a little bit of like, what do you do next, right?

Lex Fridman (1:10:46.920)

Like, how do you top that thing?

Lex Fridman (1:10:49.220)

It's the wrong question.

Kevin Systrom (1:10:50.660)

The question is like, when you wake up every day,

Lex Fridman (1:10:53.080)

what is the hardest, most interesting thing

Lex Fridman (1:10:54.900)

you can go work on?

Lex Fridman (1:10:56.500)

Because like at the end of the day,

Lex Fridman (1:10:58.180)

we all turn into dirt, it doesn't matter, right?

Lex Fridman (1:11:00.780)

But what does matter is like,

Lex Fridman (1:11:02.540)

can we really enjoy this life?

Lex Fridman (1:11:05.420)

Not in a hedonistic way, because that's those,

Kevin Systrom (1:11:08.540)

it's like the reinforcement learning,

Lex Fridman (1:11:09.780)

learning like short term versus long term objectives.

Lex Fridman (1:11:13.820)

Can you wake up every day and truly enjoy what you're doing

Lex Fridman (1:11:19.420)

knowing that it's gonna be painful?

Kevin Systrom (1:11:22.240)

Knowing that like, no matter what you choose,

Lex Fridman (1:11:24.060)

it's gonna be painful.

Kevin Systrom (1:11:25.500)

Whether you sit on a beach

Lex Fridman (1:11:26.940)

or whether you manage a thousand people or 10,000,

Kevin Systrom (1:11:29.980)

it's gonna be painful.

Lex Fridman (1:11:31.300)

So choose something that's fun to have pain.

Lex Fridman (1:11:35.180)

But yes, there was a lot of, we have arrived

Lex Fridman (1:11:38.540)

and it's a maturation process you just have to go through.

Lex Fridman (1:11:44.420)

So no matter how much success there is,

Lex Fridman (1:11:47.060)

how much money you make,

Kevin Systrom (1:11:48.100)

you have to wake up the next day and choose the hard life,

Lex Fridman (1:11:51.300)

whatever that means next, that's fun.

Kevin Systrom (1:11:53.780)

The fun slash hard life, hard life that's fun.

Lex Fridman (1:11:56.680)

I guess what I'm trying to say is slightly different,

Kevin Systrom (1:12:00.420)

which is just that no one realizes

Lex Fridman (1:12:02.980)

everything's gonna be hard.

Kevin Systrom (1:12:05.380)

Even chilling out is hard.

Lex Fridman (1:12:07.780)

And then you just start worrying about stupid stuff.

Kevin Systrom (1:12:09.980)

Like, I don't know, like did so and so

Lex Fridman (1:12:12.780)

forget to paint the house today

Lex Fridman (1:12:14.420)

or like did the gardener come or whatever?

Lex Fridman (1:12:16.260)

Like, or, oh, I'm so angry

Lex Fridman (1:12:18.400)

and my shipment of wine didn't show up

Lex Fridman (1:12:20.160)

and I'm sitting here on the beach without my wine.

Kevin Systrom (1:12:22.180)

I don't know, I'm making shit up now, but like.

Lex Fridman (1:12:24.160)

It turns out that even chilling, AKA meditation,

Kevin Systrom (1:12:27.060)

is hard work.

Lex Fridman (1:12:27.900)

Yeah, and at least meditation is like productive chilling

Kevin Systrom (1:12:31.020)

where you're like actually training yourself

Lex Fridman (1:12:32.600)

to calm down and be, but backing up for a moment,

Kevin Systrom (1:12:36.280)

everything's hard.

Lex Fridman (1:12:37.340)

You might as well be like playing the game you love to play.

Kevin Systrom (1:12:41.100)

I just like playing and winning and I'm on the,

Lex Fridman (1:12:47.500)

I'm still on the, I think the first half of life,

Kevin Systrom (1:12:49.780)

knock on wood, and I've got a lot of years

Lex Fridman (1:12:53.300)

and what am I gonna do, sit around?

Lex Fridman (1:12:54.740)

And the other way of looking at this, by the way,

Lex Fridman (1:12:59.740)

imagine you made one movie and it was great.

Lex Fridman (1:13:02.420)

Would you just like stop making movies?

Lex Fridman (1:13:05.140)

No, generally you're like, wow,

Kevin Systrom (1:13:06.960)

I really like making movies, let's make another one.

Lex Fridman (1:13:09.380)

A lot of times, by the way,

Kevin Systrom (1:13:10.380)

the second one or the third one, not that great,

Lex Fridman (1:13:11.980)

but the fourth one, awesome.

Lex Fridman (1:13:13.980)

And no one forgets the second,

Lex Fridman (1:13:15.180)

or everyone forgets the second and the third one.

Lex Fridman (1:13:17.580)

So there's just this constant process of like,

Lex Fridman (1:13:20.740)

can I produce and is that fun?

Lex Fridman (1:13:23.740)

Is that exciting?

Lex Fridman (1:13:24.580)

What else can I learn?

Lex Fridman (1:13:25.400)

So this machine learning stuff for me

Lex Fridman (1:13:26.860)

has been this awesome new chapter of being like,

Kevin Systrom (1:13:31.740)

man, that's something I didn't understand at all.

Lex Fridman (1:13:33.680)

And now I feel like I'm one 10th of the way there.

Lex Fridman (1:13:37.220)

And that feels like a big mountain to climb.

Lex Fridman (1:13:40.280)

So I distracted us from the original question.

Kevin Systrom (1:13:42.660)

No, and we'll return to the machine learning

Lex Fridman (1:13:44.580)

cause I'd love to explore your interest there.

Lex Fridman (1:13:46.560)

But I mean, speaking of sort of challenges and hard things,

Lex Fridman (1:13:50.100)

is there a possible world

Kevin Systrom (1:13:52.740)

where sitting in a room with Mark Zuckerberg

Lex Fridman (1:13:56.180)

with a $1 billion deal, you turn it down?

Kevin Systrom (1:14:00.040)

Yeah, of course.

Lex Fridman (1:14:01.180)

What does that world look like?

Lex Fridman (1:14:03.260)

Why would you turn it down?

Lex Fridman (1:14:04.660)

Why did you take it?

Lex Fridman (1:14:05.660)

What was the calculation that you were making?

Lex Fridman (1:14:08.500)

Thus enters the world of counterfactuals

Lex Fridman (1:14:11.540)

and not really knowing.

Lex Fridman (1:14:13.220)

And if only we could run that experiment.

Kevin Systrom (1:14:15.020)

Well, the universe exists,

Lex Fridman (1:14:16.500)

it's just running in parallel to our own.

Lex Fridman (1:14:18.700)

Yeah, it's so fascinating, right?

Lex Fridman (1:14:23.140)

I mean, we're talking a lot about money,

Lex Fridman (1:14:24.780)

but the real question was,

Lex Fridman (1:14:28.980)

I'm not sure you'll believe me when I say this,

Lex Fridman (1:14:30.540)

but could we strap our little company

Lex Fridman (1:14:33.700)

onto the side of a rocket ship

Lex Fridman (1:14:36.140)

and like get out to a lot of people really, really quickly

Lex Fridman (1:14:39.560)

with the support, with the talent of a place like Facebook?

Kevin Systrom (1:14:45.220)

I mean, people often ask me what I would do differently

Lex Fridman (1:14:48.220)

at Instagram today.

Lex Fridman (1:14:49.060)

And I say, well, I'd probably hire more carefully

Lex Fridman (1:14:51.220)

because we showed up just like before I knew it,

Kevin Systrom (1:14:53.740)

we had like a hundred people on the team

Lex Fridman (1:14:56.380)

and 200, then 300.

Kevin Systrom (1:14:57.780)

I don't know where all these people were coming from.

Lex Fridman (1:14:59.560)

I never had to recruit them.

Kevin Systrom (1:15:00.780)

I never had to screen them.

Lex Fridman (1:15:02.380)

They were just like internal transfers, right?

Lex Fridman (1:15:05.100)

So it's like relying on the Facebook hiring machine,

Lex Fridman (1:15:07.460)

which is quite sort of, I mean, it's an elaborate machine.

Kevin Systrom (1:15:11.780)

It's great, by the way.

Lex Fridman (1:15:12.740)

They have really talented people there.

Lex Fridman (1:15:15.260)

But my point is the choice was like, take this thing,

Lex Fridman (1:15:21.420)

put it on the side of a rocket ship

Kevin Systrom (1:15:22.900)

that you know is growing very quickly.

Lex Fridman (1:15:25.820)

Like I had seen what had happened

Kevin Systrom (1:15:27.240)

when Ev sold Blogger to Google and then Google went public.

Lex Fridman (1:15:30.860)

Remember we sold before Facebook went public.

Kevin Systrom (1:15:34.580)

There was a moment at which the stock price was $17,

Lex Fridman (1:15:37.660)

by the way, Facebook stock price was $17.

Lex Fridman (1:15:40.180)

I remember thinking, what the fuck did I just do, right?

Lex Fridman (1:15:43.340)

Now at 320 ish, I don't know where we are today.

Lex Fridman (1:15:48.220)

But like, okay, like the best thing by the way

Lex Fridman (1:15:52.940)

is like when the stock is down, everyone calls you a dope.

Lex Fridman (1:15:56.780)

And then when it's up, they also call you a dope,

Lex Fridman (1:15:58.660)

but just for a different reason, right?

Kevin Systrom (1:16:01.480)

Like you can't win.

Lex Fridman (1:16:02.540)

Less than in there somewhere.

Kevin Systrom (1:16:03.860)

Yeah.

Lex Fridman (1:16:04.700)

So, but you know, the choice is to strap yourself

Kevin Systrom (1:16:06.700)

to a rocket ship or to build your own.

Lex Fridman (1:16:08.660)

You know, Mr. Elon built his own

Kevin Systrom (1:16:11.100)

literally with a rocket ship.

Lex Fridman (1:16:13.700)

That's a difficult choice because there's a world.

Kevin Systrom (1:16:18.380)

Actually, I would say something different,

Lex Fridman (1:16:19.820)

which is Elon and others decided to sell PayPal

Kevin Systrom (1:16:23.260)

for not that much.

Lex Fridman (1:16:25.140)

I mean, how much was it about a billion dollars?

Kevin Systrom (1:16:26.920)

I can't remember.

Lex Fridman (1:16:27.760)

Something like that, yeah.

Kevin Systrom (1:16:28.580)

I mean, it was early and,

Lex Fridman (1:16:30.260)

but it's worth a lot more now.

Kevin Systrom (1:16:31.580)

To then build a new rocket ship.

Lex Fridman (1:16:33.160)

So this is the cool part, right?

Kevin Systrom (1:16:34.820)

If you are an entrepreneur

Lex Fridman (1:16:37.340)

and you own a controlling stake in the company,

Kevin Systrom (1:16:40.460)

not only is it really hard to do something else

Lex Fridman (1:16:42.740)

with your life because all of the, you know,

Kevin Systrom (1:16:45.660)

value is tied up in you as a personality

Lex Fridman (1:16:48.520)

attached to this company, right?

Lex Fridman (1:16:50.960)

But if you sell it and you're getting yourself enough capital

Lex Fridman (1:16:53.740)

and you like have enough energy,

Kevin Systrom (1:16:55.880)

you can do another thing or 10 other things

Lex Fridman (1:16:58.020)

or in Elon's case, like a bunch of other things.

Kevin Systrom (1:17:00.020)

I don't know, like I lost count at this point.

Lex Fridman (1:17:03.380)

And it might've seemed silly at the time.

Lex Fridman (1:17:05.580)

And sure, like if you look back,

Lex Fridman (1:17:07.300)

man, PayPal is worth a lot now, right?

Lex Fridman (1:17:10.500)

But I don't know.

Lex Fridman (1:17:11.340)

Like, do you think Elon like cares about like,

Lex Fridman (1:17:13.940)

are we gonna buy Pinterest or not?

Lex Fridman (1:17:15.260)

Like, I just, he is,

Kevin Systrom (1:17:17.540)

he created a massive capital

Lex Fridman (1:17:20.900)

that allowed him to do what he wants to do.

Lex Fridman (1:17:23.300)

And that's awesome.

Lex Fridman (1:17:24.120)

That's more freeing than anything

Kevin Systrom (1:17:25.500)

because when you are an entrepreneur attached to a company,

Lex Fridman (1:17:28.380)

you gotta stay at that company for a really long time.

Kevin Systrom (1:17:30.720)

It's really hard to remove yourself.

Lex Fridman (1:17:32.900)

But I'm not sure how much he loved

Kevin Systrom (1:17:34.500)

PayPal versus SpaceX and Tesla.

Lex Fridman (1:17:37.460)

I have a sense that you love Instagram.

Kevin Systrom (1:17:41.420)

Yeah, I loved enough to like work

Lex Fridman (1:17:43.620)

for six years beyond the deal.

Kevin Systrom (1:17:44.940)

Which is rare, which is very rare.

Lex Fridman (1:17:46.700)

You chose.

Lex Fridman (1:17:47.540)

But can I tell you why?

Lex Fridman (1:17:48.720)

Sure.

Kevin Systrom (1:17:49.560)

It's, please.

Lex Fridman (1:17:51.180)

There are not a lot of companies that you can be part of

Kevin Systrom (1:17:55.100)

where the Pope's like,

Lex Fridman (1:17:57.380)

I would like to sign up for your product.

Kevin Systrom (1:17:59.260)

Like I'm not a religious person at all.

Lex Fridman (1:18:01.540)

I'm really not.

Kevin Systrom (1:18:02.780)

Yeah.

Lex Fridman (1:18:03.620)

You go to the Vatican and you're like walking

Kevin Systrom (1:18:05.380)

among giant columns and you're hearing the music

Lex Fridman (1:18:07.940)

and everything and like the Pope walks in

Lex Fridman (1:18:10.540)

and he wants to press the signup button on your product.

Lex Fridman (1:18:12.620)

Like it's a moment in life, okay?

Lex Fridman (1:18:15.040)

No matter what your persuasion, okay?

Lex Fridman (1:18:19.380)

The number of doors and experiences that that opened up

Kevin Systrom (1:18:22.540)

was, it was incredible.

Lex Fridman (1:18:23.920)

I mean, the people I got to meet, the places I got to go,

Kevin Systrom (1:18:27.820)

I assume maybe like a payments company

Lex Fridman (1:18:30.140)

is slightly different, right?

Lex Fridman (1:18:32.060)

But that's why, like it was so fun.

Lex Fridman (1:18:34.780)

And plus I truly believed we were building

Kevin Systrom (1:18:37.420)

such a great product and I loved, loved the game.

Lex Fridman (1:18:40.860)

It wasn't about money.

Kevin Systrom (1:18:41.980)

It was about the game.

Lex Fridman (1:18:43.560)

Do you think you had the guts to say no?

Kevin Systrom (1:18:46.840)

Is that, so here's, I often think about this,

Lex Fridman (1:18:49.360)

like how hard is it for an entrepreneur to say no?

Kevin Systrom (1:18:52.500)

Because the peer pressure.

Lex Fridman (1:18:54.900)

So every, like basically the sea of entrepreneurs

Kevin Systrom (1:18:57.900)

in Silicon Valley are gonna tell you,

Lex Fridman (1:18:59.740)

I mean, this is their dream.

Kevin Systrom (1:19:01.680)

The thing you were sitting before is a dream.

Lex Fridman (1:19:05.200)

To walk away from that is really,

Kevin Systrom (1:19:07.460)

it seems like nearly impossible.

Lex Fridman (1:19:11.080)

Because Instagram could in 10 years be,

Kevin Systrom (1:19:16.060)

you could talk about Google.

Lex Fridman (1:19:17.420)

You could be making self driving cars

Lex Fridman (1:19:20.260)

and building rockets that go to Mars

Lex Fridman (1:19:21.920)

and compete with SpaceX.

Kevin Systrom (1:19:23.620)

Totally.

Lex Fridman (1:19:24.460)

And so that's an interesting decision to say,

Lex Fridman (1:19:28.220)

am I willing to risk it?

Lex Fridman (1:19:31.460)

And the reason I also say it's an interesting decision

Kevin Systrom (1:19:34.480)

because it feels like per our previous discussion,

Lex Fridman (1:19:37.760)

if you're launching a social network company,

Kevin Systrom (1:19:42.140)

there's going to be that meeting, whatever that number is.

Lex Fridman (1:19:45.260)

If you're successful, if you're on this rocket ship

Kevin Systrom (1:19:48.100)

of success, there's going to be a meeting

Lex Fridman (1:19:50.740)

with one of the social media, social network companies

Kevin Systrom (1:19:53.740)

that wanna buy you, whether it's Facebook or Twitter,

Lex Fridman (1:19:58.280)

but it could also very well be Google

Kevin Systrom (1:20:00.260)

who seems to have like a graveyard

Lex Fridman (1:20:04.540)

of failed social networks.

Lex Fridman (1:20:06.620)

And it's, I mean, I don't know.

Lex Fridman (1:20:08.940)

I think about that, how difficult it is

Kevin Systrom (1:20:11.500)

for an entrepreneur to make that decision.

Lex Fridman (1:20:13.700)

How many have successfully made that decision?

Kevin Systrom (1:20:15.920)

I guess, this is a big question.

Lex Fridman (1:20:18.220)

It's sad to me, to be honest,

Kevin Systrom (1:20:20.380)

that too many make that decision,

Lex Fridman (1:20:21.980)

perhaps for the wrong reason.

Kevin Systrom (1:20:23.660)

Sorry, when you say make the decision,

Lex Fridman (1:20:25.540)

you mean to the affirmative.

Kevin Systrom (1:20:26.700)

To the affirmative, yeah.

Lex Fridman (1:20:27.620)

Got it, yeah.

Lex Fridman (1:20:28.620)

There are also companies that don't sell, right?

Lex Fridman (1:20:31.420)

And take the PATH and say, we're gonna be independent.

Lex Fridman (1:20:34.860)

And then you've never heard of them again.

Lex Fridman (1:20:37.340)

Like I remember PATH, right?

Kevin Systrom (1:20:40.380)

Was one of our competitors early on.

Lex Fridman (1:20:43.340)

There was a big moment when they had,

Kevin Systrom (1:20:45.820)

I can't remember what it was,

Lex Fridman (1:20:46.860)

like $110 million offer from Google or something.

Kevin Systrom (1:20:50.780)

It might've been larger, I don't know.

Lex Fridman (1:20:54.260)

And I remember there was like this big TechCrunch article

Kevin Systrom (1:20:56.580)

that was like, they turned it down after talking deeply

Lex Fridman (1:20:59.700)

about their values and everything.

Lex Fridman (1:21:01.800)

And I don't know the inner workings of Foursquare,

Lex Fridman (1:21:05.760)

but I'm certain there were many conversations over time

Kevin Systrom (1:21:09.240)

where there were companies that wanted Foursquare as well.

Lex Fridman (1:21:11.780)

Recently, I mean, what other companies?

Lex Fridman (1:21:15.940)

There's Clubhouse, right?

Lex Fridman (1:21:16.980)

Like, I don't know.

Kevin Systrom (1:21:17.820)

Maybe people were really interested in them too.

Lex Fridman (1:21:20.580)

Like there are plenty of moments where people say no

Lex Fridman (1:21:25.580)

and we just forget that those things happen.

Lex Fridman (1:21:28.500)

We only focus on the ones where like they said yes

Lex Fridman (1:21:33.060)

and like, wow, like what if they had stayed independent?

Lex Fridman (1:21:36.300)

So I don't know.

Kevin Systrom (1:21:37.140)

I used to think a lot about this and now I just don't

Lex Fridman (1:21:39.740)

because I'm like, whatever.

Kevin Systrom (1:21:42.700)

Things have gone pretty well.

Lex Fridman (1:21:44.820)

I'm ready for the next game.

Kevin Systrom (1:21:45.980)

I mean, think about an athlete where, I don't know,

Lex Fridman (1:21:50.780)

maybe they do something wrong in the World Series

Kevin Systrom (1:21:53.700)

or whatever.

Lex Fridman (1:21:54.540)

If you let it haunt you for the rest of your career,

Kevin Systrom (1:21:57.580)

like why not just be like, I don't know, it was a game.

Lex Fridman (1:22:00.380)

Next game, next shot, right?

Lex Fridman (1:22:02.940)

And if you just move to that world,

Lex Fridman (1:22:05.100)

like at least I have a next shot, right?

Kevin Systrom (1:22:07.460)

No, that's beautiful, but I mean, just insights

Lex Fridman (1:22:10.740)

and it's funny you brought up Clubhouse.

Kevin Systrom (1:22:12.540)

It is very true.

Lex Fridman (1:22:14.220)

It seems like Clubhouse is on the downward path

Lex Fridman (1:22:19.500)

and it's very possible to see a billion plus dollar deal

Lex Fridman (1:22:23.300)

at some stage, maybe like a year ago or half a year ago

Kevin Systrom (1:22:27.220)

from Facebook, from Google.

Lex Fridman (1:22:28.580)

I think Facebook was flirting with that idea too.

Lex Fridman (1:22:30.780)

And I think a lot of companies probably were.

Lex Fridman (1:22:34.780)

I wish it was more public.

Lex Fridman (1:22:37.140)

You know what?

Lex Fridman (1:22:37.980)

There's not like a bad public story

Kevin Systrom (1:22:41.260)

about them making the decision to walk away.

Lex Fridman (1:22:43.940)

We just don't hear about it.

Lex Fridman (1:22:45.020)

And then we get to see the results of that success

Lex Fridman (1:22:47.540)

or the failure, more often failure.

Lex Fridman (1:22:49.300)

So a couple of things, one is,

Lex Fridman (1:22:52.420)

I would not assume Clubhouse is down for the count at all.

Kevin Systrom (1:22:54.900)

They're young, they have plenty of money,

Lex Fridman (1:22:56.580)

they're run by really smart people.

Kevin Systrom (1:22:59.620)

I'd give them like a very fighting chance to figure it out.

Lex Fridman (1:23:02.420)

There are a lot of times when people call Twitter

Kevin Systrom (1:23:04.180)

down for the count and they figure it out

Lex Fridman (1:23:05.860)

and they seem to be doing well, right?

Lex Fridman (1:23:08.340)

So just backing up like,

Lex Fridman (1:23:09.780)

and not knowing anything about their internals,

Kevin Systrom (1:23:11.820)

like there's a strong chance they will figure it out

Lex Fridman (1:23:15.580)

and that people are just down

Kevin Systrom (1:23:16.780)

because they like being down about companies.

Lex Fridman (1:23:18.500)

They like assuming that they're gonna fail.

Lex Fridman (1:23:20.060)

So who knows, right?

Lex Fridman (1:23:21.220)

But let's take the ones in the past

Kevin Systrom (1:23:22.580)

where like we know how it played out.

Lex Fridman (1:23:24.740)

There are plenty of examples

Kevin Systrom (1:23:25.940)

where people have turned down big offers

Lex Fridman (1:23:28.140)

and then you've just never heard from them again,

Lex Fridman (1:23:30.940)

but we never focus on the companies

Lex Fridman (1:23:32.420)

because you just forget that those were big.

Lex Fridman (1:23:34.660)

But inside your psyche,

Lex Fridman (1:23:38.260)

I think it's easy for someone with enough money

Kevin Systrom (1:23:42.220)

to say money doesn't matter, which I think is like,

Lex Fridman (1:23:45.060)

it's bullshit.

Kevin Systrom (1:23:45.900)

Of course, money matters to people, but at the moment,

Lex Fridman (1:23:50.220)

you just can't even grasp like the number of zeros

Kevin Systrom (1:23:53.100)

that you're talking about.

Lex Fridman (1:23:53.940)

It just doesn't make sense, right?

Lex Fridman (1:23:56.660)

So to think rationally in that moment

Lex Fridman (1:23:58.580)

is not something many people are equipped to do,

Kevin Systrom (1:24:01.380)

especially not people

Lex Fridman (1:24:03.020)

where I think we had founded the company a year earlier,

Kevin Systrom (1:24:05.300)

maybe two years earlier, like a year and a half,

Lex Fridman (1:24:07.540)

we were 13 people, but I will say,

Kevin Systrom (1:24:11.940)

I still don't know if it was the right decision

Lex Fridman (1:24:14.460)

because I don't have that counterfactual.

Kevin Systrom (1:24:16.020)

I don't know that other world.

Lex Fridman (1:24:18.620)

I'm just thankful that by and large,

Kevin Systrom (1:24:21.260)

most people love Instagram, still do.

Lex Fridman (1:24:23.780)

By and large, people are very happy

Kevin Systrom (1:24:25.540)

with like the time we had there

Lex Fridman (1:24:28.220)

and I'm proud of what we built.

Lex Fridman (1:24:29.700)

So like, I'm cool.

Lex Fridman (1:24:31.660)

Like now it's next shot, right?

Kevin Systrom (1:24:35.620)

Well, if we could just linger on this Yankees versus Red Sox,

Lex Fridman (1:24:40.380)

the fun of it, the competition over,

Kevin Systrom (1:24:42.460)

I would say over the space of features.

Lex Fridman (1:24:45.380)

So there are a bunch of features,

Kevin Systrom (1:24:49.220)

like there's photos, there's one minute videos on Instagram,

Lex Fridman (1:24:54.500)

there's IGTV, there's stories, there's reels, there's live.

Lex Fridman (1:24:58.860)

So that sounds like it's like a long list

Lex Fridman (1:25:01.940)

of too much stuff, but it's not

Kevin Systrom (1:25:04.340)

because it feels like they're close together,

Lex Fridman (1:25:07.540)

but they're somehow, like what we're saying,

Kevin Systrom (1:25:09.620)

fundamentally distinct,

Lex Fridman (1:25:10.780)

like each of the things I mentioned.

Kevin Systrom (1:25:13.340)

Maybe can you describe the philosophies,

Lex Fridman (1:25:15.940)

the design philosophies behind some of these,

Lex Fridman (1:25:17.860)

how you were thinking about it

Lex Fridman (1:25:19.580)

during the historic war between Snapchat and Instagram

Kevin Systrom (1:25:24.860)

or just in general,

Lex Fridman (1:25:25.820)

like this space of features that was discovered.

Kevin Systrom (1:25:30.100)

There's this great book by Clay Christensen

Lex Fridman (1:25:34.340)

called, Competing Against Luck.

Kevin Systrom (1:25:36.260)

It's like a terrible title,

Lex Fridman (1:25:39.260)

but within it, there's effectively an expression

Kevin Systrom (1:25:43.460)

of this thing called jobs to be done theory.

Lex Fridman (1:25:46.580)

And it's unclear if like he came up with it

Kevin Systrom (1:25:48.980)

or some of his colleagues,

Lex Fridman (1:25:50.020)

but there are a bunch of places you can find

Kevin Systrom (1:25:52.060)

with people claiming to have come up

Lex Fridman (1:25:53.300)

with this jobs to be done theory.

Lex Fridman (1:25:55.140)

But the idea is if you like zoom out

Lex Fridman (1:25:59.140)

and you look at your product,

Lex Fridman (1:26:00.380)

you ask yourself, why are people hiring your product?

Lex Fridman (1:26:04.940)

Like imagine every product in your life

Kevin Systrom (1:26:07.020)

is effectively an employee, you know,

Lex Fridman (1:26:10.140)

you're CEO of your life

Lex Fridman (1:26:11.700)

and you hire products to be employees effectively.

Lex Fridman (1:26:13.860)

They all have roles and jobs, right?

Lex Fridman (1:26:16.660)

Why are you hiring a product?

Lex Fridman (1:26:18.380)

Why do you want that product

Lex Fridman (1:26:19.540)

to perform something in your life?

Lex Fridman (1:26:21.020)

And like, what are the hidden reasons

Lex Fridman (1:26:22.660)

why you're in love with this product?

Lex Fridman (1:26:26.740)

Instagram was about sharing your life

Lex Fridman (1:26:28.780)

with others visually, period, right?

Lex Fridman (1:26:31.620)

Why? Because you feel connected with them.

Kevin Systrom (1:26:34.340)

You get to show off.

Lex Fridman (1:26:35.620)

You get to feel good and cared about, right?

Kevin Systrom (1:26:38.900)

With likes and it turns out that that will,

Lex Fridman (1:26:43.900)

I think forever define Instagram

Lex Fridman (1:26:46.500)

and any product that serves that job

Lex Fridman (1:26:49.740)

is going to do very well, okay?

Kevin Systrom (1:26:52.420)

Stories let's take as an example

Lex Fridman (1:26:55.180)

is very much serving that job.

Kevin Systrom (1:26:58.060)

In fact, it serves it better than the original product

Lex Fridman (1:27:00.940)

because when you're large and have an enormous audience,

Kevin Systrom (1:27:04.300)

you're worried about people seeing your stuff

Lex Fridman (1:27:08.020)

or you're worried about being permanent

Lex Fridman (1:27:09.500)

so that a college admissions person

Lex Fridman (1:27:11.540)

is going to see your photo of you doing something.

Lex Fridman (1:27:13.460)

And so it turns out that that is a more efficient way

Lex Fridman (1:27:16.620)

of performing that job than the original product was.

Kevin Systrom (1:27:19.780)

The original product still has its value,

Lex Fridman (1:27:22.620)

but at scale, these two things together

Kevin Systrom (1:27:24.660)

work really, really well.

Lex Fridman (1:27:26.380)

Now, I will claim that other parts of the product

Kevin Systrom (1:27:29.780)

over time didn't perform that job as well.

Lex Fridman (1:27:32.740)

I think IGTV probably didn't, right?

Kevin Systrom (1:27:35.580)

Shopping is like completely unrelated

Lex Fridman (1:27:38.260)

to what I just described, but it might work.

Lex Fridman (1:27:40.740)

I don't know, right?

Lex Fridman (1:27:43.060)

Products I think, products that succeed

Kevin Systrom (1:27:46.740)

are products that all share this parent node

Lex Fridman (1:27:49.500)

of like this job to be done that is in common.

Lex Fridman (1:27:52.060)

And then they're just like different ways of doing it, right?

Lex Fridman (1:27:55.380)

Apple, I think does a great job with this, right?

Kevin Systrom (1:27:57.620)

It's like managing your digital life

Lex Fridman (1:27:59.660)

and all the products just work together.

Lex Fridman (1:28:01.900)

They sync, they're like, it's beautiful, right?

Lex Fridman (1:28:06.140)

Even if they require like silly specific cords to work,

Lex Fridman (1:28:10.420)

but they're all part of a system.

Lex Fridman (1:28:12.900)

It's when you leave that system

Lex Fridman (1:28:14.220)

and you start doing something weird

Lex Fridman (1:28:15.580)

that people start scratching their head

Lex Fridman (1:28:17.780)

and I think you are less successful.

Lex Fridman (1:28:19.260)

So I think one of the challenges

Kevin Systrom (1:28:20.460)

Facebook has had throughout its life

Lex Fridman (1:28:22.460)

is that it has never fully, I think,

Kevin Systrom (1:28:24.580)

appreciated the job to be done of the main product.

Lex Fridman (1:28:28.100)

And what it's done is said,

Kevin Systrom (1:28:29.220)

ooh, there's a shiny object over there.

Lex Fridman (1:28:31.420)

That startup's getting some traction.

Kevin Systrom (1:28:32.780)

Let's go copy that thing.

Lex Fridman (1:28:34.460)

And then they're confused why it doesn't work.

Lex Fridman (1:28:36.900)

Like why doesn't it work?

Lex Fridman (1:28:37.940)

It's because the people who show up for this

Kevin Systrom (1:28:40.100)

don't want that, it's different.

Lex Fridman (1:28:42.620)

What's the purpose of Facebook?

Lex Fridman (1:28:44.300)

So I remember I was a very early Facebook user.

Lex Fridman (1:28:47.500)

The reason I was personally excited about Facebook

Kevin Systrom (1:28:50.700)

is because you can, first of all, use your real name.

Lex Fridman (1:28:55.580)

Like I can exist in this world.

Kevin Systrom (1:28:58.060)

It can be like formally exist.

Lex Fridman (1:29:00.820)

I like anonymity for certain things, Reddit and so on,

Lex Fridman (1:29:04.780)

but I want it to also exist not anonymously

Lex Fridman (1:29:09.260)

so that I can connect with other friends of mine

Kevin Systrom (1:29:12.260)

nonanonymously and there's a reliable way to know

Lex Fridman (1:29:16.020)

that I'm real and they're real and that we're connecting.

Lex Fridman (1:29:19.940)

And it's kind of like, I liked it for the reasons

Lex Fridman (1:29:24.780)

that people like LinkedIn, I guess.

Lex Fridman (1:29:27.100)

But like without the form, like not everybody is dressed up

Lex Fridman (1:29:30.900)

and being super polite, like more like with friends.

Lex Fridman (1:29:34.580)

But then it became something much bigger than that,

Lex Fridman (1:29:37.540)

I suppose, there's a feed.

Kevin Systrom (1:29:39.300)

It became this, I mean, it became a place

Lex Fridman (1:29:44.380)

to discover content, to share content

Kevin Systrom (1:29:49.940)

that's not just about connecting directly with friends.

Lex Fridman (1:29:53.740)

I mean, it became something else.

Kevin Systrom (1:29:54.740)

I don't even know what it is really.

Lex Fridman (1:29:56.540)

So you said Instagram is a place

Kevin Systrom (1:29:58.820)

where you visually share your life.

Lex Fridman (1:30:02.100)

What is Facebook?

Kevin Systrom (1:30:03.620)

Well, let's go back to the founding of Facebook

Lex Fridman (1:30:06.140)

and why it worked really well initially at Harvard

Lex Fridman (1:30:09.660)

and then Dartmouth and Stanford and I can't remember,

Lex Fridman (1:30:12.020)

probably MIT, there were like a handful of schools

Lex Fridman (1:30:14.620)

in that first tranche, right?

Lex Fridman (1:30:17.940)

It worked because there are communities

Kevin Systrom (1:30:19.820)

that exist in the world that want to transact.

Lex Fridman (1:30:23.300)

And when I say transact, I don't mean commercially,

Kevin Systrom (1:30:25.340)

I just mean they want to share, they want to coordinate,

Lex Fridman (1:30:28.380)

they want to communicate, they want a space for themselves.

Lex Fridman (1:30:32.020)

And Facebook at its best, I think is that.

Lex Fridman (1:30:36.340)

And if actually you look at the most popular products

Kevin Systrom (1:30:39.180)

that Facebook has built over time,

Lex Fridman (1:30:42.060)

if you look at things like groups, marketplace,

Kevin Systrom (1:30:45.260)

groups is enormous.

Lex Fridman (1:30:47.500)

And groups is effectively like everyone can found

Lex Fridman (1:30:50.340)

their own little Stanford or Dartmouth or MIT, right?

Lex Fridman (1:30:53.460)

And find each other and share and communicate

Kevin Systrom (1:30:57.060)

about something that matters deeply to them.

Lex Fridman (1:31:00.340)

That is the core of what Facebook was built around.

Lex Fridman (1:31:03.860)

And I think today is where it stands most strongly.

Lex Fridman (1:31:09.060)

Yeah, it's brilliant.

Kevin Systrom (1:31:09.900)

The groups, I wish groups were done better.

Lex Fridman (1:31:13.620)

It feels like it's not a first class citizen.

Kevin Systrom (1:31:16.060)

I know I may be saying something without much knowledge,

Lex Fridman (1:31:19.780)

but it feels like it's kind of bolted on

Kevin Systrom (1:31:24.020)

while being used a lot.

Lex Fridman (1:31:26.140)

It feels like there needs to be a little bit more structure

Kevin Systrom (1:31:28.940)

in terms of discovery, in terms of like.

Lex Fridman (1:31:32.820)

I mean, look at Reddit.

Kevin Systrom (1:31:33.780)

Like Reddit is basically groups of public and open

Lex Fridman (1:31:36.540)

and a little bit crazy, right?

Kevin Systrom (1:31:38.620)

In a good way.

Lex Fridman (1:31:40.740)

But there's clear product market fit

Kevin Systrom (1:31:42.420)

for that specific use case.

Lex Fridman (1:31:44.860)

And it doesn't have to be a college, it can be anything.

Kevin Systrom (1:31:47.060)

It can be a small group, a big group,

Lex Fridman (1:31:48.460)

it can be group messaging.

Kevin Systrom (1:31:50.180)

Facebook shines, I think, when it leans into that.

Lex Fridman (1:31:53.220)

I think when there are other companies

Kevin Systrom (1:31:56.620)

that just seem exciting and now all of a sudden

Lex Fridman (1:32:00.380)

the product shifts in some fundamental way

Kevin Systrom (1:32:02.380)

to go try to compete with that other thing,

Lex Fridman (1:32:05.820)

that's when I think consumers get confused.

Kevin Systrom (1:32:09.220)

Even if you can be successful,

Lex Fridman (1:32:10.660)

like even if you can compete with that other company,

Kevin Systrom (1:32:13.020)

even if you can figure out how to bolt it on,

Lex Fridman (1:32:16.220)

eventually you come back and you look at the app

Lex Fridman (1:32:17.980)

and you're like, I just don't know why I opened this app.

Lex Fridman (1:32:20.620)

Like why, like too many things going on.

Lex Fridman (1:32:23.180)

And that was always a worry.

Lex Fridman (1:32:24.580)

I mean, you listed all the things at Instagram

Lex Fridman (1:32:26.220)

and it almost gave me a heart attack,

Lex Fridman (1:32:27.580)

like way too many things.

Lex Fridman (1:32:30.460)

But I don't know, entrepreneurs get bored.

Lex Fridman (1:32:31.900)

They want to add things.

Lex Fridman (1:32:32.900)

They want to like, right?

Lex Fridman (1:32:35.980)

I don't have a good answer for it,

Kevin Systrom (1:32:37.300)

except for that I think being true to your original use case

Lex Fridman (1:32:41.500)

and not even original use case,

Lex Fridman (1:32:43.060)

but sorry, actually not use case, original job.

Lex Fridman (1:32:46.260)

There are many use cases under that job.

Kevin Systrom (1:32:49.340)

Being true to that and like being really good at it

Lex Fridman (1:32:52.220)

over time and morphing as needs change,

Kevin Systrom (1:32:57.180)

I think that's how to make a company last forever.

Lex Fridman (1:32:59.540)

And I mean, honestly, like my main thesis

Kevin Systrom (1:33:03.860)

about why Facebook is in the position it is today

Lex Fridman (1:33:07.540)

is if they have had a series of product launches

Kevin Systrom (1:33:12.540)

that delighted people over time,

Lex Fridman (1:33:16.340)

I think they'd be in a totally different world.

Lex Fridman (1:33:17.980)

So just like imagine for a moment,

Lex Fridman (1:33:20.420)

and by the way, Apple's entering this,

Lex Fridman (1:33:21.780)

but like Apple for so long,

Lex Fridman (1:33:23.860)

just like product after product,

Kevin Systrom (1:33:25.300)

you couldn't wait for it.

Lex Fridman (1:33:26.620)

You stood in line for it.

Kevin Systrom (1:33:27.900)

You talked about it.

Lex Fridman (1:33:28.740)

You got excited.

Kevin Systrom (1:33:29.940)

Amazon makes your life so easy.

Lex Fridman (1:33:31.780)

It's like, wow, I needed this thing

Lex Fridman (1:33:34.260)

and it showed up at my door two days later.

Lex Fridman (1:33:36.180)

And like both of these companies, by the way,

Lex Fridman (1:33:39.180)

Amazon, Apple have issues, right?

Lex Fridman (1:33:41.220)

There are labor issues,

Kevin Systrom (1:33:43.260)

whether it's here in the US or in China,

Lex Fridman (1:33:45.380)

there are environmental issues.

Lex Fridman (1:33:48.380)

But like when's the last time

Lex Fridman (1:33:49.620)

you heard like a large chorus being like,

Kevin Systrom (1:33:52.820)

these companies better pay

Lex Fridman (1:33:54.020)

for what they're doing on these things, right?

Kevin Systrom (1:33:56.220)

I think Facebook's main issue today is like,

Lex Fridman (1:33:59.940)

you need to produce a hit.

Kevin Systrom (1:34:01.740)

If you don't produce hits,

Lex Fridman (1:34:03.780)

it's really hard to keep consumers on your side.

Kevin Systrom (1:34:06.460)

Then people just start picking on you

Lex Fridman (1:34:08.300)

for a variety of reasons, whether it's right or wrong.

Kevin Systrom (1:34:11.300)

I'm not even going to place a judgment

Lex Fridman (1:34:12.580)

right here and right now.

Kevin Systrom (1:34:13.420)

I'm just going to say that it is way better

Lex Fridman (1:34:17.140)

to be in a world where you are producing hits

Lex Fridman (1:34:19.180)

and consumers love what you're doing

Lex Fridman (1:34:21.900)

because then they're on your side.

Lex Fridman (1:34:23.460)

And I think that's, it's the past 10 years

Lex Fridman (1:34:26.580)

for Facebook has been fairly hard on this dimension.

Kevin Systrom (1:34:29.900)

So, and by hits, it doesn't necessarily mean financial hits.

Lex Fridman (1:34:33.020)

It feels like to me, what you're saying

Kevin Systrom (1:34:34.620)

is something that brings joy.

Lex Fridman (1:34:37.300)

A product that brings joy

Kevin Systrom (1:34:38.380)

to some fraction of the population.

Lex Fridman (1:34:40.100)

Yeah, I mean, TikTok isn't just literally an algorithm.

Kevin Systrom (1:34:45.460)

In some ways, TikTok's content and algorithm

Lex Fridman (1:34:49.460)

have more sway now over the American psyche

Lex Fridman (1:34:53.420)

than Facebook's algorithm, right?

Lex Fridman (1:34:54.860)

It's visual, it's video.

Kevin Systrom (1:34:57.220)

By the way, it's not defined by who you follow.

Lex Fridman (1:34:59.500)

It's defined by some magical thing that,

Kevin Systrom (1:35:01.340)

by the way, if someone wanted to tweak

Lex Fridman (1:35:02.700)

to show you a certain type of content for some reason,

Kevin Systrom (1:35:05.540)

they could, right, but people love it.

Lex Fridman (1:35:10.780)

So as a CEO, let me ask you a question

Kevin Systrom (1:35:13.940)

because leadership matters.

Lex Fridman (1:35:17.700)

This is a complicated question.

Lex Fridman (1:35:19.900)

Why is Mark Zuckerberg distrusted, disliked

Lex Fridman (1:35:23.820)

and sometimes even hated by many people in public?

Kevin Systrom (1:35:28.220)

Right, that is a complicated question.

Lex Fridman (1:35:30.180)

Well, the premise, I'm not sure I agree with the premise.

Lex Fridman (1:35:34.740)

And I can expand that to include

Lex Fridman (1:35:38.100)

even a more mysterious question for me, Bill Gates.

Kevin Systrom (1:35:42.300)

Hmm.

Lex Fridman (1:35:44.780)

What is the Bill Gates version of the question?

Lex Fridman (1:35:47.380)

Do you think people hate Bill Gates?

Lex Fridman (1:35:49.660)

No, distrust.

Kevin Systrom (1:35:51.340)

Ah.

Lex Fridman (1:35:52.300)

So takeaway one, it's a checklist.

Kevin Systrom (1:35:56.780)

There is, I think Mark Zuckerberg's distrust

Lex Fridman (1:36:00.300)

is the primary one, but there's also like a dislike,

Kevin Systrom (1:36:04.220)

maybe hate is too strong a word,

Lex Fridman (1:36:05.580)

but it's just if you look at like the articles

Kevin Systrom (1:36:09.020)

that are being written and so on, there's a dislike.

Lex Fridman (1:36:13.500)

And it makes, it's confusing to me

Kevin Systrom (1:36:15.580)

because it's like the public picks certain individuals

Lex Fridman (1:36:18.540)

and they attach certain kinds of emotions

Kevin Systrom (1:36:21.940)

to those individuals.

Lex Fridman (1:36:22.940)

Yeah, so someone once just recently said,

Kevin Systrom (1:36:27.460)

there's a strong case that founder led companies

Lex Fridman (1:36:30.540)

have this problem and that a lot of Mark's issues

Kevin Systrom (1:36:33.980)

come today come from the fact that he is a visible founder

Lex Fridman (1:36:38.380)

with this story that people have watched in both a movie

Lex Fridman (1:36:42.940)

and they followed along and he's this boy wonder kid

Lex Fridman (1:36:45.500)

who became one of the world's richest people

Lex Fridman (1:36:47.540)

and he's no longer Mark the person,

Lex Fridman (1:36:50.540)

he's Mark this image of a person

Kevin Systrom (1:36:53.220)

with enormous wealth and power.

Lex Fridman (1:36:55.620)

And in today's world, we have issues

Kevin Systrom (1:36:59.020)

with enormous wealth and power for a variety of reasons.

Lex Fridman (1:37:02.260)

One of which is we've been stuck inside

Kevin Systrom (1:37:05.260)

for a year and a half, two years.

Lex Fridman (1:37:07.380)

One of which is a lot of people were really unhappy

Kevin Systrom (1:37:10.380)

about not the last election, but the last, last election.

Lex Fridman (1:37:13.860)

And where do you take out that anger?

Lex Fridman (1:37:15.220)

Who do you blame but the people in charge?

Lex Fridman (1:37:18.380)

That's one example or one reason why I think

Kevin Systrom (1:37:22.260)

a lot of people express anger or resentment

Lex Fridman (1:37:26.540)

or unhappiness with Mark.

Kevin Systrom (1:37:29.660)

At the same time, I don't know,

Lex Fridman (1:37:31.620)

I pointed out to that person, I was like, well,

Kevin Systrom (1:37:34.020)

I don't know, I think a lot of people really like Elon.

Lex Fridman (1:37:38.260)

Elon arguably, he kept his factory open here

Kevin Systrom (1:37:41.340)

throughout COVID protocols, which arguably

Lex Fridman (1:37:44.380)

a lot of people would be against.

Kevin Systrom (1:37:47.180)

While saying a bunch of crazy offensive things

Lex Fridman (1:37:50.660)

on the internet, they still.

Lex Fridman (1:37:52.300)

And basically gives the middle finger to the SEC

Lex Fridman (1:37:56.940)

on Twitter and I don't know, I'm like,

Kevin Systrom (1:37:59.500)

well, there's a founder and like people kind of like him.

Lex Fridman (1:38:02.900)

So I do think that the founder and slash CEO

Kevin Systrom (1:38:09.100)

of a company that's a social network company

Lex Fridman (1:38:11.660)

is like an extra level of difficulty.

Kevin Systrom (1:38:13.820)

If life is a video game,

Lex Fridman (1:38:15.420)

you just chose the harder video game.

Kevin Systrom (1:38:17.700)

So, I mean, that's why it's interesting to ask you

Lex Fridman (1:38:20.100)

because you were the founder and CEO of a social network.

Kevin Systrom (1:38:24.780)

I challenge it because.

Lex Fridman (1:38:25.860)

Exactly, but you're one of the rare examples.

Kevin Systrom (1:38:30.780)

Even Jack Dorsey's disliked, not to the degree,

Lex Fridman (1:38:34.500)

but it just seems harder

Kevin Systrom (1:38:35.860)

when you're running a social media company.

Lex Fridman (1:38:38.900)

It's interesting.

Kevin Systrom (1:38:39.740)

I never thought of Jack as just like,

Lex Fridman (1:38:41.380)

I think generally he's well respected.

Kevin Systrom (1:38:45.100)

Yeah, I think so.

Lex Fridman (1:38:45.940)

I think you're right, but he's not loved.

Kevin Systrom (1:38:50.540)

Yeah.

Lex Fridman (1:38:51.380)

And I feel like you, I mean, to me,

Kevin Systrom (1:38:53.260)

Twitter is an incredible thing.

Lex Fridman (1:38:54.860)

Yeah, again, can I just come back to this point,

Kevin Systrom (1:38:57.500)

which seems over simplistic,

Lex Fridman (1:38:59.180)

but I really do think how a product makes someone feel,

Kevin Systrom (1:39:04.940)

they ascribe that feeling to the founder.

Lex Fridman (1:39:07.420)

Yep.

Lex Fridman (1:39:09.900)

So make people feel good.

Lex Fridman (1:39:11.860)

So think about it.

Kevin Systrom (1:39:13.020)

Let's just go with this thesis for a second.

Lex Fridman (1:39:14.860)

Sure, I like it though.

Lex Fridman (1:39:17.140)

Amazon's pretty utilitarian, right?

Lex Fridman (1:39:19.580)

It delivers brown boxes to your front door.

Kevin Systrom (1:39:21.540)

Sure, you can have Alexa

Lex Fridman (1:39:23.420)

and you can have all these things, right?

Lex Fridman (1:39:25.540)

But in general, it delivers stuff quickly to you

Lex Fridman (1:39:28.980)

at a reasonable price, right?

Kevin Systrom (1:39:31.460)

I think Jeff Bezos is wonderfully wealthy,

Lex Fridman (1:39:34.220)

thoughtful, smart guy, right?

Lex Fridman (1:39:36.660)

But people kind of feel that way about him.

Lex Fridman (1:39:38.820)

They're like, wow, this is really big.

Kevin Systrom (1:39:40.300)

We're impressed that this is really big.

Lex Fridman (1:39:42.820)

But he's doing the same space stuff Elon's doing,

Lex Fridman (1:39:45.820)

but they don't necessarily ascribe

Lex Fridman (1:39:47.220)

the same sense of wonder, right?

Kevin Systrom (1:39:50.060)

Now let's take Elon.

Lex Fridman (1:39:51.300)

And again, this is pet theory.

Kevin Systrom (1:39:52.860)

I don't have much proof other than my own intuition.

Lex Fridman (1:39:56.460)

He is literally about living the future.

Kevin Systrom (1:39:59.740)

Mars, it's about wonder.

Lex Fridman (1:40:01.980)

It's about going back to that feeling as a kid

Kevin Systrom (1:40:04.060)

when you looked up to the stars and asked,

Lex Fridman (1:40:05.900)

is there life out there?

Kevin Systrom (1:40:08.140)

People get behind that because it's a sense of hope

Lex Fridman (1:40:11.060)

and excitement and innovation.

Lex Fridman (1:40:13.460)

And like, you can say whatever you want,

Lex Fridman (1:40:15.340)

but we ascribe that emotion to that person.

Kevin Systrom (1:40:18.860)

Now, let's say you're on a social network

Lex Fridman (1:40:21.140)

and people make you kind of angry

Kevin Systrom (1:40:23.220)

because they disagree with you

Lex Fridman (1:40:24.380)

or they say something ridiculous

Kevin Systrom (1:40:25.820)

or they're living a FOMO type life where you're like,

Lex Fridman (1:40:28.580)

wow, I wish I was doing that thing.

Kevin Systrom (1:40:31.780)

I think Instagram, if I were to think back,

Lex Fridman (1:40:33.980)

by and large, when I was there,

Kevin Systrom (1:40:35.900)

was not about FOMO, was not about this influencer economy,

Lex Fridman (1:40:39.780)

although it certainly became that way closer to the end.

Kevin Systrom (1:40:43.540)

It was about the sense of wonder and happiness

Lex Fridman (1:40:45.820)

and beautiful things in the world.

Lex Fridman (1:40:47.220)

And I don't know, I mean, like,

Lex Fridman (1:40:49.380)

I don't want to have a blind spot,

Lex Fridman (1:40:50.540)

but I don't think anyone had a strong opinion

Lex Fridman (1:40:52.500)

about me one way or the other.

Kevin Systrom (1:40:53.620)

For the longest time, the way people explained to me,

Lex Fridman (1:40:55.660)

I mean, if you want to go for toxicity,

Kevin Systrom (1:40:57.580)

you go to Facebook or Twitter.

Lex Fridman (1:40:59.380)

If you want to go to feel good about life,

Kevin Systrom (1:41:01.820)

you go to Instagram to enjoy, celebrate life.

Lex Fridman (1:41:04.460)

And my experience been talking to people

Kevin Systrom (1:41:05.860)

is they gave me the benefit of the doubt because of that.

Lex Fridman (1:41:08.340)

But if your experience of the product

Kevin Systrom (1:41:10.020)

is kind of makes you angry, it's where you argue.

Lex Fridman (1:41:13.660)

I mean, a big part of Jack might be

Kevin Systrom (1:41:15.200)

that he wasn't actually the CEO for a very long time

Lex Fridman (1:41:17.780)

and only became recently.

Lex Fridman (1:41:19.260)

So I'm not sure how much of the connection got made.

Lex Fridman (1:41:23.660)

But in general, I mean, if you hate,

Kevin Systrom (1:41:28.140)

I'm just thinking about other companies

Lex Fridman (1:41:29.620)

that aren't tech companies.

Kevin Systrom (1:41:30.720)

If you hate like what a company is doing

Lex Fridman (1:41:32.580)

or it makes you not feel happy.

Kevin Systrom (1:41:36.340)

I don't know, like people are really angry

Lex Fridman (1:41:37.860)

about Comcast or whatever.

Lex Fridman (1:41:39.140)

Are they even called Comcast anymore?

Lex Fridman (1:41:40.660)

It's like Xfinity or something, right?

Kevin Systrom (1:41:42.100)

They had to rebrand.

Lex Fridman (1:41:43.300)

They became meta, right?

Lex Fridman (1:41:44.620)

And it's like, but my point is if it makes you angry.

Lex Fridman (1:41:48.580)

That's beautiful, yeah.

Lex Fridman (1:41:50.020)

But the thing is, this is me saying this.

Lex Fridman (1:41:54.940)

I think your thesis is very strong and correct,

Kevin Systrom (1:41:58.780)

has elements of correctness,

Lex Fridman (1:42:00.100)

but I still personally put some blame on individuals.

Kevin Systrom (1:42:05.540)

Of course.

Lex Fridman (1:42:06.380)

I think you said, Elon, looking up,

Kevin Systrom (1:42:10.360)

there's something about Childlike Wander to him,

Lex Fridman (1:42:13.500)

like to his personality, his character,

Kevin Systrom (1:42:16.400)

something about, I think more so than others

Lex Fridman (1:42:19.900)

where people can trust them.

Lex Fridman (1:42:21.260)

And there's, I don't know,

Lex Fridman (1:42:23.060)

Sondra Prachai is an example of somebody who's like,

Kevin Systrom (1:42:26.460)

there's some kind, it's hard to put into words,

Lex Fridman (1:42:29.660)

but there's something about the human being

Kevin Systrom (1:42:32.120)

where he's trustworthy.

Lex Fridman (1:42:34.300)

He's human in a way that connects to us.

Lex Fridman (1:42:38.260)

And the same with Sajid Nadella.

Lex Fridman (1:42:40.820)

I mean, some of these folks, something about us

Kevin Systrom (1:42:46.460)

is drawn to them, even when they're flawed.

Lex Fridman (1:42:48.500)

Even like, so your thesis really holds up for Steve Jobs

Kevin Systrom (1:42:53.100)

because I think people didn't like Steve Jobs,

Lex Fridman (1:42:55.080)

but he delivered products

Lex Fridman (1:42:57.460)

and then they fell in love every time.

Lex Fridman (1:43:01.060)

I guess you could say that the CEO,

Kevin Systrom (1:43:03.040)

the leader is also a product.

Lex Fridman (1:43:05.820)

And if they keep delivering a product that people like

Kevin Systrom (1:43:08.980)

by being in public and saying things that people like,

Lex Fridman (1:43:11.460)

that's also a way to make people happy.

Lex Fridman (1:43:14.660)

But from a social network perspective,

Lex Fridman (1:43:16.260)

it makes me wonder how difficult it is

Kevin Systrom (1:43:19.860)

to explain to people why certain things happen,

Lex Fridman (1:43:22.100)

like to explain machine learning,

Kevin Systrom (1:43:24.500)

to explain why certain,

Lex Fridman (1:43:30.460)

the woke mob effect happens

Kevin Systrom (1:43:32.480)

or the certain kinds of like bullying happens,

Lex Fridman (1:43:36.420)

which is like, it's human nature combined with algorithm.

Lex Fridman (1:43:40.780)

And it's very difficult to control for

Lex Fridman (1:43:42.560)

how the spread of quote unquote misinformation happens.

Kevin Systrom (1:43:45.620)

It's very difficult to control for that.

Lex Fridman (1:43:47.700)

And so you try to decelerate certain parts

Lex Fridman (1:43:50.500)

and you create more problems than you solve.

Lex Fridman (1:43:53.700)

And anything that looks at all like censorship

Kevin Systrom (1:43:55.820)

can create huge amounts of problems as a slippery slope.

Lex Fridman (1:43:58.820)

And then you have to inject humans

Kevin Systrom (1:44:01.100)

to oversee the machine learning algorithms.

Lex Fridman (1:44:03.180)

And anytime you inject humans into the system,

Kevin Systrom (1:44:05.540)

it's gonna create a huge number of problems.

Lex Fridman (1:44:07.220)

And I feel like it's up to the leader

Kevin Systrom (1:44:08.780)

to communicate that effectively, to be transparent.

Lex Fridman (1:44:12.780)

First of all, design products

Kevin Systrom (1:44:13.960)

that don't have those problems.

Lex Fridman (1:44:15.180)

And second of all, when they have those problems,

Kevin Systrom (1:44:17.500)

to be able to communicate with them.

Lex Fridman (1:44:19.060)

I guess that's all going to,

Kevin Systrom (1:44:21.100)

when you run a social network company, your job is hard.

Lex Fridman (1:44:25.060)

Yeah, I will say the one element that you haven't named

Kevin Systrom (1:44:28.780)

that I think you're getting at is just bedside manner,

Lex Fridman (1:44:31.660)

which Steve Jobs, I never worked for him.

Kevin Systrom (1:44:35.460)

I never met him in person.

Lex Fridman (1:44:38.580)

Had an uncanny ability in public to have bedside manner.

Kevin Systrom (1:44:43.620)

I mean, some of the best clips of Steve Jobs

Lex Fridman (1:44:46.500)

from like, I would say maybe the 80s

Kevin Systrom (1:44:49.500)

when he's on the stage and getting questions

Lex Fridman (1:44:51.980)

from the audience about life.

Lex Fridman (1:44:54.460)

And he'll take this question that is like,

Lex Fridman (1:44:56.600)

how are you gonna compete with blah?

Lex Fridman (1:44:58.420)

And it's super boring.

Lex Fridman (1:44:59.460)

And I don't even know the name of the company.

Lex Fridman (1:45:01.480)

And his answer is as if you just asked your grandfather,

Lex Fridman (1:45:05.360)

the meaning of life.

Lex Fridman (1:45:07.420)

And you sit there and you're just like, what?

Lex Fridman (1:45:10.220)

And there's that bedside manner.

Lex Fridman (1:45:13.220)

And if you lack that, or if that's just not intuitive to you,

Lex Fridman (1:45:16.980)

I think that it can be a lot harder

Kevin Systrom (1:45:20.260)

to gain the trust of people.

Lex Fridman (1:45:21.500)

And then add on top of that, missteps of companies.

Kevin Systrom (1:45:27.320)

I don't know if you have any friends from the past

Lex Fridman (1:45:29.060)

where maybe they crossed you once,

Kevin Systrom (1:45:31.360)

or maybe you get back together and you're friends again,

Lex Fridman (1:45:34.380)

but you just never really forget that thing.

Kevin Systrom (1:45:36.900)

It's human nature not to forget.

Lex Fridman (1:45:38.880)

I'm Russian, you crossed me once.

Kevin Systrom (1:45:41.780)

We solved the problem.

Lex Fridman (1:45:43.340)

So my point is, humans don't forget.

Lex Fridman (1:45:48.900)

And if there are times in the past

Lex Fridman (1:45:50.380)

where they feel like they don't trust the company

Kevin Systrom (1:45:52.180)

or the company hasn't had their back,

Lex Fridman (1:45:55.180)

that is really hard to earn back,

Kevin Systrom (1:45:57.060)

especially if you don't have that bedside manner.

Lex Fridman (1:46:00.000)

And again, I'm not attributing this specifically to Marc

Kevin Systrom (1:46:02.900)

because I think a lot of the companies have this issue

Lex Fridman (1:46:06.380)

where one, you have to be trustworthy as a company

Lex Fridman (1:46:10.620)

and live by it and live by those actions.

Lex Fridman (1:46:12.620)

And then two, I think you need to be able

Kevin Systrom (1:46:14.540)

to be really relatable in a way that's very difficult

Lex Fridman (1:46:18.100)

if you're worth what these people are.

Kevin Systrom (1:46:20.560)

It's really hard.

Lex Fridman (1:46:22.060)

Yeah.

Kevin Systrom (1:46:23.220)

Jack does a pretty good job of this by being a monk.

Lex Fridman (1:46:27.020)

But I also, Jack issues attention.

Kevin Systrom (1:46:29.740)

He's not out there almost on purpose.

Lex Fridman (1:46:32.100)

He's just working hard, doing square, right?

Kevin Systrom (1:46:35.300)

I literally shared a desk like this with him at Odeo.

Lex Fridman (1:46:38.140)

I mean, this normal guy who likes painting,

Kevin Systrom (1:46:41.220)

I remember he would leave early on Wednesdays or something

Lex Fridman (1:46:45.000)

to go to a painting class.

Lex Fridman (1:46:47.340)

And he's creative, he's thoughtful.

Lex Fridman (1:46:49.380)

I mean, money makes people more creative and more thoughtful,

Lex Fridman (1:46:53.340)

extreme versions of themselves, right?

Lex Fridman (1:46:55.980)

And this was a long, long time ago.

Kevin Systrom (1:46:58.300)

You mentioned that he asked you

Lex Fridman (1:47:00.540)

to do some kind of JavaScript thing.

Kevin Systrom (1:47:02.740)

We were working on some JavaScript together.

Lex Fridman (1:47:05.220)

That's hilarious, like pre Twitter, early Twitter days,

Kevin Systrom (1:47:08.020)

you and Jack Dorsey are in a room together

Lex Fridman (1:47:11.220)

talking about JavaScript,

Kevin Systrom (1:47:12.780)

solving some kind of menial problem.

Lex Fridman (1:47:14.780)

Terrible problems, yeah.

Kevin Systrom (1:47:15.860)

I mean, not terrible, just like boring widget problem.

Lex Fridman (1:47:18.420)

I think it was the Odeo widget

Kevin Systrom (1:47:19.660)

we were working on at the time.

Lex Fridman (1:47:21.660)

I'm surprised anyone paid me to be in the room as an intern

Kevin Systrom (1:47:24.800)

because I didn't really provide any value.

Lex Fridman (1:47:26.500)

I'm very thankful to anyone who included me back in the day.

Kevin Systrom (1:47:30.440)

It was very helpful.

Lex Fridman (1:47:32.220)

So thank you for listening.

Lex Fridman (1:47:33.780)

I mean, is there Odeo that's a precursor to Twitter?

Lex Fridman (1:47:38.820)

First of all, did you have any anticipation

Kevin Systrom (1:47:40.700)

that this Jack Dorsey guy could be also

Lex Fridman (1:47:43.580)

a head of a major social network?

Lex Fridman (1:47:46.180)

And second, did you learn anything from the guy that,

Lex Fridman (1:47:49.420)

like, do you think it's a coincidence

Lex Fridman (1:47:52.460)

that you two were in the room together?

Lex Fridman (1:47:55.220)

And it's the coincidence meaning like,

Lex Fridman (1:47:58.900)

why does the world play its game in a certain way

Lex Fridman (1:48:01.700)

where these two founders of social networks?

Kevin Systrom (1:48:04.100)

I don't know.

Lex Fridman (1:48:04.940)

It's so weird, right?

Kevin Systrom (1:48:05.820)

Like, I mean, it's also weird that Mark showed up

Lex Fridman (1:48:11.900)

in our fraternity my sophomore year

Lex Fridman (1:48:13.940)

and we got to know each other then,

Lex Fridman (1:48:16.340)

like long before Instagram.

Kevin Systrom (1:48:18.060)

It's a small world,

Lex Fridman (1:48:20.620)

but let me tell a fun story about Jack.

Kevin Systrom (1:48:25.100)

We're at Odeo and I don't know,

Lex Fridman (1:48:27.260)

I think Ev was feeling like people

Kevin Systrom (1:48:28.900)

weren't working hard enough or something.

Lex Fridman (1:48:31.340)

Nice.

Lex Fridman (1:48:32.180)

And I can't remember exactly what he,

Lex Fridman (1:48:35.100)

he created this thing where every Friday,

Kevin Systrom (1:48:39.180)

I don't know if it was every Friday,

Lex Fridman (1:48:40.380)

I only remember this happening once,

Lex Fridman (1:48:43.140)

but he had us like a statuette, it's like of Mary.

Lex Fridman (1:48:47.540)

And in the bottom, it's hollow, right?

Lex Fridman (1:48:50.740)

And I remember on a Friday,

Lex Fridman (1:48:54.260)

Ev decided he was going to let everyone vote for

Kevin Systrom (1:48:57.300)

who had worked the hardest that week.

Lex Fridman (1:48:59.600)

We all voted, closed ballot, right?

Kevin Systrom (1:49:01.380)

We all put it in a bucket and he tallied the votes.

Lex Fridman (1:49:04.420)

And then whoever got the most votes, as I recall,

Kevin Systrom (1:49:07.820)

got the statuette.

Lex Fridman (1:49:09.860)

And in the statuette was a thousand bucks,

Kevin Systrom (1:49:12.280)

or I recall there was a thousand bucks in it.

Lex Fridman (1:49:14.380)

It might've been a hundred bucks,

Lex Fridman (1:49:15.280)

but let's call it a thousand.

Lex Fridman (1:49:16.220)

It's more exciting that way.

Kevin Systrom (1:49:17.060)

It felt like a thousand, yeah.

Lex Fridman (1:49:18.020)

It did to me for sure.

Kevin Systrom (1:49:19.420)

I actually got two votes.

Lex Fridman (1:49:20.560)

I was very happy.

Kevin Systrom (1:49:21.400)

We were a small company, but as the intern,

Lex Fridman (1:49:23.180)

I got at least two votes.

Lex Fridman (1:49:24.020)

So everybody knew how many votes they got individually?

Lex Fridman (1:49:26.260)

Yeah, yeah.

Lex Fridman (1:49:27.100)

And I think it was one of these self accountability things.

Lex Fridman (1:49:29.060)

Anyway, I remember Jack just getting

Kevin Systrom (1:49:31.420)

like the vast majority of votes from everyone.

Lex Fridman (1:49:35.100)

And I remember just thinking like,

Kevin Systrom (1:49:37.100)

like I couldn't imagine he would become what he'd become

Lex Fridman (1:49:39.740)

and do what he would do,

Lex Fridman (1:49:41.140)

but I had a profound respect that the new guy

Lex Fridman (1:49:45.640)

who I really liked worked that hard.

Lex Fridman (1:49:48.820)

And you could see his dedication even then

Lex Fridman (1:49:51.720)

and that people respected him.

Kevin Systrom (1:49:53.860)

That's the one story that I remember of him

Lex Fridman (1:49:55.660)

like working with him specifically from that summer.

Lex Fridman (1:49:58.000)

Can I take a small tangent on that?

Lex Fridman (1:49:59.500)

Of course.

Kevin Systrom (1:50:00.340)

There's kind of a pushback in Silicon Valley

Lex Fridman (1:50:02.020)

a little bit against hard work.

Lex Fridman (1:50:04.380)

Can you speak to the sort of the thing you admire

Lex Fridman (1:50:08.380)

to see the new guy working so hard, that thing?

Lex Fridman (1:50:11.420)

What is the value of that thing in a company?

Lex Fridman (1:50:13.740)

See, this is like, just to be very frank,

Kevin Systrom (1:50:16.420)

it drives me nuts.

Lex Fridman (1:50:17.720)

Like I saw this really funny video on TikTok.

Lex Fridman (1:50:21.860)

Was it on TikTok?

Lex Fridman (1:50:22.680)

It was like, I'm taking a break from my mental health

Kevin Systrom (1:50:24.940)

to work on my career.

Lex Fridman (1:50:26.300)

I thought that was funny.

Kevin Systrom (1:50:27.460)

Um, so I was like, oh, it is kind of phrased that way,

Lex Fridman (1:50:31.700)

the opposite often, right?

Kevin Systrom (1:50:32.980)

Yeah.

Lex Fridman (1:50:33.820)

Okay, so a couple of things.

Kevin Systrom (1:50:35.180)

Uh, I, uh, I have worked so hard

Lex Fridman (1:50:42.380)

to do the things that I did.

Kevin Systrom (1:50:43.680)

Like Mike and I lost years off of our lives,

Lex Fridman (1:50:47.580)

staying up late, figuring things out,

Kevin Systrom (1:50:50.140)

the stress that comes with the job.

Lex Fridman (1:50:51.700)

I have a lot more gray hair now than I did back then.

Kevin Systrom (1:50:54.800)

It requires an enormous amount of work

Lex Fridman (1:50:57.100)

and most people aren't successful, right?

Lex Fridman (1:50:59.460)

But even the ones that do don't skate by.

Lex Fridman (1:51:04.300)

I am okay if people choose not to work hard

Kevin Systrom (1:51:07.220)

because I don't actually think there's anything

Lex Fridman (1:51:09.020)

in this world that says you have to work hard.

Lex Fridman (1:51:13.140)

But I do think that great things require a lot of hard work.

Lex Fridman (1:51:16.420)

So there's no way you can expect to change the world

Kevin Systrom (1:51:18.640)

without working really hard.

Lex Fridman (1:51:19.700)

By the way, even changing the world,

Kevin Systrom (1:51:22.560)

you know, the folks that I respect the most

Lex Fridman (1:51:24.260)

have nudged the world in like a slight direction,

Kevin Systrom (1:51:27.220)

slight, very, very slight.

Lex Fridman (1:51:30.840)

Like even if Elon accomplishes all the things

Kevin Systrom (1:51:33.940)

he wants to accomplish,

Lex Fridman (1:51:34.920)

we will have nudged the world in a slight direction,

Lex Fridman (1:51:38.420)

but it requires enormous amount.

Lex Fridman (1:51:40.120)

There was an interview with him where he was just like,

Kevin Systrom (1:51:43.140)

he was interviewed, I think, at the Tesla factory

Lex Fridman (1:51:45.020)

and he was like, work is really hard.

Kevin Systrom (1:51:47.500)

This is actually unhealthy.

Lex Fridman (1:51:49.380)

And I can't recall the exact,

Lex Fridman (1:51:51.220)

but he was like visibly shaken

Lex Fridman (1:51:52.860)

about how hard he had been working.

Lex Fridman (1:51:54.220)

And he was like, this is bad.

Lex Fridman (1:51:55.860)

And unfortunately, I think to have great outcomes,

Kevin Systrom (1:51:57.620)

you actually do need to work

Lex Fridman (1:51:58.900)

at like three standard deviations above the mean,

Lex Fridman (1:52:01.840)

but there's nothing saying that people have to go for that.

Lex Fridman (1:52:03.980)

See, the thing is, but what I would argue,

Kevin Systrom (1:52:06.420)

this is my personal opinion,

Lex Fridman (1:52:08.080)

is nobody has to do anything, first of all.

Kevin Systrom (1:52:10.460)

They certainly don't have to work hard.

Lex Fridman (1:52:12.340)

But I think hard work in a company should be admired.

Lex Fridman (1:52:17.340)

And you should not feel like,

Lex Fridman (1:52:22.340)

you shouldn't feel good about yourself for not working hard.

Kevin Systrom (1:52:26.340)

Like, so for example, I don't have to work out.

Lex Fridman (1:52:30.140)

I don't have to run.

Kevin Systrom (1:52:31.860)

I hate running, but like, I certainly don't feel good

Lex Fridman (1:52:35.580)

if I don't run because I know for my health,

Kevin Systrom (1:52:37.540)

like there's certain values,

Lex Fridman (1:52:39.020)

I guess is what I'm trying to get at.

Kevin Systrom (1:52:40.340)

There's certain values that you have in life.

Lex Fridman (1:52:42.180)

It feels like there's certain values

Kevin Systrom (1:52:43.580)

that companies should have in hard work.

Lex Fridman (1:52:45.380)

Certain values that companies should have in hard work

Kevin Systrom (1:52:47.700)

is one of the things I think that should be admired.

Lex Fridman (1:52:51.460)

I often ask this kind of silly question,

Kevin Systrom (1:52:54.540)

just to get a sense of people,

Lex Fridman (1:52:56.560)

like if I'm hiring and so on.

Kevin Systrom (1:52:58.580)

I just ask if they think it's better

Lex Fridman (1:53:00.860)

to work hard or work smart.

Kevin Systrom (1:53:03.140)

It was helpful for me to get a sense of people from that.

Lex Fridman (1:53:07.420)

Because you think like the right.

Kevin Systrom (1:53:08.260)

The answer's both.

Lex Fridman (1:53:09.740)

What's that?

Kevin Systrom (1:53:10.560)

The answer's both.

Lex Fridman (1:53:11.400)

The answer's both.

Kevin Systrom (1:53:12.240)

I usually try not to give them that,

Lex Fridman (1:53:13.900)

but sometimes I'll say both if that's an option.

Lex Fridman (1:53:16.900)

But a lot of people kind of,

Lex Fridman (1:53:19.780)

a surprising number will say work smart.

Lex Fridman (1:53:22.060)

And there are usually people

Lex Fridman (1:53:23.380)

who don't know how to work smart.

Lex Fridman (1:53:26.700)

And they're literally just lazy.

Lex Fridman (1:53:29.580)

Not just, there's two effects behind that.

Kevin Systrom (1:53:32.260)

One is laziness and the other is ego.

Lex Fridman (1:53:37.300)

When you're younger and you say it's better to work smart,

Kevin Systrom (1:53:40.420)

it means you think you know what it means

Lex Fridman (1:53:44.220)

to work smart at this early stage.

Kevin Systrom (1:53:46.820)

To me, people that say work hard or both,

Lex Fridman (1:53:49.940)

they have the humility to understand like,

Kevin Systrom (1:53:52.360)

I'm going to have to work my ass off

Lex Fridman (1:53:54.340)

because I'm too dumb to know how to work smart.

Lex Fridman (1:53:56.820)

And people who are self critical in this way,

Lex Fridman (1:53:59.140)

in some small amount, you have to have some confidence.

Lex Fridman (1:54:01.480)

But if you have humility,

Lex Fridman (1:54:03.860)

that means you're going to actually eventually figure out

Lex Fridman (1:54:06.420)

what it means to work smart.

Lex Fridman (1:54:07.500)

And then to actually be successful, you should do both.

Lex Fridman (1:54:11.220)

So I have a very particular take on this,

Lex Fridman (1:54:14.060)

which is that no one's forcing you to do anything.

Kevin Systrom (1:54:19.820)

All choices have consequences.

Lex Fridman (1:54:22.260)

So if you major in, I don't know, theoretical literature,

Kevin Systrom (1:54:27.540)

I don't even know if that's a major.

Lex Fridman (1:54:28.900)

I'm just making something up.

Kevin Systrom (1:54:30.020)

That's supposed to regular literature, applied literature.

Lex Fridman (1:54:33.580)

Yeah, think about like theoretical Spanish lit

Kevin Systrom (1:54:38.300)

from the 14th century.

Lex Fridman (1:54:39.420)

Like just make up your esoteric thing.

Lex Fridman (1:54:41.740)

And then the number of people I went to Stanford with

Lex Fridman (1:54:44.300)

who get out in the world and they're like,

Lex Fridman (1:54:45.360)

wait, what, I can't find a job?

Lex Fridman (1:54:47.260)

Like no one wants a theoretical,

Kevin Systrom (1:54:49.020)

like there are plenty of counter examples

Lex Fridman (1:54:51.660)

of people who have majored in esoteric things

Lex Fridman (1:54:53.380)

and gone on to be very successful.

Lex Fridman (1:54:54.840)

So I just want to be clear, it's not about the major.

Lex Fridman (1:54:56.700)

But every choice you make, whether it's to have kids,

Lex Fridman (1:55:00.820)

like I love my children.

Kevin Systrom (1:55:02.900)

It's so awesome to have two kids.

Lex Fridman (1:55:04.660)

And it is so hard to work really hard and also have kids.

Kevin Systrom (1:55:08.060)

It's really hard.

Lex Fridman (1:55:09.420)

And there's a reason why certain very successful people

Kevin Systrom (1:55:12.640)

like don't have, or not successful,

Lex Fridman (1:55:14.560)

but people who run very, very large companies or startups

Kevin Systrom (1:55:17.460)

have chosen not to have kids for a while

Lex Fridman (1:55:19.180)

or chosen not to like prioritize them.

Kevin Systrom (1:55:21.620)

Everything's a choice.

Lex Fridman (1:55:22.620)

And like I choose to prioritize my children

Lex Fridman (1:55:24.900)

because like I want to do that, right?

Lex Fridman (1:55:27.780)

So everything's a choice.

Kevin Systrom (1:55:29.420)

Now, once you've made that choice,

Lex Fridman (1:55:33.120)

I think it's important that the contract is clear,

Kevin Systrom (1:55:36.160)

which is to say,

Lex Fridman (1:55:37.000)

let's imagine you were joining a new startup.

Kevin Systrom (1:55:40.760)

It's important that that startup communicate

Lex Fridman (1:55:43.520)

that like the expectation is like,

Kevin Systrom (1:55:44.720)

we're all working really, really hard right now.

Lex Fridman (1:55:46.440)

You don't have to join the startup.

Lex Fridman (1:55:48.760)

But like, if you do just know like you're,

Lex Fridman (1:55:51.480)

it's almost as if you join, I don't know,

Kevin Systrom (1:55:53.720)

pick your like sports team.

Lex Fridman (1:55:57.960)

Like let's go back to the Yankees for a second.

Kevin Systrom (1:56:00.840)

You want to join the Yankees,

Lex Fridman (1:56:01.880)

but you don't really want to work that hard.

Kevin Systrom (1:56:03.640)

You don't really want to do batting practice

Lex Fridman (1:56:05.780)

or pitching practice or whatever for your position, right?

Kevin Systrom (1:56:09.840)

That to me is wacko.

Lex Fridman (1:56:11.400)

And that's actually the world that it feels like we live in

Kevin Systrom (1:56:13.920)

in tech sometimes,

Lex Fridman (1:56:15.240)

where people both want to work for the Yankees

Kevin Systrom (1:56:17.080)

because it pays a lot,

Lex Fridman (1:56:18.480)

but like don't actually want to work that hard.

Kevin Systrom (1:56:21.100)

That I don't fully understand.

Lex Fridman (1:56:23.300)

Because if you sign up for some of these things,

Kevin Systrom (1:56:26.000)

just sign up for it.

Lex Fridman (1:56:26.840)

But it's okay if you don't want to sign up for it.

Kevin Systrom (1:56:29.020)

There's so many wonderful careers in this world

Lex Fridman (1:56:31.720)

that don't require 80 hours a week.

Lex Fridman (1:56:33.780)

But when I read about companies going to like

Lex Fridman (1:56:35.280)

four day work weeks and stuff,

Kevin Systrom (1:56:36.520)

I just like, I chuckle because I can't get enough done

Lex Fridman (1:56:39.800)

with a seven day week.

Kevin Systrom (1:56:41.080)

I don't know how.

Lex Fridman (1:56:42.820)

And people will say, oh, you're just not working smart.

Lex Fridman (1:56:44.920)

And it's like, no, I work pretty smart,

Lex Fridman (1:56:46.600)

I think in general.

Kevin Systrom (1:56:47.560)

Like I wouldn't have gotten to this point

Lex Fridman (1:56:49.340)

if I hadn't like some amount of working smart.

Lex Fridman (1:56:52.760)

And there is balance though.

Lex Fridman (1:56:53.840)

So I used to be like a pretty big cyclist.

Kevin Systrom (1:56:55.880)

I don't do it much anymore just because of kids

Lex Fridman (1:56:58.240)

and like prioritizing other things, right?

Lex Fridman (1:57:01.760)

But one of the most important things to learn

Lex Fridman (1:57:03.380)

as a cyclist is to take a rest day.

Lex Fridman (1:57:06.140)

But to me and to cyclists,

Lex Fridman (1:57:07.640)

like resting is a function of optimizing for the long run.

Kevin Systrom (1:57:12.680)

It's not like a thing that you do for its own merits.

Lex Fridman (1:57:16.240)

It's actually like, if you don't rest,

Kevin Systrom (1:57:17.880)

your muscles don't recover.

Lex Fridman (1:57:18.880)

And then you're just not as,

Kevin Systrom (1:57:19.880)

like you're not training as efficiently.

Lex Fridman (1:57:21.520)

You should probably, the successful people I've known

Kevin Systrom (1:57:24.680)

in terms of athletes, they hate rest days,

Lex Fridman (1:57:27.160)

but they know they have to do it for the long term.

Kevin Systrom (1:57:29.800)

They think their opposition is getting stronger and stronger

Lex Fridman (1:57:32.560)

and that's the feeling,

Lex Fridman (1:57:34.360)

but you know it's the right thing

Lex Fridman (1:57:36.000)

and usually you need a coach to help you.

Kevin Systrom (1:57:38.200)

Yeah, totally.

Lex Fridman (1:57:39.040)

So, I mean, I use this thing called training peaks

Lex Fridman (1:57:41.240)

and it's interesting

Lex Fridman (1:57:42.080)

because it actually mathematically shows

Kevin Systrom (1:57:44.060)

like where you are on the curve and all this stuff,

Lex Fridman (1:57:46.800)

but you have to have that rest,

Lex Fridman (1:57:49.400)

but it's a function of going harder for longer.

Lex Fridman (1:57:52.320)

Again, it's this reinforcement learning,

Kevin Systrom (1:57:53.840)

like planning the aggregate and the long,

Lex Fridman (1:57:56.480)

but a lot of people will hide behind laziness

Kevin Systrom (1:57:58.600)

by saying that they're trying to optimize for the long run

Lex Fridman (1:58:00.880)

and they're not, they're just not working very hard.

Lex Fridman (1:58:03.200)

But again, you don't have to sign up for it.

Lex Fridman (1:58:05.040)

It's totally cool.

Kevin Systrom (1:58:05.880)

Like, I don't think less of people

Lex Fridman (1:58:07.420)

for like not working super hard.

Kevin Systrom (1:58:08.960)

It's just like, don't sign up for things

Lex Fridman (1:58:10.280)

that require working super hard.

Lex Fridman (1:58:11.880)

And some of that requires for the leadership

Lex Fridman (1:58:13.920)

to have the guts, the boldness to communicate effectively

Kevin Systrom (1:58:17.080)

at the very beginning.

Lex Fridman (1:58:17.920)

I mean, sometimes I think most of the problems arise

Kevin Systrom (1:58:20.760)

in the fact that the leadership is kind of hesitant

Lex Fridman (1:58:24.600)

to communicate the socially difficult truth

Kevin Systrom (1:58:30.960)

of what it takes to be at this company.

Lex Fridman (1:58:33.840)

So they kind of say, hey, come with us.

Kevin Systrom (1:58:36.520)

There's, we have snacks, you know, but.

Lex Fridman (1:58:39.480)

Unlimited vacation and.

Kevin Systrom (1:58:40.800)

Yeah.

Lex Fridman (1:58:41.640)

You know, Ray at Bridgewater is always fascinating

Kevin Systrom (1:58:44.240)

because, you know, people,

Lex Fridman (1:58:45.640)

it's been called like a cult on the outside

Kevin Systrom (1:58:47.920)

or cultish and,

Lex Fridman (1:58:49.860)

but what's fascinating is like,

Kevin Systrom (1:58:51.000)

they just don't give on their principles.

Lex Fridman (1:58:52.800)

They're like, listen, this is what it's like to work here.

Kevin Systrom (1:58:55.240)

We record every meeting.

Lex Fridman (1:58:57.320)

We're like brutally honest

Lex Fridman (1:58:59.280)

and that's not going to feel right to everyone.

Lex Fridman (1:59:00.920)

And if it doesn't feel right to you, totally cool.

Kevin Systrom (1:59:03.440)

Just go work somewhere else.

Lex Fridman (1:59:04.880)

But if you work here, you are signing up for this.

Lex Fridman (1:59:08.040)

And that's, that's been fascinating to me

Lex Fridman (1:59:10.400)

because it's honesty upfront.

Kevin Systrom (1:59:12.200)

It's a system in which you operate.

Lex Fridman (1:59:15.480)

And if it's not for you,

Lex Fridman (1:59:16.720)

like no one's forcing you to work there, right?

Lex Fridman (1:59:19.600)

I actually did.

Lex Fridman (1:59:20.440)

So I did a conversation with him

Lex Fridman (1:59:22.280)

and kind of got stuck in a funny moment,

Kevin Systrom (1:59:25.140)

which is at the end I asked him, you know,

Lex Fridman (1:59:28.360)

to give me honest feedback of how I did on the interview.

Lex Fridman (1:59:31.420)

And I was.

Lex Fridman (1:59:32.260)

Did he?

Kevin Systrom (1:59:33.440)

I don't think so.

Lex Fridman (1:59:34.760)

He was super nice.

Kevin Systrom (1:59:36.200)

He asked me, he's like, well,

Lex Fridman (1:59:38.760)

tell me, did you accomplish

Lex Fridman (1:59:40.120)

what you were hoping to accomplish?

Lex Fridman (1:59:42.040)

I was like, that's not, that's not.

Kevin Systrom (1:59:45.080)

I'm asking you as an objective observer

Lex Fridman (1:59:47.880)

of two people talking, how do we do today?

Lex Fridman (1:59:52.000)

And then he's like, well,

Lex Fridman (1:59:53.760)

he gave me this politician answer.

Kevin Systrom (1:59:55.240)

Well, I feel like we've accomplished

Lex Fridman (1:59:57.200)

successful communication of like ideas,

Lex Fridman (20:00.580)

or are you just gonna have to try it out?

Lex Fridman (20:03.820)

Because to me, my intuition would be

Kevin Systrom (20:07.380)

those three seconds would kill the app.

Lex Fridman (20:09.940)

Like I would try to do the OpenGL thing.

Kevin Systrom (20:12.300)

Right, so I wish I were that smart at the time.

Lex Fridman (20:17.600)

I wasn't, I just knew how to do what I knew how to do.

Lex Fridman (20:21.680)

And I decided, okay, like,

Lex Fridman (20:23.580)

why don't I just iterate over the values and change them?

Lex Fridman (20:27.220)

And what's interesting is that

Lex Fridman (20:31.820)

compared to the alternatives, no one else used OpenGL.

Lex Fridman (20:35.340)

So everyone else was doing it the dumb way.

Lex Fridman (20:37.140)

And in fact, they were doing it at a high resolution.

Kevin Systrom (20:40.460)

Now comes in the small resolution

Lex Fridman (20:42.500)

that we'll talk about in a second.

Kevin Systrom (20:45.140)

By choosing 512 pixels by 512 pixels,

Lex Fridman (20:48.620)

which I believe it was at the time,

Kevin Systrom (20:51.000)

we iterated over a lot fewer pixels than our competitors

Lex Fridman (20:54.340)

who were trying to do these enormous output like images.

Lex Fridman (20:58.180)

So instead of taking 20 seconds,

Lex Fridman (21:00.940)

I mean, three seconds feels pretty good, right?

Lex Fridman (21:03.320)

So on a relative basis, we were winning like a lot.

Lex Fridman (21:06.340)

Okay, so that's answer number one.

Kevin Systrom (21:08.520)

Answer number two is we actually focused

Lex Fridman (21:11.580)

on latency in the right places.

Lex Fridman (21:13.140)

So we did this really wonderful thing when you uploaded.

Lex Fridman (21:17.460)

So the way it would work is you'd take your phone,

Kevin Systrom (21:20.900)

you'd take the photo and then you'd go to the,

Lex Fridman (21:25.040)

you'd go to the edit screen where you would caption it.

Lex Fridman (21:28.820)

And on that caption screen, you'd start typing

Lex Fridman (21:31.180)

and you'd think, okay, like what's a clever caption?

Lex Fridman (21:34.180)

And I said to Mike, hey, when I worked on the Gmail team,

Lex Fridman (21:36.380)

you know what they did?

Kevin Systrom (21:37.440)

When you typed in your username or your email address,

Lex Fridman (21:40.820)

even before you've entered in your password,

Kevin Systrom (21:43.220)

like the probability once you enter in your username

Lex Fridman (21:47.100)

that you're going to actually sign in is extremely high.

Lex Fridman (21:51.100)

So why not just start loading your account

Lex Fridman (21:52.620)

in the background?

Kevin Systrom (21:53.460)

Not like sending it down to the desktop,

Lex Fridman (21:55.700)

that would be a security issue,

Lex Fridman (21:59.380)

but like loaded into memory on the server,

Lex Fridman (22:01.340)

like get it ready, prepare it.

Kevin Systrom (22:03.600)

I always thought that was so fascinating and unintuitive.

Lex Fridman (22:06.980)

And I was like, Mike, why don't we just do that?

Lex Fridman (22:08.900)

But like, we'll just upload the photo

Lex Fridman (22:10.440)

and like assume you're going to upload the photo.

Lex Fridman (22:14.140)

And if you don't, forget about it, we'll delete it, right?

Lex Fridman (22:17.620)

So what ended up happening was people would caption

Kevin Systrom (22:20.340)

their photo, they'd press done or upload

Lex Fridman (22:24.660)

and you'd see this little progress bar just go foop.

Lex Fridman (22:27.940)

It was lightning fast, okay?

Lex Fridman (22:30.180)

We were no faster than anyone else at the time,

Lex Fridman (22:32.960)

but by choosing 512 by 512 and doing it in the background,

Lex Fridman (22:37.420)

it almost guaranteed that it was done

Kevin Systrom (22:39.680)

by the time you captioned.

Lex Fridman (22:41.340)

And everyone when they used it was like,

Lex Fridman (22:43.980)

how the hell is this thing so fast?

Lex Fridman (22:47.000)

But we were slow, we just hid the slowness.

Kevin Systrom (22:49.340)

It wasn't like, these things are just like,

Lex Fridman (22:51.460)

it's a shell game, you're just hiding the latency.

Kevin Systrom (22:55.180)

That mattered to people like a lot.

Lex Fridman (22:58.720)

And I think that, so you were willing to put up

Kevin Systrom (23:01.020)

with a slow filter if it meant

Lex Fridman (23:03.500)

you could share it immediately.

Lex Fridman (23:04.700)

And of course we added sharing options

Lex Fridman (23:06.580)

which let you distribute it really quickly,

Kevin Systrom (23:08.220)

that was the third part.

Lex Fridman (23:11.400)

So latency matters, but relative to what?

Lex Fridman (23:14.520)

And then there's some like tricks,

Lex Fridman (23:17.500)

you get around to just hiding the latency.

Kevin Systrom (23:20.580)

Like I don't know if Spotify starts downloading

Lex Fridman (23:23.020)

the next song eagerly, I'm assuming they do.

Kevin Systrom (23:25.740)

There are a bunch of ideas here that are not rocket science

Lex Fridman (23:29.300)

that really help.

Lex Fridman (23:31.740)

And all of that was stuff you were explicitly

Lex Fridman (23:35.100)

having a discussion about, like those designs

Lex Fridman (23:37.820)

and you were having like arguments, discussions.

Lex Fridman (23:41.660)

I'm not sure it was arguments, I mean,

Kevin Systrom (23:43.540)

I'm not sure if you've met my co founder Mike,

Lex Fridman (23:45.480)

but he's a pretty nice guy and he's very reasonable.

Lex Fridman (23:48.460)

And we both just saw eye to eye and we're like,

Lex Fridman (23:51.220)

yeah, it's like, make this fast or at least seem fast,

Kevin Systrom (23:56.500)

it'll be great.

Lex Fridman (23:57.340)

Honestly, I think the most contentious thing

Lex Fridman (23:59.460)

and he would say this too initially,

Lex Fridman (24:02.740)

was I was on an iPhone 3G, so like the not so fast one.

Lex Fridman (24:08.300)

And he had a brand new iPhone 4, that was cheap.

Lex Fridman (24:10.860)

Nice.

Lex Fridman (24:12.500)

And his feed loaded super smoothly,

Lex Fridman (24:15.140)

like when he would scroll from photo to photo,

Lex Fridman (24:18.760)

buttery smooth, right?

Lex Fridman (24:20.380)

But on my phone, every time you got to a new photo,

Kevin Systrom (24:22.780)

it was like, kachunk, kachunk, allocate memory,

Lex Fridman (24:25.420)

like all this stuff, right?

Kevin Systrom (24:27.460)

I was like, Mike, that's unacceptable.

Lex Fridman (24:29.300)

He's like, oh, come on, man, just like upgrade your phone.

Kevin Systrom (24:31.860)

Basically, he didn't actually say that,

Lex Fridman (24:33.260)

he was nicer than that.

Lex Fridman (24:35.780)

But I could tell he wished like,

Lex Fridman (24:36.900)

I would just stop being cheap and just get a new phone.

Lex Fridman (24:39.680)

But what's funny is we actually sat there working

Lex Fridman (24:41.540)

on that little detail for a few days before launch.

Lex Fridman (24:45.420)

And that polished experience,

Lex Fridman (24:48.620)

plus the fact that uploading seemed fast

Kevin Systrom (24:51.060)

for all these people who didn't have nice phones,

Lex Fridman (24:54.380)

I think meant a lot because far too often,

Kevin Systrom (24:57.820)

you see teams focus not on performance,

Lex Fridman (25:00.560)

they focus on what's the cool computer science problem

Lex Fridman (25:03.980)

they can solve, right?

Lex Fridman (25:05.940)

Can we scale this thing to a billion users

Lex Fridman (25:08.500)

and they've got like a hundred, right?

Lex Fridman (25:11.040)

Yeah.

Kevin Systrom (25:12.780)

You talked about loss function,

Lex Fridman (25:14.020)

so I want to come back to that.

Kevin Systrom (25:16.060)

Like the loss function is like,

Lex Fridman (25:17.780)

do you provide a great, happy, magical,

Lex Fridman (25:20.700)

whatever experience for the consumer?

Lex Fridman (25:23.160)

And listen, if it happens to involve something complex

Lex Fridman (25:25.980)

and technical, then great.

Lex Fridman (25:27.760)

But it turns out, I think most of the time,

Kevin Systrom (25:32.260)

those experiences are just sitting there waiting to be built

Lex Fridman (25:34.820)

with like not that complex solutions.

Lex Fridman (25:38.620)

But everyone is just like so stuck in their own head

Lex Fridman (25:40.860)

that they have to overengineer everything

Lex Fridman (25:42.660)

and then they forget about the easy stuff.

Lex Fridman (25:45.020)

I mean, also, maybe to flip the loss function there is,

Kevin Systrom (25:48.460)

you're trying to minimize the number of times

Lex Fridman (25:50.780)

there's unpleasant experience, right?

Kevin Systrom (25:54.040)

Like the one you mentioned where when you go

Lex Fridman (25:56.100)

to the next photo, it freezes for a little bit.

Lex Fridman (25:58.460)

So it's almost, as opposed to maximizing pleasure,

Lex Fridman (26:00.960)

it's probably easier to minimize the number of like,

Kevin Systrom (26:04.720)

the friction.

Lex Fridman (26:05.860)

Yeah.

Lex Fridman (26:06.700)

And as we all know, you just make the pleasure negative

Lex Fridman (26:10.420)

and then minimize everything, so.

Kevin Systrom (26:13.060)

We're mapping this all back to neural networks.

Lex Fridman (26:14.980)

But actually, can I say one thing on that,

Kevin Systrom (26:16.560)

which is I don't know a lot about machine learning,

Lex Fridman (26:19.760)

but I feel like I've tried studying a bunch.

Kevin Systrom (26:23.060)

That whole idea of reinforcement learning

Lex Fridman (26:26.880)

and planning out more than the greedy single experience,

Kevin Systrom (26:30.620)

I think is the closest you can get

Lex Fridman (26:34.060)

to like ideal product design thinking,

Kevin Systrom (26:38.580)

where you're not saying,

Lex Fridman (26:39.980)

hey, like, can we have a great experience just this one time?

Lex Fridman (26:43.740)

But like, what is the right way to onboard someone?

Lex Fridman (26:46.340)

What series of experiences correlate most with them

Lex Fridman (26:50.060)

hanging on long term, right?

Lex Fridman (26:52.660)

So not just saying, oh, did the photo load slowly

Kevin Systrom (26:55.620)

a couple of times, or did they get a great photo

Lex Fridman (26:57.940)

at the top of their feed?

Lex Fridman (27:00.060)

But like, what are the things that are gonna make

Lex Fridman (27:01.860)

this person come back over the next week,

Lex Fridman (27:04.940)

over the next month?

Lex Fridman (27:06.380)

And as a product designer asking yourself,

Kevin Systrom (27:08.460)

okay, I wanna optimize, not just minimize bad experiences

Lex Fridman (27:11.720)

in the short run, but like,

Lex Fridman (27:13.060)

how do I get someone to engage over the next month?

Lex Fridman (27:17.900)

And I'm not gonna claim at all that I thought that way

Kevin Systrom (27:20.780)

at all at the time, because I certainly didn't.

Lex Fridman (27:23.120)

But if I were going back and giving myself any advice,

Kevin Systrom (27:25.540)

it would be thinking, what are those second order effects

Lex Fridman (27:28.700)

that you can create?

Lex Fridman (27:30.380)

And it turns out having your friends on the service

Lex Fridman (27:33.380)

is an enormous win.

Lex Fridman (27:34.860)

So starting with a very small group of people

Lex Fridman (27:37.580)

that produce content that you wanted to see, which we did,

Kevin Systrom (27:40.520)

we seeded the community very well, I think.

Lex Fridman (27:43.860)

Ended up mattering, and so.

Kevin Systrom (27:46.620)

Yeah, you said that community is one

Lex Fridman (27:48.200)

of the most important things.

Lex Fridman (27:49.660)

So it's from a metrics perspective,

Lex Fridman (27:52.020)

from maybe a philosophy perspective,

Kevin Systrom (27:55.220)

building a certain kind of community within the app.

Lex Fridman (27:57.900)

See, I wasn't sure what exactly you meant by that

Kevin Systrom (28:00.580)

when I first heard you say that.

Lex Fridman (28:01.940)

Maybe you can elaborate, but as I understand now,

Kevin Systrom (28:04.420)

it can literally mean get your friends onto the app.

Lex Fridman (28:09.420)

Yeah, think of it this way.

Kevin Systrom (28:12.900)

You can build an amazing restaurant or bar or whatever,

Lex Fridman (28:16.700)

right, but if you show up and you're the only one there,

Lex Fridman (28:20.300)

is it like, does it matter how good the food is?

Lex Fridman (28:24.020)

The drinks, whatever?

Kevin Systrom (28:25.620)

No, these are inherently social experiences

Lex Fridman (28:28.820)

that we were working on.

Lex Fridman (28:30.460)

So the idea of having people there,

Lex Fridman (28:35.580)

like you needed to have that,

Kevin Systrom (28:36.900)

otherwise it was just a filter out.

Lex Fridman (28:39.220)

But by the way, part of the genius,

Kevin Systrom (28:41.380)

I'm gonna say genius, even though it wasn't really genius,

Lex Fridman (28:43.900)

was starting to be marauding as a filter app was awesome.

Kevin Systrom (28:50.300)

The fact that you could,

Lex Fridman (28:51.220)

so we talk about single player mode a lot,

Lex Fridman (28:53.020)

which is like, can you play the game alone?

Lex Fridman (28:56.280)

And Instagram, you could totally play alone.

Kevin Systrom (28:57.980)

You could filter your photos,

Lex Fridman (28:59.380)

and a lot of people would tell me,

Kevin Systrom (29:00.900)

I didn't even realize that this thing was a social network

Lex Fridman (29:04.580)

until my friend showed up.

Kevin Systrom (29:06.140)

It totally worked as a single player game.

Lex Fridman (29:09.580)

And then when your friends showed up,

Kevin Systrom (29:10.940)

all of a sudden it was like,

Lex Fridman (29:11.780)

oh, not only was this great alone,

Lex Fridman (29:14.100)

but now I actually have this trove of photos

Lex Fridman (29:16.940)

that people can look at and start liking,

Lex Fridman (29:19.380)

and then I can like theirs.

Lex Fridman (29:20.500)

And so it was this bootstrap method

Kevin Systrom (29:23.180)

of how do you make the thing not suck

Lex Fridman (29:25.180)

when the restaurant is empty?

Kevin Systrom (29:27.040)

Yeah, but the thing is, when you say friends,

Lex Fridman (29:29.660)

we're not necessarily referring to friends

Kevin Systrom (29:31.700)

in the physical space.

Lex Fridman (29:32.780)

So you're not bringing your physical friends with you.

Kevin Systrom (29:35.340)

You're also making new friends.

Lex Fridman (29:36.740)

So you're finding new community.

Lex Fridman (29:38.380)

So it's not immediately obvious to me

Lex Fridman (29:40.140)

that it's almost like building any kind of community.

Kevin Systrom (29:45.020)

It was both.

Lex Fridman (29:46.540)

And what we learned very early on

Kevin Systrom (29:48.620)

was what made Instagram special

Lex Fridman (29:50.160)

and the reason why you would sign up for it

Kevin Systrom (29:51.860)

versus say, just sit on Facebook

Lex Fridman (29:53.500)

and look at your friends photos.

Kevin Systrom (29:55.820)

Of course we were live,

Lex Fridman (29:56.900)

and of course it was interesting

Kevin Systrom (29:58.220)

to see what your friends were doing now.

Lex Fridman (2:00:00.280)

which is I'd love to spread some of the ideas

Kevin Systrom (2:00:03.600)

in that, like in principles and so on.

Lex Fridman (2:00:06.880)

I was like.

Kevin Systrom (2:00:07.720)

Back to my original point,

Lex Fridman (2:00:08.760)

it's really hard to get.

Kevin Systrom (2:00:10.000)

Even for Radalia.

Lex Fridman (2:00:11.880)

It's really hard to give feedback.

Lex Fridman (2:00:13.880)

And one of the other things I learned from him

Lex Fridman (2:00:15.720)

and just people in that world is like,

Kevin Systrom (2:00:19.760)

man, humans really like to pretend

Lex Fridman (2:00:22.120)

like they've come to,

Kevin Systrom (2:00:24.700)

that they've come to some kind of meeting of the minds.

Lex Fridman (2:00:27.080)

Like if there's conflict, if you and I have conflict,

Lex Fridman (2:00:30.120)

it's always better to meet face to face, right?

Lex Fridman (2:00:32.620)

We're on the phone.

Lex Fridman (2:00:33.720)

Slack is not great, right?

Lex Fridman (2:00:35.240)

Email is not great, but face to face.

Kevin Systrom (2:00:37.160)

What's crazy is you and I get together

Lex Fridman (2:00:38.680)

and we actively try to,

Kevin Systrom (2:00:40.660)

even if we're not actually solving the conflict,

Lex Fridman (2:00:43.080)

we actively try to paper over the conflict.

Kevin Systrom (2:00:45.880)

Oh yeah, it didn't really bother me that much.

Lex Fridman (2:00:48.620)

Oh yeah, I'm sure you didn't mean it.

Lex Fridman (2:00:50.840)

But like, no, in our minds we're still there.

Lex Fridman (2:00:54.320)

So this is one of the things that as a leader,

Kevin Systrom (2:00:56.200)

you always have to be digging,

Lex Fridman (2:00:58.420)

especially as you ascend.

Kevin Systrom (2:00:59.260)

Like straight to the conflict.

Lex Fridman (2:01:00.880)

Yeah, but as you ascend,

Kevin Systrom (2:01:02.040)

no one wants to tell you you're crazy.

Lex Fridman (2:01:03.600)

No one wants to tell you your idea is bad.

Lex Fridman (2:01:06.720)

And you can, you're like, oh,

Lex Fridman (2:01:08.120)

oh, I'm going to be a leader.

Lex Fridman (2:01:09.160)

And the idea is, well, I'm just going to ask people.

Lex Fridman (2:01:12.280)

No one tells you.

Lex Fridman (2:01:13.600)

So like you have to look for the markers,

Lex Fridman (2:01:16.560)

knowing that literally just people

Kevin Systrom (2:01:18.240)

aren't going to tell you along the way and be paranoid.

Lex Fridman (2:01:21.280)

I mean, you asked about selling the company.

Kevin Systrom (2:01:24.000)

I think one of the biggest differences between me

Lex Fridman (2:01:25.960)

and a lot of other entrepreneurs is like,

Kevin Systrom (2:01:28.240)

I wasn't completely confident we could do it.

Lex Fridman (2:01:30.440)

Like we could be alone and actually be great.

Lex Fridman (2:01:35.920)

And if any entrepreneur is honest with you,

Lex Fridman (2:01:37.560)

they also feel that way.

Lex Fridman (2:01:39.120)

But a lot of people are like,

Lex Fridman (2:01:40.080)

well, I have to be cocky and just say,

Kevin Systrom (2:01:42.600)

I can do this on my own.

Lex Fridman (2:01:43.960)

We're going to be fine.

Kevin Systrom (2:01:44.800)

We're going to crush everyone.

Lex Fridman (2:01:46.760)

Some people do say that and then it's not right.

Lex Fridman (2:01:49.160)

And they, and they fail.

Lex Fridman (2:01:50.760)

But being honest in that moment with yourself,

Kevin Systrom (2:01:55.480)

with those close to you.

Lex Fridman (2:01:57.440)

And also you talked about the personality of leaders

Lex Fridman (2:02:00.840)

and who resonates and who doesn't.

Lex Fridman (2:02:04.320)

It's rare that I see leaders be vulnerable, rare.

Lex Fridman (2:02:09.320)

And one thing I tried to do at Instagram,

Lex Fridman (2:02:12.500)

at least internally was like, say when I screwed up

Lex Fridman (2:02:16.960)

and like point out how I was wrong about things

Lex Fridman (2:02:19.700)

and point out where my judgment was off.

Lex Fridman (2:02:22.700)

Everyone thinks they have to bat a thousand, right?

Lex Fridman (2:02:25.360)

Like that's crazy.

Kevin Systrom (2:02:27.060)

The best quant hedge funds in the world,

Lex Fridman (2:02:29.060)

bat 50.001%.

Lex Fridman (2:02:31.660)

They just take a lot of bets, right?

Lex Fridman (2:02:34.020)

Renaissance, they might, they might bat 51%, right?

Lex Fridman (2:02:37.800)

But holy hell, like the question isn't,

Lex Fridman (2:02:41.620)

are you right every single time

Lex Fridman (2:02:43.340)

and you have to seem invincible.

Lex Fridman (2:02:46.380)

The question is how many at bats do you get?

Lex Fridman (2:02:48.220)

And on average, are you better on average, right?

Lex Fridman (2:02:53.700)

With enough bets and enough at bats

Kevin Systrom (2:02:55.620)

that your aggregate can be very high.

Lex Fridman (2:02:58.700)

I mean, Steve Jobs was wrong at a lot of stuff.

Lex Fridman (2:03:01.980)

The Newton too early, right?

Lex Fridman (2:03:03.780)

Next, not quite right.

Kevin Systrom (2:03:05.280)

There was even a time where he said like,

Lex Fridman (2:03:08.200)

no one will ever wanna watch a video on the iPod.

Kevin Systrom (2:03:13.340)

Totally wrong.

Lex Fridman (2:03:14.400)

But who cares if you come around

Lex Fridman (2:03:16.200)

and realize your mistake and fix it.

Lex Fridman (2:03:18.600)

It becomes just like you said, harder and harder

Kevin Systrom (2:03:20.360)

when your ego grows and the number of people around you

Lex Fridman (2:03:23.000)

that say positive things towards you grows.

Kevin Systrom (2:03:26.240)

I actually think it's really valuable that,

Lex Fridman (2:03:29.400)

like let's imagine a counterfactual

Kevin Systrom (2:03:31.560)

where Instagram became worth like $300 billion

Lex Fridman (2:03:34.760)

or something crazy, right?

Kevin Systrom (2:03:37.280)

I kind of like that my life is relatively normal now.

Lex Fridman (2:03:40.240)

When I say relatively, you get what I mean.

Kevin Systrom (2:03:41.840)

I'm not making a claim that I live a normal life,

Lex Fridman (2:03:44.260)

but like I certainly don't live in a world

Kevin Systrom (2:03:46.180)

where there are like 15 Sherpas following me,

Lex Fridman (2:03:49.460)

like fetching me water, like that's not how it works.

Kevin Systrom (2:03:52.600)

I actually like that I have a sense of humility of like,

Lex Fridman (2:03:56.540)

I may not found another thing that's nearly as big

Lex Fridman (2:03:59.160)

so I have to work twice as hard

Lex Fridman (2:04:01.140)

or I have to like learn twice as much.

Kevin Systrom (2:04:04.420)

I have to, we haven't talked about machine learning yet,

Lex Fridman (2:04:07.320)

but my favorite thing is all these like famous,

Kevin Systrom (2:04:11.260)

you know, tech guys who have worked in the industry,

Lex Fridman (2:04:14.980)

pontificating about the future of machine learning

Lex Fridman (2:04:17.280)

and how it's gonna kill us all, right?

Lex Fridman (2:04:19.080)

And like, I'm pretty sure they've never tried

Kevin Systrom (2:04:22.340)

to build anything with machine learning themselves.

Lex Fridman (2:04:24.400)

Yes, so there's a nice line between people

Kevin Systrom (2:04:27.920)

that actually build stuff with machine,

Lex Fridman (2:04:29.640)

like actually program something

Kevin Systrom (2:04:31.720)

or at least understand some of those fundamentals

Lex Fridman (2:04:33.720)

and the people that are just saying philosophical stuff

Kevin Systrom (2:04:36.840)

or journalists and so on.

Lex Fridman (2:04:38.460)

It's an interesting line to walk

Kevin Systrom (2:04:40.840)

because the people who program are often not philosophers.

Lex Fridman (2:04:44.900)

Or don't have the attention,

Kevin Systrom (2:04:45.840)

they can't write an op ed for the Wall Street Journal,

Lex Fridman (2:04:48.080)

like it doesn't work.

Lex Fridman (2:04:49.080)

So like, it's nice to be both a little bit,

Lex Fridman (2:04:51.380)

like to have elements of both.

Kevin Systrom (2:04:52.880)

My point is the fact that I have to learn stuff from scratch

Lex Fridman (2:04:56.220)

or that I choose to are like.

Kevin Systrom (2:04:58.160)

It's humbling.

Lex Fridman (2:04:59.760)

Yeah, I mean, again, I have a lot of advantages.

Kevin Systrom (2:05:02.960)

I like, but my point is it's awesome

Lex Fridman (2:05:06.560)

to be back in a game where you have to fight.

Kevin Systrom (2:05:11.800)

That is, that's fun.

Lex Fridman (2:05:13.180)

So being humble, being vulnerable,

Kevin Systrom (2:05:16.300)

it's an important aspect of a leader

Lex Fridman (2:05:18.040)

and I hope it serves me well,

Lex Fridman (2:05:19.240)

but like, I can't fast forward 10 years to now.

Lex Fridman (2:05:22.000)

I've just, that's my game plan.

Kevin Systrom (2:05:24.040)

Before I forget, I have to ask you one last thing

Lex Fridman (2:05:26.120)

on Instagram.

Lex Fridman (2:05:28.920)

What do you think about the whistleblower,

Lex Fridman (2:05:30.760)

Frances Haugen, recently coming out

Lex Fridman (2:05:33.400)

and saying that Facebook is aware of Instagram's

Lex Fridman (2:05:36.360)

harmful effect on teenage girls

Lex Fridman (2:05:39.960)

as per their own internal research studies on the matter?

Lex Fridman (2:05:43.320)

What do you think about this baby of yours,

Kevin Systrom (2:05:46.040)

Instagram being under fire now,

Lex Fridman (2:05:48.240)

as we've been talking about under the leadership of Facebook?

Lex Fridman (2:05:54.260)

You know, I often question, where does the blame lie?

Lex Fridman (2:05:59.260)

Is the blame at the people that originated the network me?

Kevin Systrom (2:06:06.780)

Is the blame at like the decision to combine the network

Lex Fridman (2:06:10.540)

with another network with a certain set of values?

Lex Fridman (2:06:15.780)

Is the blame at how it gets run after I left?

Lex Fridman (2:06:20.500)

Like, is it the driver or is it the car, right?

Lex Fridman (2:06:25.820)

Is it that someone enabled these devices in the first place?

Lex Fridman (2:06:29.180)

If you go to an extreme, right?

Lex Fridman (2:06:31.780)

Or is it the users themselves, just human nature?

Lex Fridman (2:06:35.740)

Is it just the way of human nature?

Kevin Systrom (2:06:37.740)

Sure, and like the idea that we're gonna find

Lex Fridman (2:06:39.840)

a mutually exclusive answer here is crazy.

Kevin Systrom (2:06:42.020)

There's not one place that's a combination

Lex Fridman (2:06:43.980)

of a lot of these things.

Lex Fridman (2:06:45.660)

And then the question is like, is it true at all, right?

Lex Fridman (2:06:48.020)

Like I'm not actually saying that's not true

Kevin Systrom (2:06:50.820)

or that it's true, but there's always more nuance here.

Lex Fridman (2:06:55.520)

Do I believe that social media has an effect

Lex Fridman (2:06:59.260)

on young people?

Lex Fridman (2:07:00.260)

Well, it's got it, they use it a lot.

Lex Fridman (2:07:02.380)

And I bet you there are a lot of positive effects

Lex Fridman (2:07:04.300)

and I bet you there are negative effects,

Kevin Systrom (2:07:05.740)

just like any technology.

Lex Fridman (2:07:07.620)

And where I've come to in my thinking on this

Kevin Systrom (2:07:10.180)

is that I think any technology has negative side effects.

Lex Fridman (2:07:13.420)

The question is, as a leader, what do you do about them?

Lex Fridman (2:07:16.540)

And are you actively working on them

Lex Fridman (2:07:18.180)

or do you just like not really believe in them?

Kevin Systrom (2:07:20.620)

If you're a leader that sits there and says,

Lex Fridman (2:07:22.380)

well, we're gonna put an enormous amount

Kevin Systrom (2:07:24.060)

of resources against this.

Lex Fridman (2:07:26.280)

We're gonna acknowledge when there are true criticisms,

Kevin Systrom (2:07:29.580)

we're gonna be vulnerable and that we're not perfect

Lex Fridman (2:07:32.620)

and we're gonna go fix them

Lex Fridman (2:07:33.700)

and we're gonna be held accountable along the way.

Lex Fridman (2:07:37.220)

I think that people generally really respect that.

Lex Fridman (2:07:41.540)

But I think that where Facebook I think has had issues

Lex Fridman (2:07:44.340)

in the past is where they say things like,

Kevin Systrom (2:07:46.880)

can't remember what Mark said about misinformation

Lex Fridman (2:07:49.500)

during the election.

Kevin Systrom (2:07:50.340)

There was that like famous quote where he was like,

Lex Fridman (2:07:52.640)

it's pretty crazy to think that Facebook had anything

Kevin Systrom (2:07:54.780)

to do with this election.

Lex Fridman (2:07:55.660)

Like that was something like that quote.

Lex Fridman (2:07:57.140)

And I don't remember what stage he was on and yeah.

Lex Fridman (2:08:00.340)

But ooh, that did not age well, right?

Kevin Systrom (2:08:02.580)

Like you have to be willing to say,

Lex Fridman (2:08:05.660)

well, maybe there's something there and wow,

Kevin Systrom (2:08:09.700)

like I wanna go look into it

Lex Fridman (2:08:11.060)

and truly believe it in your gut.

Lex Fridman (2:08:12.900)

But if people look at you and how you act

Lex Fridman (2:08:14.580)

and what you say and don't believe you truly feel that way.

Kevin Systrom (2:08:18.680)

It's not just the words you say, but how you say them

Lex Fridman (2:08:20.860)

and that people believe they actually feel the pain

Kevin Systrom (2:08:23.780)

of having caused any suffering in the world.

Lex Fridman (2:08:25.380)

So to me, it's much more about your actions

Lex Fridman (2:08:29.320)

and your posture post event

Lex Fridman (2:08:31.540)

than it is about debugging the why.

Kevin Systrom (2:08:33.060)

Cause I don't know, is it like, I don't know this research.

Lex Fridman (2:08:36.780)

It was written well after I left, right?

Lex Fridman (2:08:38.440)

Like, is it the algorithm?

Lex Fridman (2:08:41.060)

Is it the explore page?

Kevin Systrom (2:08:42.360)

Is it the people you might know unit connecting you

Lex Fridman (2:08:45.460)

to ideas that are dangerous?

Kevin Systrom (2:08:47.860)

Like I really don't know.

Lex Fridman (2:08:51.060)

So we'd have to have a much deeper,

Kevin Systrom (2:08:53.580)

I think dive to understand where the blame lies.

Lex Fridman (2:08:56.580)

What's very unpleasant to me to consider,

Kevin Systrom (2:08:58.700)

now, I don't know if this is true,

Lex Fridman (2:09:00.060)

but to consider the very fact that there might be

Kevin Systrom (2:09:03.660)

some complicated games being played here.

Lex Fridman (2:09:07.020)

For example, as somebody, I really love psychology

Lex Fridman (2:09:10.780)

and I love it enough to know that the field

Lex Fridman (2:09:13.540)

is pretty broken in the following way.

Kevin Systrom (2:09:15.660)

It's very difficult to study human beings well at scale

Lex Fridman (2:09:19.980)

because the questions you ask affect the results.

Kevin Systrom (2:09:22.820)

You can basically get any results you want.

Lex Fridman (2:09:25.060)

And so you have an internal Facebook study

Kevin Systrom (2:09:27.300)

that asks some question of which we don't know

Lex Fridman (2:09:29.680)

the full details and there's some kind of analysis,

Lex Fridman (2:09:32.580)

but that's just the one little tiny slice

Lex Fridman (2:09:34.820)

into some much bigger picture.

Lex Fridman (2:09:37.060)

And so you can have thousands of employees at Facebook.

Lex Fridman (2:09:40.940)

One of them comes out and picks whatever narrative,

Kevin Systrom (2:09:44.360)

knowing that they become famous,

Lex Fridman (2:09:46.460)

coupled with the other really uncomfortable thing

Kevin Systrom (2:09:49.200)

I see in the world, which is journalists seem to understand

Lex Fridman (2:09:53.280)

they get a lot of clickbait attention

Kevin Systrom (2:09:55.840)

from saying something negative about social networks.

Lex Fridman (2:09:58.420)

Certain companies, like they even get some clickbait stuff

Kevin Systrom (2:10:03.780)

about Tesla or about, especially when it's like,

Lex Fridman (2:10:07.700)

when there's a public famous CEO type of person,

Kevin Systrom (2:10:11.060)

if they get a lot of views on the negative, not the positive,

Lex Fridman (2:10:14.500)

the positive, they'll get, I mean,

Kevin Systrom (2:10:16.380)

it actually goes to the thing you were saying before,

Lex Fridman (2:10:18.300)

if there's a hot, sexy new product,

Kevin Systrom (2:10:20.780)

that's great to look forward to, they get positive on that,

Lex Fridman (2:10:23.540)

but absent a product, it's nice to have like the CEO

Kevin Systrom (2:10:28.580)

messing up in some kind of way.

Lex Fridman (2:10:30.260)

And so couple that with the whistleblower

Lex Fridman (2:10:33.580)

and with this whole dynamic of journalism and so on,

Lex Fridman (2:10:38.300)

you know, with Social Dilemma being really popular,

Kevin Systrom (2:10:41.180)

documentary, it's like, all right,

Lex Fridman (2:10:44.620)

my concern is there's deep flaws in human nature here

Kevin Systrom (2:10:48.800)

in terms of things we need to deal with,

Lex Fridman (2:10:51.300)

like the nature of hate, of bullying,

Kevin Systrom (2:10:54.420)

all those kinds of things.

Lex Fridman (2:10:56.280)

And then there's people who are trying to use that

Kevin Systrom (2:11:00.060)

potentially to become famous and make money

Lex Fridman (2:11:02.420)

off of blaming others for causing more of the problem

Kevin Systrom (2:11:06.980)

as opposed to helping solve the problem.

Lex Fridman (2:11:08.860)

So I don't know what to think.

Kevin Systrom (2:11:10.220)

I'm not saying this is like, I'm just uncomfortable

Lex Fridman (2:11:12.900)

with, I guess, not knowing what to think about any of this

Kevin Systrom (2:11:16.740)

because a bunch of folks I know that work at Facebook

Lex Fridman (2:11:19.500)

on the machine learning side, so Yann LeCun,

Kevin Systrom (2:11:21.500)

I mean, they're quite upset about what's happening

Lex Fridman (2:11:25.160)

because there's a lot of really brilliant,

Kevin Systrom (2:11:26.860)

good people inside Facebook that are trying to do good.

Lex Fridman (2:11:30.700)

And so like all of this press, Yann is one of them,

Lex Fridman (2:11:33.420)

and he has an amazing team

Lex Fridman (2:11:34.500)

with the machine learning researchers.

Kevin Systrom (2:11:35.740)

Like he's really upset with the fact

Lex Fridman (2:11:38.540)

that people don't seem to understand

Kevin Systrom (2:11:40.540)

that the portrayal does not represent

Lex Fridman (2:11:43.660)

the full nature of efforts that's going on at Facebook.

Lex Fridman (2:11:46.620)

So I don't know what to think about that.

Lex Fridman (2:11:48.220)

Well, you just, I think, very helpfully explained

Kevin Systrom (2:11:52.700)

the nuance of the situation

Lex Fridman (2:11:54.060)

and why it's so hard to understand.

Lex Fridman (2:11:56.300)

But a couple of things.

Lex Fridman (2:11:57.140)

One is I think I have been surprised

Kevin Systrom (2:12:02.140)

at the scale with which some product manager

Lex Fridman (2:12:11.940)

can do an enormous amount of harm

Kevin Systrom (2:12:14.380)

to a very, very large company

Lex Fridman (2:12:17.260)

by releasing a trove of documents.

Kevin Systrom (2:12:19.260)

Like I think I read a couple of them when they got published

Lex Fridman (2:12:21.940)

and I haven't even spent any time going deep.

Kevin Systrom (2:12:24.140)

Part of it's like, I don't really feel like reliving

Lex Fridman (2:12:26.040)

a previous life, but wow.

Kevin Systrom (2:12:30.780)

Like talk about challenging the idea of open culture

Lex Fridman (2:12:34.420)

and like what that does to Facebook internally.

Kevin Systrom (2:12:37.540)

If Facebook was built, like I remember like my office,

Lex Fridman (2:12:43.340)

we had this like no visitors rule around my office

Kevin Systrom (2:12:45.700)

because we always had like confidential stuff up

Lex Fridman (2:12:47.500)

on the walls and everyone was super angry

Kevin Systrom (2:12:50.340)

because they're like, that goes against our culture

Lex Fridman (2:12:52.040)

of transparency and like marks in the fish cube

Kevin Systrom (2:12:54.700)

or whatever they call it, the aquarium,

Lex Fridman (2:12:56.260)

I think they called it, where like literally anyone could see

Lex Fridman (2:12:59.300)

what he was doing at any point and I don't know.

Lex Fridman (2:13:03.300)

I mean, other companies like Apple have been quiet

Kevin Systrom (2:13:06.540)

slash locked down, Snapchat's the same way for a reason.

Lex Fridman (2:13:10.240)

And I don't know what this does to transparency

Kevin Systrom (2:13:14.260)

on the inside of startups that value that.

Lex Fridman (2:13:16.820)

I think that it's a seminal moment and you can say,

Lex Fridman (2:13:20.140)

well, you should have nothing to hide, right?

Lex Fridman (2:13:22.300)

But to your point, you can pick out documents

Lex Fridman (2:13:24.620)

that show anything, right?

Lex Fridman (2:13:27.660)

But I don't know.

Lex Fridman (2:13:28.500)

What happens to transparency inside of startups

Lex Fridman (2:13:31.380)

and the culture that startups or companies

Kevin Systrom (2:13:35.440)

in the future will grow, like the startup of the future

Lex Fridman (2:13:37.980)

that becomes the next Facebook will be locked down

Lex Fridman (2:13:40.420)

and what does that do, right?

Lex Fridman (2:13:42.580)

So that's part one.

Kevin Systrom (2:13:44.220)

Part two, like I don't think that you could design

Lex Fridman (2:13:49.500)

a more like a well orchestrated handful of events

Kevin Systrom (2:13:54.500)

from the like 16 minutes to releasing the documents

Lex Fridman (2:13:59.820)

in the way that they were released at the right time.

Kevin Systrom (2:14:02.420)

That takes a lot of planning and partnership.

Lex Fridman (2:14:05.580)

And it seems like she has a partner at some firm, right?

Kevin Systrom (2:14:09.060)

That probably helped a lot with this, but man,

Lex Fridman (2:14:12.940)

at a personal level, if you're her,

Kevin Systrom (2:14:15.340)

you'd have to really believe in what you are doing,

Lex Fridman (2:14:19.860)

really believe in it because you are personally

Lex Fridman (2:14:22.620)

putting your ass on the line, right?

Lex Fridman (2:14:25.140)

Like you've got a very large company

Lex Fridman (2:14:29.020)

that doesn't like enemies, right?

Lex Fridman (2:14:33.700)

It takes a lot of guts and I don't love

Kevin Systrom (2:14:37.940)

these conspiracy theories about like,

Lex Fridman (2:14:39.700)

oh, she's being financed from some person or people.

Kevin Systrom (2:14:42.380)

Like I don't love them because that's

Lex Fridman (2:14:43.700)

like the easy thing to say.

Kevin Systrom (2:14:45.680)

I think the Occam's razor here is like someone thought

Lex Fridman (2:14:49.340)

they were doing something wrong

Lex Fridman (2:14:51.300)

and was like very, very courageous.

Lex Fridman (2:14:54.540)

And I don't know if courageous is the word,

Lex Fridman (2:14:57.500)

but like, so without getting into like,

Lex Fridman (2:15:00.540)

is she a martyr?

Lex Fridman (2:15:01.800)

Is she courageous?

Lex Fridman (2:15:02.640)

Is she right?

Kevin Systrom (2:15:03.480)

Like, let's put that aside for a second.

Lex Fridman (2:15:05.860)

Then there are the documents themselves.

Kevin Systrom (2:15:07.460)

They say what they say.

Lex Fridman (2:15:09.580)

To your point, a lot of the things that like people

Kevin Systrom (2:15:12.420)

have been worried about already in the documents

Lex Fridman (2:15:15.060)

or they're already been said externally.

Lex Fridman (2:15:17.260)

And I don't know, I'm just like, I'm thankful

Lex Fridman (2:15:22.340)

that I am focused on new things with my life.

Kevin Systrom (2:15:25.820)

Well, let me just say, I just think it's a really

Lex Fridman (2:15:27.860)

hard problem that probably Facebook and Twitter

Kevin Systrom (2:15:30.860)

are trying to solve.

Lex Fridman (2:15:32.280)

I'm actually just fascinated by how hard this problem is.

Kevin Systrom (2:15:35.340)

There are fundamental issues at Facebook in tone

Lex Fridman (2:15:38.380)

and in an approach of how product gets built

Lex Fridman (2:15:41.740)

and the objective functions.

Lex Fridman (2:15:43.100)

And since people, organizations are not people.

Lex Fridman (2:15:48.280)

So yawn and fair, right?

Lex Fridman (2:15:50.000)

Like there are a lot of really great people

Kevin Systrom (2:15:51.740)

who like literally just want to push

Lex Fridman (2:15:53.380)

reinforcement learning forward.

Kevin Systrom (2:15:55.220)

They literally just want to teach a robot

Lex Fridman (2:15:56.940)

to touch, feel, lift, right?

Kevin Systrom (2:15:59.380)

Like they're not thinking about political misinformation,

Lex Fridman (2:16:02.820)

right?

Lex Fridman (2:16:03.880)

But there's a strong connection between what funds

Lex Fridman (2:16:07.440)

that research and an enormously profitable machine

Kevin Systrom (2:16:10.740)

that has trade offs.

Lex Fridman (2:16:13.360)

And one cannot separate the two.

Kevin Systrom (2:16:18.200)

You are not completely separate from the system.

Lex Fridman (2:16:21.400)

So I agree, it can feel really frustrating

Kevin Systrom (2:16:24.520)

to feel if you're internal there,

Lex Fridman (2:16:27.220)

that you're working on something completely unrelated

Lex Fridman (2:16:29.600)

and you feel like your group's good.

Lex Fridman (2:16:31.200)

I can understand that.

Lex Fridman (2:16:32.840)

But there's some responsibility still.

Lex Fridman (2:16:34.440)

You have to acknowledge, it's like the Ray Dalio thing.

Kevin Systrom (2:16:36.440)

You have to look in the mirror and see if there's problems

Lex Fridman (2:16:38.960)

and you have to fix those problems.

Kevin Systrom (2:16:40.280)

Yeah.

Lex Fridman (2:16:43.360)

You mentioned machine learning reinforcement quite a bit.

Kevin Systrom (2:16:46.720)

I mean, to me, social networks is one of the exciting places,

Lex Fridman (2:16:50.200)

recommender systems where machine learning is applied.

Kevin Systrom (2:16:52.840)

Where else in the world, in the space of possibilities

Lex Fridman (2:16:56.440)

over the next five, 10, 20 years,

Lex Fridman (2:16:58.200)

do you think we're going to see impact of machine learning

Lex Fridman (2:17:02.140)

when you try, on the philosophical level,

Lex Fridman (2:17:05.000)

on a technical level, what do you think?

Lex Fridman (2:17:07.280)

Or within social networks themselves?

Kevin Systrom (2:17:11.720)

Well, I think the obvious answers are climate change.

Lex Fridman (2:17:17.120)

Think about how much fuel

Kevin Systrom (2:17:19.800)

or just waste there is in energy consumption today

Lex Fridman (2:17:24.520)

because we don't plan accordingly,

Kevin Systrom (2:17:27.460)

because we take the least efficient route or...

Lex Fridman (2:17:30.960)

The logistics and stuff, the supply chain,

Kevin Systrom (2:17:32.640)

all that kind of stuff.

Lex Fridman (2:17:33.480)

Yeah, I mean, listen, if we're gonna fight climate change,

Kevin Systrom (2:17:36.600)

like one really way, one awesome way to do it

Lex Fridman (2:17:39.920)

is figure out how to optimize how we operate as a species

Lex Fridman (2:17:43.280)

and minimize the amount of energy we consume

Lex Fridman (2:17:47.220)

to maximize whatever economic impact we wanna have.

Kevin Systrom (2:17:51.080)

Because right now those two are very much tied together.

Lex Fridman (2:17:53.900)

And I don't believe that that has to be the case.

Kevin Systrom (2:17:56.520)

There's this really interesting, you've read it.

Lex Fridman (2:18:00.000)

For people who are listening,

Kevin Systrom (2:18:00.840)

there's this really interesting paper

Lex Fridman (2:18:02.680)

on reinforcement learning

Lex Fridman (2:18:04.000)

and energy consumption inside buildings.

Lex Fridman (2:18:06.200)

It's like one of the seminal ones, right?

Lex Fridman (2:18:08.280)

But imagine that at massive scale.

Lex Fridman (2:18:10.280)

That's super interesting.

Kevin Systrom (2:18:11.280)

I mean, they've done resource planning for servers

Lex Fridman (2:18:15.300)

for peak load using reinforcement learning.

Kevin Systrom (2:18:17.800)

I don't know if that was at Google or somewhere else,

Lex Fridman (2:18:19.560)

but like, okay, great, you do it for servers,

Lex Fridman (2:18:21.880)

but what if you could do it for just capacity

Lex Fridman (2:18:24.040)

and general energy capacity for cities

Lex Fridman (2:18:26.200)

and planning for traffic?

Lex Fridman (2:18:27.520)

And of course there's all the self driving cars

Lex Fridman (2:18:30.960)

and I don't know, like I'm not gonna pontificate

Lex Fridman (2:18:34.060)

on like crazy ideas using reinforcement learning

Kevin Systrom (2:18:39.320)

or machine learning.

Lex Fridman (2:18:40.160)

It's just so clear to me

Kevin Systrom (2:18:41.240)

that humans don't think quickly enough.

Lex Fridman (2:18:43.640)

So it's interesting to think about machine learning

Kevin Systrom (2:18:46.520)

helping a little bit at scale.

Lex Fridman (2:18:49.680)

So a little bit to a large number of people

Kevin Systrom (2:18:52.000)

that has a huge impact.

Lex Fridman (2:18:53.460)

So if you optimize, say Google Maps, something like that,

Kevin Systrom (2:18:57.320)

trajectory planning or what a map quest first.

Lex Fridman (2:19:01.200)

Getting here, I looked and it was like,

Kevin Systrom (2:19:02.960)

here's the most energy efficient route.

Lex Fridman (2:19:04.520)

And I was like, I'm gonna be late.

Kevin Systrom (2:19:05.680)

I need to take the fastest route.

Lex Fridman (2:19:07.280)

As opposed to unrolling the map.

Kevin Systrom (2:19:09.160)

Yeah, yeah.

Lex Fridman (2:19:10.640)

Like, and then that's going to be very inefficient

Kevin Systrom (2:19:12.960)

no matter what.

Lex Fridman (2:19:13.800)

I was definitely the other day,

Kevin Systrom (2:19:14.800)

like part of the Epsilon of Epsilon Greedy

Lex Fridman (2:19:17.820)

with Waze where like I was sent on like a weird route

Kevin Systrom (2:19:21.800)

that I could tell they're like,

Lex Fridman (2:19:22.960)

we just need to collect data at this road.

Kevin Systrom (2:19:24.960)

Like we just, Kevin's.

Lex Fridman (2:19:26.920)

You were the ant they sent out for exploration.

Kevin Systrom (2:19:28.760)

Kevin's definitely gonna be the guinea pig.

Lex Fridman (2:19:30.400)

And great, now we have.

Lex Fridman (2:19:32.120)

Did you at least feel pride?

Lex Fridman (2:19:34.160)

Oh, going through it, I was like, oh, this is fun.

Kevin Systrom (2:19:36.000)

Like now they get data about this weird shortcut.

Lex Fridman (2:19:38.280)

And actually I hit all the green lights networked.

Kevin Systrom (2:19:40.180)

I'm like, this is a problem.

Lex Fridman (2:19:41.880)

Bad data.

Kevin Systrom (2:19:42.720)

Bad data, they're just gonna imagine.

Lex Fridman (2:19:44.680)

I could see you slowing down and stopping at a green light

Kevin Systrom (2:19:47.520)

just to give them the right kind of data.

Lex Fridman (2:19:50.080)

But to answer your question,

Kevin Systrom (2:19:51.000)

like I feel like that was fairly unsatisfying

Lex Fridman (2:19:53.400)

and it's easy to say climate change.

Lex Fridman (2:19:55.000)

But what I would say is at Instagram,

Lex Fridman (2:19:58.640)

everything we applied machine learning to

Kevin Systrom (2:20:02.120)

got better for users and it got better for the company.

Lex Fridman (2:20:04.800)

I saw the power.

Kevin Systrom (2:20:06.200)

I didn't fully understand it as an executive.

Lex Fridman (2:20:08.200)

And I think that's actually one of the issues

Kevin Systrom (2:20:10.640)

that, and when I say understand,

Lex Fridman (2:20:12.860)

I mean the mathematics of it.

Kevin Systrom (2:20:14.020)

Like I understand what it does.

Lex Fridman (2:20:15.320)

I understand that it helps.

Lex Fridman (2:20:17.040)

But there are a lot of executives now

Lex Fridman (2:20:20.120)

that talk about it in the way

Kevin Systrom (2:20:22.080)

that they talk about the internet

Lex Fridman (2:20:23.680)

or they talked about the internet like 10 years ago.

Kevin Systrom (2:20:25.560)

They're like, we're gonna build mobile.

Lex Fridman (2:20:27.040)

And you're like, what does that mean?

Kevin Systrom (2:20:27.940)

They're like, we're just gonna do mobile.

Lex Fridman (2:20:29.400)

And you're like, okay.

Lex Fridman (2:20:31.220)

So my sense is the next generation of leaders

Lex Fridman (2:20:33.900)

will have grown up having had classes

Kevin Systrom (2:20:37.160)

in reinforcement learning, supervised learning, whatever.

Lex Fridman (2:20:40.160)

And they will be able to thoughtfully apply it

Kevin Systrom (2:20:42.880)

to their companies and the places that it is needed most.

Lex Fridman (2:20:46.280)

And that's really cool.

Kevin Systrom (2:20:47.660)

Cause I mean, talk about efficiency gains.

Lex Fridman (2:20:53.040)

That's what excites me the most about it.

Kevin Systrom (2:20:54.760)

Yeah, so there's, it's interesting just to get a fundamental

Lex Fridman (2:20:58.080)

first principles understanding

Kevin Systrom (2:20:59.440)

of certain concepts of machine learning.

Lex Fridman (2:21:01.000)

So supervised learning from an executive perspective,

Kevin Systrom (2:21:04.400)

supervised learning, you have to have a lot of humans

Lex Fridman (2:21:07.020)

label a lot of data.

Lex Fridman (2:21:08.440)

So the question there is, okay,

Lex Fridman (2:21:09.960)

can we gather a large amount of data

Lex Fridman (2:21:12.380)

that can be labeled well?

Lex Fridman (2:21:14.360)

And that's the question Tesla asked,

Kevin Systrom (2:21:16.080)

like can we create a data engine

Lex Fridman (2:21:17.760)

that keeps sending an imperfect machine learning system

Kevin Systrom (2:21:22.040)

out there, whenever it fails, it gives us data back.

Lex Fridman (2:21:25.000)

We label it by human and we send it back and forth

Kevin Systrom (2:21:27.640)

to this way.

Lex Fridman (2:21:28.640)

Then there's Yann LeCun's excited

Kevin Systrom (2:21:30.440)

about the self supervised learning

Lex Fridman (2:21:32.380)

where you do much less human labeling

Lex Fridman (2:21:36.140)

and there's some kind of mechanism for the system

Lex Fridman (2:21:38.880)

to learn it by itself on the human generated data.

Lex Fridman (2:21:42.760)

And then there's the reinforcement learning,

Lex Fridman (2:21:44.760)

which is like basically allowing,

Kevin Systrom (2:21:47.800)

it's applying the alpha zero technology

Lex Fridman (2:21:52.080)

that allow through self play to learn how to solve

Kevin Systrom (2:21:56.080)

the game of Go and achieve incredible levels

Lex Fridman (2:21:59.920)

at the game of chess.

Lex Fridman (2:22:02.580)

Can you formulate the problem you're trying to solve

Lex Fridman (2:22:05.160)

in a way that's amenable to reinforcement learning?

Lex Fridman (2:22:07.600)

And can you get the right kind of signal at scale?

Lex Fridman (2:22:09.960)

Cause you need a lot of signal.

Lex Fridman (2:22:11.840)

And that's kind of fascinating to see which part

Lex Fridman (2:22:15.480)

of a social network can you convert

Kevin Systrom (2:22:17.640)

into reinforcement learning problem.

Lex Fridman (2:22:19.440)

The fascinating thing about reinforcement learning,

Kevin Systrom (2:22:22.280)

I think, is that we now have learned

Lex Fridman (2:22:25.960)

to apply neural networks to guess the Q function,

Kevin Systrom (2:22:34.260)

basically the values for any state in action.

Lex Fridman (2:22:37.920)

And that is fascinating cause we used to just like,

Kevin Systrom (2:22:40.480)

I don't know, have like a linear regression,

Lex Fridman (2:22:42.160)

like hope it worked and that was the fanciest version of it.

Lex Fridman (2:22:45.320)

But now you look at it, I'm like trying to learn this stuff

Lex Fridman (2:22:47.400)

and I look at it and I'm like,

Kevin Systrom (2:22:48.440)

there are like 17 different acronyms of different ways

Lex Fridman (2:22:51.640)

you can try to apply this.

Kevin Systrom (2:22:52.760)

No one quite agrees, like what's the best.

Lex Fridman (2:22:56.080)

Generally, if you're trying to like build a neural network,

Kevin Systrom (2:22:58.880)

there are pretty well trodden ways of doing that.

Lex Fridman (2:23:02.120)

You use Adam, you use ReLU,

Kevin Systrom (2:23:04.520)

like there's just like general good ideas.

Lex Fridman (2:23:07.360)

And in reinforcement learning,

Kevin Systrom (2:23:08.840)

I feel like the consensus is like, it totally depends.

Lex Fridman (2:23:13.720)

And by the way, it's really hard to get it to converge

Lex Fridman (2:23:16.440)

and it's noisy and it like,

Lex Fridman (2:23:18.640)

so there are all these really interesting ideas

Kevin Systrom (2:23:20.640)

around building simulators.

Lex Fridman (2:23:23.600)

You know, like for instance, in self driving, right?

Kevin Systrom (2:23:25.700)

Like you don't want to like actually have someone

Lex Fridman (2:23:29.300)

getting in an accident to learn that an accident is bad.

Lex Fridman (2:23:31.880)

So you start simulating accidents,

Lex Fridman (2:23:33.920)

simulating aggressive drivers,

Kevin Systrom (2:23:35.440)

just simulating crazy dogs that run into the street and,

Lex Fridman (2:23:39.120)

wow, fascinating, right?

Kevin Systrom (2:23:41.280)

Like my mind starts racing and then the question is,

Lex Fridman (2:23:43.600)

okay, forget about self driving cars.

Kevin Systrom (2:23:45.920)

Let's talk about social networks.

Lex Fridman (2:23:49.780)

How can you produce a better, more thoughtful experience

Lex Fridman (2:23:52.960)

using these types of algorithms?

Lex Fridman (2:23:55.840)

And honestly, in talking to some of the people

Kevin Systrom (2:23:58.200)

that work at Facebook and old Instagrammers,

Lex Fridman (2:24:01.640)

most people are like, yeah, we tried a lot of things,

Kevin Systrom (2:24:04.160)

didn't quite ever make it work.

Lex Fridman (2:24:05.600)

I mean, for the longest time,

Lex Fridman (2:24:06.720)

Facebook ads was effectively a logistic regression, okay?

Lex Fridman (2:24:10.600)

I don't know what it is now,

Lex Fridman (2:24:11.740)

but like if you look at this paper

Lex Fridman (2:24:13.160)

that they published back in the day,

Kevin Systrom (2:24:14.080)

it was literally just a logistic regression.

Lex Fridman (2:24:16.120)

Made a lot of money.

Lex Fridman (2:24:18.960)

So even at these like extremely large scales,

Lex Fridman (2:24:21.760)

if we are not yet touching

Lex Fridman (2:24:23.320)

what reinforcement learning can truly do,

Lex Fridman (2:24:25.520)

imagine what the next 10 years looks like.

Lex Fridman (2:24:27.400)

How cool is that?

Lex Fridman (2:24:28.600)

It's amazing.

Lex Fridman (2:24:29.620)

So I really liked the use of reinforcement learning

Lex Fridman (2:24:32.200)

as part of the simulation, for example,

Kevin Systrom (2:24:34.760)

like with self driving cars, it's modeling pedestrians.

Lex Fridman (2:24:37.760)

So the nice thing about reinforcement learning,

Kevin Systrom (2:24:40.440)

it can be used to learn agents within the world.

Lex Fridman (2:24:45.040)

So they can learn to behave properly.

Kevin Systrom (2:24:47.080)

Like you can teach pedestrians to,

Lex Fridman (2:24:49.360)

like you don't hard code the way they behave,

Kevin Systrom (2:24:51.640)

they learn how to behave.

Lex Fridman (2:24:53.280)

In that same way, I do have a hope,

Kevin Systrom (2:24:55.200)

was it Jack Dorsey talks about healthy conversations.

Lex Fridman (2:24:58.280)

You talked about meaningful interactions, I believe.

Kevin Systrom (2:25:03.520)

Like simulating interactions.

Lex Fridman (2:25:06.940)

So you can learn how to manage that, it's fascinating.

Lex Fridman (2:25:09.480)

So where most of your algorithm development happens

Lex Fridman (2:25:13.040)

in virtual worlds, and then you can really learn

Lex Fridman (2:25:16.480)

how to design the interface,

Lex Fridman (2:25:18.120)

how you design a bunch of aspects of the experience

Kevin Systrom (2:25:21.400)

in terms of how you select what's shown in the feed,

Lex Fridman (2:25:24.480)

all those kinds of things.

Kevin Systrom (2:25:26.120)

It feels like if you can connect reinforcement learning

Lex Fridman (2:25:28.480)

to that, that's super exciting.

Kevin Systrom (2:25:30.080)

Yep, and I think if you have a company and leadership

Lex Fridman (2:25:35.080)

that believe in doing the right things

Lex Fridman (2:25:36.560)

and can apply this technology in the right way,

Lex Fridman (2:25:38.920)

some really special stuff can happen.

Kevin Systrom (2:25:41.440)

It is mostly likely going to be a group of people

Lex Fridman (2:25:44.440)

we've never heard about, start up from scratch, right?

Lex Fridman (2:25:49.360)

And you asked if like new social networks could be built,

Lex Fridman (2:25:52.920)

I've got to imagine they will be.

Lex Fridman (2:25:54.300)

And whoever starts it, it might be some kids in a garage

Lex Fridman (2:25:58.400)

that took these classes from these people, you, right?

Lex Fridman (2:26:02.200)

And they're building all of these things

Lex Fridman (2:26:04.400)

with this tech at the core.

Lex Fridman (2:26:06.320)

So I'm trying not to be someone who just like throws

Lex Fridman (2:26:08.400)

around reinforcement learning as a buzzword.

Kevin Systrom (2:26:11.880)

I truly believe that it is the most cutting edge

Lex Fridman (2:26:16.240)

in what can happen in social networks.

Lex Fridman (2:26:18.320)

And I also believe it's super hard.

Lex Fridman (2:26:20.880)

Like it's super hard to make it work.

Kevin Systrom (2:26:22.480)

It's super hard to do it at scale.

Lex Fridman (2:26:24.040)

It's super hard to find people that truly understand it.

Lex Fridman (2:26:26.840)

So I'm not gonna say that like,

Lex Fridman (2:26:30.120)

I think it'll be applied in social networks

Kevin Systrom (2:26:31.760)

before we have true self driving.

Lex Fridman (2:26:33.840)

Yeah, we could argue about this for a long time,

Lex Fridman (2:26:36.920)

but yes, I agree with you.

Lex Fridman (2:26:38.000)

I think self driving is way harder than people realize.

Kevin Systrom (2:26:40.480)

Oh, absolutely.

Lex Fridman (2:26:41.840)

Let me ask you in terms of that kid in the garage

Kevin Systrom (2:26:44.040)

or those couple of kids in the garage,

Lex Fridman (2:26:45.520)

what advice would you give to them

Lex Fridman (2:26:47.040)

if they wanna start a new social network or a business?

Lex Fridman (2:26:50.760)

What advice would you give to somebody

Lex Fridman (2:26:52.280)

with a big dream and a young startup?

Lex Fridman (2:26:56.880)

To me, you have to choose to do something

Lex Fridman (2:26:59.500)

that even if it fails, like it was so fun, right?

Lex Fridman (2:27:03.800)

Like we never started Instagram knowing

Kevin Systrom (2:27:06.680)

it was going to be big.

Lex Fridman (2:27:07.760)

We started Instagram because we loved photography.

Kevin Systrom (2:27:10.480)

We loved social networks.

Lex Fridman (2:27:12.480)

I had seen what other social networks had done

Lex Fridman (2:27:14.520)

and I thought, hmm, maybe we did a spin on this,

Lex Fridman (2:27:17.640)

but like nowhere was our feet predestined.

Kevin Systrom (2:27:20.400)

Like it wasn't like, it wasn't written out anywhere

Lex Fridman (2:27:23.640)

that everything was gonna go great.

Lex Fridman (2:27:25.640)

And I often think the counterfactual,

Lex Fridman (2:27:27.000)

like what if it had not gone well?

Kevin Systrom (2:27:28.820)

I would have been like, I don't know, that was fun.

Lex Fridman (2:27:30.360)

We raised some money, we learned some stuff

Lex Fridman (2:27:32.280)

and does it position you well for the next experience?

Lex Fridman (2:27:37.280)

That's the advice that I would give

Kevin Systrom (2:27:39.240)

to anyone wanting to start something today,

Lex Fridman (2:27:41.200)

which is like, does this meet with your ultimate goals?

Kevin Systrom (2:27:45.120)

Not wealth, not fame, none of that,

Lex Fridman (2:27:47.200)

because all of that, by the way, is bullshit.

Kevin Systrom (2:27:48.840)

Like you can get super famous and super wealthy.

Lex Fridman (2:27:52.360)

And I think generally those are not things that,

Kevin Systrom (2:27:57.400)

again, it's easy to say with like a lot of money

Lex Fridman (2:27:59.420)

that somehow like it's not good to have a lot of money.

Kevin Systrom (2:28:01.800)

It's just, I think that complicates life enormously

Lex Fridman (2:28:04.920)

in a way that people don't fully comprehend.

Lex Fridman (2:28:06.900)

So I think it is way more interesting to shoot for,

Lex Fridman (2:28:09.520)

can I make something that people love,

Kevin Systrom (2:28:11.760)

that provides value in the world,

Lex Fridman (2:28:13.600)

that I love building, that I love working on, right?

Kevin Systrom (2:28:17.520)

That's what I would do if I were starting from scratch.

Lex Fridman (2:28:21.920)

And by the way, like in some ways

Kevin Systrom (2:28:22.960)

that I will do that personally,

Lex Fridman (2:28:25.200)

which is like choose the thing

Kevin Systrom (2:28:26.420)

that you get up every morning and you're like,

Lex Fridman (2:28:27.920)

I love this, even when it's painful.

Kevin Systrom (2:28:33.160)

Even when it's painful.

Lex Fridman (2:28:34.480)

What about a social network specifically?

Kevin Systrom (2:28:36.780)

If you were to imagine, put yourself in the mind of some.

Lex Fridman (2:28:40.600)

I can't compete against myself.

Kevin Systrom (2:28:42.000)

I can't give out ideas.

Lex Fridman (2:28:43.600)

Okay, I got you.

Kevin Systrom (2:28:44.440)

No, but it's like high level.

Lex Fridman (2:28:45.840)

You can focus on community.

Kevin Systrom (2:28:47.600)

Yeah.

Lex Fridman (2:28:50.360)

I said that as a half joke.

Kevin Systrom (2:28:54.160)

In all honesty, I think these things are so hard to build

Lex Fridman (2:28:56.960)

that like ideas are a dime a dozen, but.

Kevin Systrom (2:28:59.960)

You have talked about keeping it simple.

Lex Fridman (2:29:02.920)

Can I tell you?

Kevin Systrom (2:29:03.760)

Which is a liberating idea.

Lex Fridman (2:29:04.920)

My model is it's three circles and they overlap.

Lex Fridman (2:29:08.280)

One circle is what do I have experience at?

Lex Fridman (2:29:11.360)

Slash, what am I good at?

Kevin Systrom (2:29:12.800)

I don't like saying what am I good at

Lex Fridman (2:29:14.140)

because it just like seems like,

Lex Fridman (2:29:16.040)

what do I have experience in, right?

Lex Fridman (2:29:18.280)

What can I bring to the table?

Lex Fridman (2:29:19.720)

What am I excited about is the other circle.

Lex Fridman (2:29:22.040)

What gets it?

Lex Fridman (2:29:22.880)

What's just super cool, right?

Lex Fridman (2:29:24.640)

That I want to work on because even when this is hard,

Kevin Systrom (2:29:29.160)

I think it's so cool.

Lex Fridman (2:29:30.000)

I want to stick with it.

Lex Fridman (2:29:31.720)

And the last circle is like, what does the world need?

Lex Fridman (2:29:34.960)

And if that circle ain't there,

Kevin Systrom (2:29:36.200)

it doesn't matter what you work on.

Lex Fridman (2:29:37.720)

Cause there are a lot of startups that exist

Kevin Systrom (2:29:39.280)

that just no one needs or very small markets need.

Lex Fridman (2:29:43.200)

But if you want to be successful,

Kevin Systrom (2:29:44.840)

I think if you're like, if you're good at it,

Lex Fridman (2:29:47.240)

you have, sorry, if you're good at it,

Kevin Systrom (2:29:49.720)

you're passionate about it and the world needs it.

Lex Fridman (2:29:51.640)

I mean, this sounds simple,

Kevin Systrom (2:29:53.040)

not enough people sit down and just think

Lex Fridman (2:29:54.920)

about those circles and think, do these things overlap?

Lex Fridman (2:29:58.120)

And then can I get that middle section?

Lex Fridman (2:29:59.480)

It's small, but can I get that middle section?

Kevin Systrom (2:30:02.840)

I think a lot about that personally.

Lex Fridman (2:30:05.960)

And then you have to be really honest about the circle

Kevin Systrom (2:30:09.040)

that you're good at and really honest about the circle

Lex Fridman (2:30:14.400)

that the world needs.

Lex Fridman (2:30:17.080)

And as opposed to really honest about the passion,

Lex Fridman (2:30:20.280)

like what do you actually love?

Kevin Systrom (2:30:22.520)

As opposed to like some kind of dream of making money,

Lex Fridman (2:30:24.600)

all those kinds of stuff, like literally love doing.

Kevin Systrom (2:30:26.600)

I had a former engineer who decided to start a startup

Lex Fridman (2:30:29.960)

and I was like, are you sure you want to start a company

Lex Fridman (2:30:32.080)

versus like join something else?

Lex Fridman (2:30:34.440)

Because being a coach of an NBA team and playing basketball

Kevin Systrom (2:30:39.640)

are two very, very different things.

Lex Fridman (2:30:42.320)

And like not everyone fully understands the difference.

Kevin Systrom (2:30:45.880)

I think you can kind of do it both.

Lex Fridman (2:30:50.160)

And I don't know, jury's out on that one

Kevin Systrom (2:30:51.960)

because like they're in the middle of it now.

Lex Fridman (2:30:54.600)

But it's really important to figure out

Lex Fridman (2:30:57.640)

what you're good at, not be full of yourself,

Lex Fridman (2:30:59.880)

like truly look at your track record.

Kevin Systrom (2:31:03.720)

What's the saying like, it ain't bragging if you can do it.

Lex Fridman (2:31:09.200)

But too many people are delusional

Lex Fridman (2:31:12.280)

and like think they're better at things

Lex Fridman (2:31:14.280)

than they actually are,

Kevin Systrom (2:31:15.480)

or think there's a bigger market than there actually is.

Lex Fridman (2:31:18.640)

When you confuse your passion for things with a big market,

Lex Fridman (2:31:21.080)

that's really scary, right?

Lex Fridman (2:31:23.600)

Like just because you think it's cool

Kevin Systrom (2:31:25.000)

doesn't mean that it's a big business opportunity.

Lex Fridman (2:31:27.120)

So like, what evidence do you have?

Kevin Systrom (2:31:28.760)

Again, I'm a fairly like, I'm a strict rationalist on this.

Lex Fridman (2:31:32.600)

And like sometimes people don't like working with me

Kevin Systrom (2:31:35.000)

because I'm pretty pragmatic about things.

Lex Fridman (2:31:37.160)

Like I'm not Elon, like I don't sit

Lex Fridman (2:31:40.440)

and make bold proclamations about visiting Mars.

Lex Fridman (2:31:44.160)

Like that's just not how I work.

Kevin Systrom (2:31:46.160)

I'm like, okay, I want to build this really cool thing

Lex Fridman (2:31:48.480)

that's fairly practical and I think we could do it.

Lex Fridman (2:31:50.840)

And it's in this way.

Lex Fridman (2:31:52.320)

And what's cool though is like, that's just my sweet spot.

Kevin Systrom (2:31:55.040)

I'm not like, I just, I can't with a straight face

Lex Fridman (2:31:58.760)

talk about the metaphors.

Kevin Systrom (2:31:59.680)

I can't, I just, it's not me.

Lex Fridman (2:32:01.960)

What do you think about the Facebook renaming itself to?

Kevin Systrom (2:32:05.400)

I didn't mean that as a dig.

Lex Fridman (2:32:06.440)

I just literally mean like, I'm fairly,

Kevin Systrom (2:32:09.440)

I like to live in the next five years.

Lex Fridman (2:32:11.200)

And like, what things can I get out in a year

Lex Fridman (2:32:13.520)

that people will use at scale?

Lex Fridman (2:32:15.040)

And so it's just, again, those circles I think are different

Kevin Systrom (2:32:20.360)

for different people, but it's important to realize

Lex Fridman (2:32:22.200)

that like market matters, you being good at it matters

Lex Fridman (2:32:26.040)

and having passion for it matters.

Lex Fridman (2:32:27.960)

Your question, sorry.

Kevin Systrom (2:32:29.440)

Well, on that last, on this topic in terms of funding,

Lex Fridman (2:32:33.760)

is there, by way of advice,

Lex Fridman (2:32:39.200)

was funding in your own journey helpful, unhelpful?

Lex Fridman (2:32:44.840)

Like is there a right time to get funding, venture funding

Lex Fridman (2:32:48.360)

or anything, borrow some money from your parents?

Lex Fridman (2:32:51.120)

I don't know.

Lex Fridman (2:32:51.960)

Like is money getting in the way?

Lex Fridman (2:32:54.880)

Does it help?

Lex Fridman (2:32:56.040)

Is the timing important?

Lex Fridman (2:32:57.440)

Is there some kind of wisdom you can give there

Lex Fridman (2:33:00.280)

because you were exceptionally successful very quickly?

Lex Fridman (2:33:06.480)

Funding helps as long as it's from the right people.

Kevin Systrom (2:33:09.720)

That includes yourself.

Lex Fridman (2:33:10.640)

And I'll talk about myself funding myself in a second,

Kevin Systrom (2:33:13.640)

which is like, because I can fund myself

Lex Fridman (2:33:15.800)

doing whatever projects I can do,

Kevin Systrom (2:33:18.000)

I don't really have another person putting pressure on me

Lex Fridman (2:33:20.600)

except for myself and that creates strange dynamics, right?

Lex Fridman (2:33:24.440)

But let's like talk about people getting funding

Lex Fridman (2:33:27.080)

from a venture capitalist initially.

Kevin Systrom (2:33:30.440)

We raised money from Matt Kohler at Benchmark.

Lex Fridman (2:33:32.480)

He's brilliant, amazing guy, very thoughtful.

Lex Fridman (2:33:36.480)

And he was very helpful early on.

Lex Fridman (2:33:39.440)

But I have stories from entrepreneurs

Kevin Systrom (2:33:41.080)

where they raised money from the wrong person

Lex Fridman (2:33:42.880)

or the wrong firm where incentives weren't aligned.

Kevin Systrom (2:33:46.760)

They didn't think in the same way

Lex Fridman (2:33:48.600)

and bad things happened because of that.

Kevin Systrom (2:33:51.320)

The boardroom was always noisy.

Lex Fridman (2:33:53.480)

There were fights, like we just never had that.

Kevin Systrom (2:33:55.520)

Matt was great.

Lex Fridman (2:33:57.080)

I think like capital these days

Lex Fridman (2:33:59.880)

is kind of a dime a dozen, right?

Lex Fridman (2:34:01.720)

Like as long as you're fundable,

Kevin Systrom (2:34:03.120)

like it seems like there's money out there

Lex Fridman (2:34:05.080)

is what I'm hearing.

Kevin Systrom (2:34:08.120)

It's really important that you are aligned

Lex Fridman (2:34:10.400)

and that you think of raising money

Kevin Systrom (2:34:12.000)

as hiring someone for your team rather than taking money

Lex Fridman (2:34:15.920)

if capital is plentiful, right?

Kevin Systrom (2:34:18.320)

It provides a certain amount of pressure

Lex Fridman (2:34:20.880)

to do the right thing that I think is healthy

Kevin Systrom (2:34:23.080)

for any startup.

Lex Fridman (2:34:24.120)

And it keeps you real and honest

Kevin Systrom (2:34:25.720)

because they don't wanna lose their money.

Lex Fridman (2:34:27.400)

They're paid to not lose their money.

Kevin Systrom (2:34:29.640)

The problem, maybe I could depersonalize it,

Lex Fridman (2:34:32.720)

but like I remember having lunch with Elon.

Kevin Systrom (2:34:35.800)

It's only happened once.

Lex Fridman (2:34:37.640)

And I asked him, like I was trying to figure out

Lex Fridman (2:34:39.960)

what I was doing after Instagram, right?

Lex Fridman (2:34:42.560)

And I asked him something about like angel investing.

Lex Fridman (2:34:44.880)

And he looked at me with a straight face.

Lex Fridman (2:34:46.320)

He was like, why the F would I do that?

Lex Fridman (2:34:48.120)

Like, why?

Lex Fridman (2:34:48.960)

I was like, I don't know.

Kevin Systrom (2:34:50.760)

Like you're connected.

Lex Fridman (2:34:51.600)

Like seems like he's like, I only invest in myself.

Kevin Systrom (2:34:54.760)

I was like, Ooh, okay.

Lex Fridman (2:34:56.880)

You know, like not the confidence.

Kevin Systrom (2:34:59.240)

I was just like, what a novel idea.

Lex Fridman (2:35:00.880)

It's like, yeah, if you have money,

Kevin Systrom (2:35:03.080)

like why not just put it against your bag

Lex Fridman (2:35:06.000)

and like enable you're visiting Mars or something, right?

Kevin Systrom (2:35:11.160)

Like that's awesome, great.

Lex Fridman (2:35:12.800)

But I had never really thought of it that way.

Lex Fridman (2:35:14.920)

But also with that comes an interesting dynamic

Lex Fridman (2:35:17.600)

where you don't actually have people

Kevin Systrom (2:35:22.200)

who are gonna lose that money telling you,

Lex Fridman (2:35:23.920)

hey, don't do this or, hey, you need to face this reality.

Lex Fridman (2:35:27.000)

So you need to create other versions of that truth teller.

Lex Fridman (2:35:32.560)

And whatever I do next,

Kevin Systrom (2:35:34.560)

that's gonna be one of the interesting challenges

Lex Fridman (2:35:36.720)

is how do you create that truth telling situation?

Lex Fridman (2:35:40.640)

And that's part of why, by the way,

Lex Fridman (2:35:41.720)

I think someone like Jack, when you start Square,

Kevin Systrom (2:35:43.800)

you have money, but you still, you bring on partners

Lex Fridman (2:35:46.800)

because I think it creates

Kevin Systrom (2:35:48.680)

a truth telling type environment.

Lex Fridman (2:35:51.400)

I'm still trying to figure this out.

Kevin Systrom (2:35:52.440)

Like it's an interesting dynamic.

Lex Fridman (2:35:56.440)

So you're thinking of perhaps launching some kind of venture

Lex Fridman (2:35:59.840)

where you're investing in yourself?

Lex Fridman (2:36:01.720)

I mean, I'm 37 going on 38 next month.

Kevin Systrom (2:36:06.440)

I have a long life to live.

Lex Fridman (2:36:07.840)

I'm definitely not gonna sit on the beach, right?

Lex Fridman (2:36:10.720)

So I'm gonna do something at some point

Lex Fridman (2:36:13.520)

and I gotta imagine I will like help fund it, right?

Lex Fridman (2:36:18.200)

So the other way of thinking about this

Lex Fridman (2:36:20.040)

is you can park your money in the SMP,

Lex Fridman (2:36:21.760)

and this is bad

Lex Fridman (2:36:22.600)

because the SMP has done wonderfully well last year, right?

Kevin Systrom (2:36:26.000)

Or you can invest in yourself.

Lex Fridman (2:36:27.520)

And if you're not gonna invest in yourself,

Kevin Systrom (2:36:30.800)

you probably shouldn't do a startup.

Lex Fridman (2:36:32.120)

It's kind of the way of thinking about it.

Lex Fridman (2:36:35.800)

And you can invest in yourself in the way Elon does,

Lex Fridman (2:36:37.760)

which is basically go all in on this investment.

Kevin Systrom (2:36:41.320)

Maybe that's one way to achieve accountability

Lex Fridman (2:36:43.320)

is like you're kind of screwed if you fail.

Kevin Systrom (2:36:46.360)

Yeah, that's, yeah.

Lex Fridman (2:36:49.960)

I personally like that.

Kevin Systrom (2:36:50.800)

I like burning bridges behind me

Lex Fridman (2:36:52.640)

so that I'm fucked if it fails.

Kevin Systrom (2:36:57.360)

It's really important though.

Lex Fridman (2:37:00.480)

One of the things I think Mark said to me early on

Kevin Systrom (2:37:03.680)

that sticks with me that I think is true.

Lex Fridman (2:37:06.320)

We were talking about people

Kevin Systrom (2:37:07.560)

who had left like operating roles

Lex Fridman (2:37:10.480)

and started doing venture or something.

Kevin Systrom (2:37:11.680)

He was like, a lot of people convince themselves

Lex Fridman (2:37:13.360)

they work really hard.

Kevin Systrom (2:37:14.320)

Like they think they work really hard

Lex Fridman (2:37:15.800)

and they put on the show

Lex Fridman (2:37:17.480)

and in their minds they work really hard,

Lex Fridman (2:37:19.400)

but they don't work very hard.

Kevin Systrom (2:37:22.120)

There is something about lighting a fire underneath you

Lex Fridman (2:37:24.880)

and burning bridges such that you can't turn back.

Kevin Systrom (2:37:28.240)

That I think, we didn't talk about this specifically,

Lex Fridman (2:37:31.200)

but I think you're right.

Kevin Systrom (2:37:32.360)

There is, you need to have that

Lex Fridman (2:37:34.920)

because there's the self delusion at a certain scale.

Kevin Systrom (2:37:39.360)

Oh, I have so many board calls.

Lex Fridman (2:37:40.920)

Oh, like we have all these things to figure out.

Kevin Systrom (2:37:43.120)

It's like, this is one of the hard parts

Lex Fridman (2:37:45.960)

about it being an operator.

Kevin Systrom (2:37:47.320)

It's like, there are so many people

Lex Fridman (2:37:50.160)

that have made a lot of money not operating,

Lex Fridman (2:37:52.760)

but operating is just one of the hardest things on earth.

Lex Fridman (2:37:55.400)

It is just so effing hard.

Kevin Systrom (2:37:58.160)

It is stressful.

Lex Fridman (2:37:59.320)

It is, you're dealing with real humans,

Kevin Systrom (2:38:01.440)

not just like throwing capital in and hoping it grows.

Lex Fridman (2:38:04.800)

I'm not undermining the VC mindset.

Kevin Systrom (2:38:06.720)

I think it's a wonderful thing and needed

Lex Fridman (2:38:08.480)

and so many wonderful VCs I've worked with.

Lex Fridman (2:38:11.520)

But yeah, like when your ass is on the line

Lex Fridman (2:38:14.760)

and it's your money, it's...

Kevin Systrom (2:38:18.560)

Talk to me in 10 years, we'll see how it goes.

Lex Fridman (2:38:21.720)

Yeah, but like you were saying, that is a source.

Kevin Systrom (2:38:23.440)

When you wake up in the morning

Lex Fridman (2:38:24.680)

and you look forward to the day full of challenges,

Kevin Systrom (2:38:28.880)

that's also where you can find happiness.

Lex Fridman (2:38:31.040)

Let me ask you about love and friendship.

Kevin Systrom (2:38:32.760)

Sure.

Lex Fridman (2:38:33.600)

What's the role in this heck of a difficult journey

Lex Fridman (2:38:36.680)

you have been on of love, of friendship?

Lex Fridman (2:38:41.680)

What's the role of love in the human condition?

Kevin Systrom (2:38:45.680)

Well, first things first,

Lex Fridman (2:38:47.280)

the woman I married, my wife, Nicole,

Kevin Systrom (2:38:49.480)

there's no way I could do what I do if we weren't together.

Lex Fridman (2:38:53.840)

She had the filter idea.

Kevin Systrom (2:38:55.120)

Yeah, yeah, exactly.

Lex Fridman (2:38:56.680)

We didn't go over that story.

Lex Fridman (2:38:59.360)

Everything is a partnership, right?

Lex Fridman (2:39:01.480)

And to achieve great things,

Kevin Systrom (2:39:03.840)

it's not about like someone pulling their weight in places.

Lex Fridman (2:39:06.480)

Like it's not like someone's supporting you

Lex Fridman (2:39:08.840)

so that you could do this other thing.

Lex Fridman (2:39:11.160)

It's literally like,

Kevin Systrom (2:39:14.440)

Mike and I and our partnership as cofounders is fascinating

Lex Fridman (2:39:18.920)

because I don't think Instagram would have happened

Kevin Systrom (2:39:20.440)

without that partnership.

Lex Fridman (2:39:21.920)

Like either him or me alone, no way.

Kevin Systrom (2:39:25.800)

We pushed and pulled each other in a way

Lex Fridman (2:39:28.600)

that allowed us to build a better thing because of it.

Kevin Systrom (2:39:32.360)

Nicole, she pushed me to work on the filters early on.

Lex Fridman (2:39:35.840)

And yes, that's exciting.

Lex Fridman (2:39:36.760)

It's a fun story, right?

Lex Fridman (2:39:38.560)

But the truth of it is being able to like level

Kevin Systrom (2:39:41.920)

with someone about how hard the process is

Lex Fridman (2:39:44.720)

and have someone see you for who you are before Instagram

Lex Fridman (2:39:49.320)

and know that there's a constant you throughout all of this

Lex Fridman (2:39:53.240)

and be able to call you when you're drifting from that,

Lex Fridman (2:39:55.960)

but also support you when you're trying to stick with that.

Lex Fridman (2:39:58.360)

That's, I mean, that's true friendship slash love,

Kevin Systrom (2:40:02.920)

whatever you want to call it.

Lex Fridman (2:40:04.320)

But also it was for someone not to care.

Kevin Systrom (2:40:07.120)

I remember Nicole saying,

Lex Fridman (2:40:08.600)

hey, like I know you're going to do this Instagram thing.

Kevin Systrom (2:40:10.480)

You should, I guess it was bourbon at the time.

Lex Fridman (2:40:12.880)

You should do it because, you know,

Kevin Systrom (2:40:15.720)

even if it doesn't work,

Lex Fridman (2:40:16.840)

we can move to like a smaller apartment and it'll be fine.

Kevin Systrom (2:40:20.160)

Like we'll make it work.

Lex Fridman (2:40:22.360)

How beautiful is that, right?

Kevin Systrom (2:40:24.200)

That's almost like a superpower

Lex Fridman (2:40:25.720)

that gives you permission to fail.

Lex Fridman (2:40:27.640)

And somehow that actually leads to success.

Lex Fridman (2:40:29.760)

But also she's like the least impressed

Kevin Systrom (2:40:31.760)

about Instagram of anyone.

Lex Fridman (2:40:33.520)

She's like, yeah, it's great.

Lex Fridman (2:40:34.760)

But like, I love you for you.

Lex Fridman (2:40:36.200)

Like, I like that you're like a decent cook.

Kevin Systrom (2:40:38.000)

That's beautiful.

Lex Fridman (2:40:39.520)

That's beautiful with the Gantt chart and Thanksgiving,

Kevin Systrom (2:40:42.720)

which I still think is a brilliant effing idea.

Lex Fridman (2:40:44.880)

Thank you.

Kevin Systrom (2:40:46.240)

Big, ridiculous question.

Lex Fridman (2:40:48.280)

Have you, you're old and wise at this stage.

Lex Fridman (2:40:53.040)

So have you discovered meaning to this whole thing?

Lex Fridman (2:40:55.480)

Why the hell are we descendants of apes here on earth?

Lex Fridman (2:40:59.600)

What's the meaning of it?

Lex Fridman (2:41:00.680)

What's the meaning of life?

Kevin Systrom (2:41:01.960)

I haven't.

Lex Fridman (2:41:02.920)

And I am, so the crazy,

Lex Fridman (2:41:06.040)

so the best learning for me has been like,

Lex Fridman (2:41:09.280)

no matter what level of success you achieve,

Kevin Systrom (2:41:12.600)

you're still worried about similar things,

Lex Fridman (2:41:14.840)

just maybe on a slightly different scale.

Kevin Systrom (2:41:16.320)

You're still concerned about the same thing.

Lex Fridman (2:41:18.920)

You're still self conscious about the same things.

Kevin Systrom (2:41:21.120)

Just like, and actually that moment going through that

Lex Fridman (2:41:26.680)

is what makes you believe there's gotta be like

Kevin Systrom (2:41:29.000)

more machinery to life or purpose to life.

Lex Fridman (2:41:31.720)

And that we're all chasing these materialistic things,

Lex Fridman (2:41:35.240)

but you start realizing like,

Lex Fridman (2:41:38.200)

it's almost like, you know, the Truman Show

Kevin Systrom (2:41:39.760)

when he gets to the edge and he like knocks against it.

Lex Fridman (2:41:42.920)

He's like, what?

Kevin Systrom (2:41:43.760)

Like there's this awakening that happens

Lex Fridman (2:41:45.840)

when you get to that edge that you realize,

Kevin Systrom (2:41:47.640)

oh, like sure, it's great.

Lex Fridman (2:41:49.640)

It's great that we all chase money and fame and success.

Lex Fridman (2:41:53.760)

But you hit the edge and I'm not even claiming

Lex Fridman (2:41:56.120)

I hit an edge like Elon's hit an edge.

Kevin Systrom (2:41:58.240)

Like there's clearly larger scales.

Lex Fridman (2:42:00.000)

But what's cool is you learn that,

Kevin Systrom (2:42:02.360)

like it doesn't actually matter

Lex Fridman (2:42:03.840)

and that there are all these other things that truly matter.

Kevin Systrom (2:42:07.200)

That's not a case for working less hard.

Lex Fridman (2:42:09.280)

That's not a case for taking it easy.

Kevin Systrom (2:42:11.480)

That's not a case for the four day work week.

Lex Fridman (2:42:13.920)

What that is a case for is designing your life

Kevin Systrom (2:42:16.640)

exactly the way you want to design it.

Lex Fridman (2:42:18.960)

Cause I don't know, I think we go around the earth,

Kevin Systrom (2:42:22.440)

you know, the sun a certain number of times

Lex Fridman (2:42:25.040)

and then we die and then that's it.

Kevin Systrom (2:42:27.640)

That's me.

Lex Fridman (2:42:28.520)

Are you afraid of that moment?

Kevin Systrom (2:42:30.200)

No, not at all.

Lex Fridman (2:42:31.040)

In fact, or at least not yet.

Kevin Systrom (2:42:36.200)

Listen, I'm like a pilot, like I do crazy things

Lex Fridman (2:42:39.560)

and I like, no, I like, if anything, I'm like,

Kevin Systrom (2:42:43.440)

oh, I got to choose mindfully and purposefully

Lex Fridman (2:42:49.640)

the thing I am doing right now and not just fall into it

Kevin Systrom (2:42:54.080)

because you're going to wake up one day and ask yourself

Lex Fridman (2:42:55.800)

why the hell you spent the last 10 years doing X, Y or Z?

Kevin Systrom (2:42:58.400)

Yeah.

Lex Fridman (2:42:59.240)

So I guess my like shorter answer to this is

Kevin Systrom (2:43:03.720)

doing things on purpose because you choose to do them.

Lex Fridman (2:43:08.320)

So important in life and not just like floating

Kevin Systrom (2:43:11.640)

down the river of life, hitting branches along the way

Lex Fridman (2:43:14.680)

cause you will hit branches, right?

Lex Fridman (2:43:17.160)

But rather like literally plotting a course

Lex Fridman (2:43:19.680)

and not having a 10 year plan,

Lex Fridman (2:43:21.080)

but just choosing every day to opt in.

Lex Fridman (2:43:23.740)

That I think has been more like,

Kevin Systrom (2:43:28.540)

I haven't figured out the meaning of life

Lex Fridman (2:43:29.900)

by any stretch of the imagination,

Lex Fridman (2:43:31.620)

but it certainly isn't money and it certainly isn't fame

Lex Fridman (2:43:33.820)

and it certainly isn't travel.

Lex Fridman (2:43:35.080)

And it's like, and it's way more of like opting

Lex Fridman (2:43:37.740)

into the game you love playing.

Kevin Systrom (2:43:40.260)

Every day opting in.

Lex Fridman (2:43:41.980)

Just opting in and like, don't let it happen.

Kevin Systrom (2:43:44.820)

You opt in.

Lex Fridman (2:43:46.920)

Kevin, it's great to end on love and the meaning of life.

Kevin Systrom (2:43:51.920)

This was an amazing conversation.

Lex Fridman (2:43:53.920)

It was a lot of fun, thank you.

Kevin Systrom (2:43:54.760)

You gave me like a light into some fascinating aspects

Lex Fridman (2:43:57.880)

of this technical world.

Lex Fridman (2:44:00.160)

And I can't honestly wait to see what you do next.

Lex Fridman (2:44:04.040)

Thank you so much.

Kevin Systrom (2:44:05.280)

Thanks for having me.

Lex Fridman (2:44:07.380)

Thanks for listening to this conversation

Kevin Systrom (2:44:09.040)

with Kevin Systrom.

Lex Fridman (2:44:10.480)

To support this podcast, please check out our sponsors

Kevin Systrom (2:44:13.280)

in the description.

Lex Fridman (2:44:14.640)

And now let me leave you with some words

Kevin Systrom (2:44:16.880)

from Kevin Systrom himself.

Lex Fridman (2:44:19.520)

Focusing on one thing and doing it really, really well

Kevin Systrom (2:44:24.240)

can get you very far.

Lex Fridman (2:44:25.620)

Thank you for listening and hope to see you next time.

Lex Fridman (30:00.340)

But the fact that you could connect with people

Lex Fridman (30:02.140)

who like took really beautiful photos in a certain style

Kevin Systrom (30:05.300)

all around the world, whether they were travelers,

Lex Fridman (30:07.320)

it was the beginning of the influencer economy.

Kevin Systrom (30:11.380)

It was these people

Lex Fridman (30:12.220)

who became professional Instagrammers way back when.

Lex Fridman (30:16.660)

But they took these amazing photos

Lex Fridman (30:18.800)

and some of them were photographers professionally.

Lex Fridman (30:24.100)

And all of a sudden you had this moment in the day

Lex Fridman (30:25.900)

when you could open up this app

Lex Fridman (30:27.140)

and sure you could see what your friends were doing,

Lex Fridman (30:28.900)

but also it was like, oh my God,

Kevin Systrom (30:30.320)

that's a beautiful waterfall or oh my God,

Lex Fridman (30:33.080)

I didn't realize there was that corner of England

Kevin Systrom (30:35.100)

or like really cool stuff.

Lex Fridman (30:38.700)

And the beauty about Instagram early on

Kevin Systrom (30:40.280)

was that it was international by default.

Lex Fridman (30:43.940)

You didn't have to speak English to use it, right?

Kevin Systrom (30:46.740)

You could just look at the photos, works great.

Lex Fridman (30:49.780)

We did translate, we had some pretty bad translations,

Lex Fridman (30:52.360)

but we did translate the app.

Lex Fridman (30:54.420)

And even if our translations were pretty poor,

Kevin Systrom (30:57.860)

the idea that you could just connect with other people

Lex Fridman (31:00.100)

through their images was pretty powerful.

Lex Fridman (31:02.280)

So how much technical difficulties

Lex Fridman (31:05.760)

there with the programming?

Lex Fridman (31:07.120)

Like what programming language you were talking about?

Lex Fridman (31:09.120)

What was?

Kevin Systrom (31:09.960)

Zero, like maybe it was hard for us,

Lex Fridman (31:12.040)

but I mean, there was nothing.

Kevin Systrom (31:16.280)

The only thing that was complex about Instagram

Lex Fridman (31:18.700)

at the beginning technically was making it scale.

Kevin Systrom (31:23.680)

We were just plain old Objective C for the client.

Lex Fridman (31:27.920)

So it was iPhone only at first?

Kevin Systrom (31:29.480)

iPhone only, yep.

Lex Fridman (31:30.880)

As an Android person, I'm deeply offended, but go ahead.

Kevin Systrom (31:34.080)

This was 2010.

Lex Fridman (31:35.880)

Oh, sure, sure.

Kevin Systrom (31:36.720)

Sorry, sorry.

Lex Fridman (31:37.560)

Android's getting a lot better, right?

Kevin Systrom (31:39.680)

So.

Lex Fridman (31:40.960)

I take it back, you're right.

Kevin Systrom (31:41.840)

If I were to do something today,

Lex Fridman (31:43.160)

I think it would be very different

Lex Fridman (31:44.440)

in terms of launch strategy, right?

Lex Fridman (31:46.680)

Android's enormous too.

Lex Fridman (31:49.180)

But anyway, back to that moment,

Lex Fridman (31:51.160)

it was Objective C and then we were Python based,

Kevin Systrom (31:56.920)

which is just like, this is before Python was really cool.

Lex Fridman (32:00.220)

Like now it's cool

Kevin Systrom (32:01.200)

because it's all these machine learning libraries

Lex Fridman (32:03.080)

like support Python and right.

Kevin Systrom (32:05.240)

Now it's super, now it's like cool to be Python.

Lex Fridman (32:08.120)

Back then it was like, oh, Google uses Python.

Kevin Systrom (32:10.800)

Like maybe you should use Python.

Lex Fridman (32:12.600)

Facebook was PHP.

Kevin Systrom (32:14.360)

Like I had worked at a small startup

Lex Fridman (32:16.760)

of some ex Googlers that used Python.

Lex Fridman (32:19.320)

So we used it and we used a framework called Django,

Lex Fridman (32:22.480)

still exists and people use for basically the backend.

Lex Fridman (32:27.260)

And then you threw a couple interesting things in there.

Lex Fridman (32:30.120)

I mean, we used Postgres, which was kind of fun.

Kevin Systrom (32:32.160)

It was a little bit like Hipster database

Lex Fridman (32:34.120)

at the time, right?

Kevin Systrom (32:34.960)

Versus MySQL.

Lex Fridman (32:36.160)

MySQL, like everyone used MySQL.

Lex Fridman (32:37.920)

So like using Postgres was like an interesting decision.

Lex Fridman (32:40.520)

Right?

Lex Fridman (32:42.080)

But we used it because it had a bunch of geo features

Lex Fridman (32:45.320)

built in because we thought we were gonna be

Lex Fridman (32:46.840)

a check and app, remember?

Lex Fridman (32:47.880)

It's also super cool now.

Lex Fridman (32:49.280)

So you were into Python before it was cool

Lex Fridman (32:51.480)

and you were into Postgres before it was cool.

Kevin Systrom (32:53.480)

Yeah, we were basically like,

Lex Fridman (32:55.060)

not only Hipster photo company,

Lex Fridman (32:57.960)

Hipster tech company, right?

Lex Fridman (33:00.100)

We also adopted Redis early and like loved it.

Kevin Systrom (33:04.000)

I mean, it solves so many problems for us

Lex Fridman (33:06.960)

and turns out that's still pretty cool.

Lex Fridman (33:09.580)

But the programming was very easy.

Lex Fridman (33:11.260)

It was like, sign up a user, have a feed.

Kevin Systrom (33:13.920)

There was nothing, no machine learning at all, zero.

Lex Fridman (33:17.440)

Can you get some context?

Lex Fridman (33:18.400)

How many users at each of these stages?

Lex Fridman (33:20.720)

Are we talking about a hundred users, a thousand users?

Lex Fridman (33:24.100)

So the stage I just described,

Lex Fridman (33:25.640)

I mean that technical stack lasted

Kevin Systrom (33:27.900)

through probably 50 million users.

Lex Fridman (33:32.720)

I mean, seriously, like you can get away with a lot

Kevin Systrom (33:36.120)

with a pretty basic stack.

Lex Fridman (33:38.360)

Like I think a lot of startups try to overengineer

Kevin Systrom (33:41.060)

their solutions from the beginning to like really scale

Lex Fridman (33:43.640)

and you can get away with a lot.

Kevin Systrom (33:45.640)

That being said, most of the first two years of Instagram

Lex Fridman (33:48.560)

was literally just trying to make that stack scale.

Lex Fridman (33:51.280)

And it wasn't, it was not a Python problem.

Lex Fridman (33:54.120)

It was like literally just like, where do we put the data?

Kevin Systrom (33:58.040)

Like it's all coming in too fast.

Lex Fridman (33:59.920)

Like how do we store it?

Lex Fridman (34:01.000)

How do we make sure to be up?

Lex Fridman (34:02.360)

How do we like, how do we make sure we're

Lex Fridman (34:05.840)

on the right size boxes that they have enough memory?

Lex Fridman (34:09.320)

Those were the issues, but.

Lex Fridman (34:11.200)

Can you speak to the choices you make at that stage

Lex Fridman (34:14.280)

when you're growing so quickly?

Lex Fridman (34:16.200)

Do you use something like somebody else's

Lex Fridman (34:18.480)

computer infrastructure or do you build in house?

Kevin Systrom (34:22.640)

I'm only laughing because we, when we launched

Lex Fridman (34:25.640)

we had a single computer that we had rented

Kevin Systrom (34:30.700)

in some color space in LA.

Lex Fridman (34:32.160)

I don't even remember what it was called.

Kevin Systrom (34:34.720)

Cause I thought that's what you did.

Lex Fridman (34:35.940)

When I worked at a company called Odio that became Twitter.

Kevin Systrom (34:38.520)

I remember visiting our space in San Francisco.

Lex Fridman (34:41.200)

You walked in, you had to wear the ear things.

Lex Fridman (34:43.480)

It was cold and fans everywhere, right?

Lex Fridman (34:46.880)

And we had to, you know, plug one out, replace one.

Lex Fridman (34:49.940)

And I was the intern, so I just like held things.

Lex Fridman (34:52.720)

But I thought to myself, oh, this is how it goes.

Lex Fridman (34:54.960)

And then I remember being in a VC's office.

Lex Fridman (34:58.480)

I think it was a benchmarks office.

Lex Fridman (35:00.040)

And I think we ran into another entrepreneur

Lex Fridman (35:02.280)

and they were like, oh, how are things going?

Kevin Systrom (35:03.560)

We're like, ah, you know, try to scale this thing.

Lex Fridman (35:06.840)

And they were like, well, I mean

Lex Fridman (35:08.420)

can't you just add more instances?

Lex Fridman (35:09.960)

And I was like, what do you mean?

Lex Fridman (35:11.540)

And they're like instances on Amazon.

Lex Fridman (35:13.560)

I was like, what are those?

Lex Fridman (35:15.520)

And it was this moment where we realized how deep

Lex Fridman (35:18.240)

in it we were because we had no idea that AWS existed

Kevin Systrom (35:22.980)

nor should we be using it.

Lex Fridman (35:24.420)

Anyway, that night we went back to the office

Lex Fridman (35:27.000)

and we got on AWS, but we did this really dumb thing.

Lex Fridman (35:29.840)

We're so sorry to people listening.

Lex Fridman (35:32.960)

But we brought up an instance, which was our database.

Lex Fridman (35:38.300)

It's gonna be a replacement for our database.

Lex Fridman (35:40.920)

But we had it talking over the public internet

Lex Fridman (35:43.720)

to our little box in LA that was our app server.

Kevin Systrom (35:46.560)

Very nice. Yeah.

Lex Fridman (35:48.380)

That's how sophisticated we were.

Lex Fridman (35:49.960)

And obviously that was very, very slow.

Lex Fridman (35:52.840)

Didn't work at all.

Kevin Systrom (35:53.840)

I mean, it worked, but didn't work.

Lex Fridman (35:55.960)

Only like later that night did we realize

Kevin Systrom (35:58.080)

we had to have it all together.

Lex Fridman (36:00.160)

But at least like if you're listening right now

Lex Fridman (36:02.120)

and you're thinking, you know, I have no chance.

Lex Fridman (36:04.740)

I'm gonna start to start, but I have no chance.

Kevin Systrom (36:06.920)

I don't know.

Lex Fridman (36:07.760)

We did it and we made a bunch

Kevin Systrom (36:08.580)

of really dumb mistakes initially.

Lex Fridman (36:10.640)

I think the question is how quickly do you learn

Lex Fridman (36:12.540)

that you're making a mistake?

Lex Fridman (36:13.560)

And do you do the right thing immediately right after?

Lex Fridman (36:16.360)

So you didn't pay for those mistakes by failure.

Lex Fridman (36:20.320)

So yeah, how quickly did you fix it?

Kevin Systrom (36:23.840)

I guess there's a lot of ways to sneak up to this question

Lex Fridman (36:26.280)

of how the hell do you scale the thing?

Kevin Systrom (36:28.440)

Other startups, if you have an idea,

Lex Fridman (36:30.360)

how do you scale the thing?

Kevin Systrom (36:31.640)

Is it just AWS and you try to write the kind of code

Lex Fridman (36:37.680)

that's easy to spread across a large number of instances,

Lex Fridman (36:41.820)

and then the rest is just put money into it?

Lex Fridman (36:45.180)

Basically, I would say a couple of things.

Kevin Systrom (36:48.860)

First off, don't even ask the question,

Lex Fridman (36:51.580)

just find product market fit, duct tape it together, right?

Kevin Systrom (36:55.460)

Like if you have to.

Lex Fridman (36:56.620)

I think there's a big caveat here, which I want to get to,

Lex Fridman (37:00.400)

but generally all that matters is product market fit.

Lex Fridman (37:03.800)

That's all that matters.

Kevin Systrom (37:04.860)

If people like your product,

Lex Fridman (37:06.880)

do not worry about when 50,000 people use your product

Kevin Systrom (37:10.300)

because you will be happy that you have that problem

Lex Fridman (37:12.760)

when you get there.

Kevin Systrom (37:13.920)

I actually can't name many startups

Lex Fridman (37:18.660)

where they go from nothing to something overnight

Lex Fridman (37:22.100)

and they can't figure out how to scale it.

Lex Fridman (37:24.260)

There are some, but I think nowadays,

Kevin Systrom (37:27.620)

it's a, when I say a solved problem,

Lex Fridman (37:29.640)

like there are ways of solving it.

Kevin Systrom (37:33.260)

The base case is typically that startups

Lex Fridman (37:35.060)

worry way too much about scaling way too early

Lex Fridman (37:38.000)

and forget that they actually have to make something

Lex Fridman (37:39.800)

that people like.

Kevin Systrom (37:40.640)

That's the default mistake case.

Lex Fridman (37:43.740)

But what I'll say is once you start scaling,

Kevin Systrom (37:48.020)

I mean, hiring quickly,

Lex Fridman (37:49.820)

people who have seen the game before

Lex Fridman (37:51.460)

and just know how to do it,

Lex Fridman (37:52.980)

it becomes a bit of like, yeah,

Lex Fridman (37:56.660)

just throw instances of the problem, right?

Lex Fridman (37:59.340)

But the last thing I'll say on this

Kevin Systrom (38:00.740)

that I think did save us,

Lex Fridman (38:03.420)

we were pretty rigorous about writing tests

Kevin Systrom (38:06.700)

from the beginning.

Lex Fridman (38:08.560)

That helped us move very, very quickly

Kevin Systrom (38:11.420)

when we wanted to rewrite parts of the product

Lex Fridman (38:14.620)

and know that we weren't breaking something else.

Kevin Systrom (38:17.260)

Tests are one of those things where it's like,

Lex Fridman (38:18.700)

you go slow to go fast.

Lex Fridman (38:20.660)

And they suck when you have to write them

Lex Fridman (38:22.940)

because you have to figure it out.

Lex Fridman (38:24.180)

And there are always those ones that break

Lex Fridman (38:27.000)

when you don't want them to break and they're annoying

Lex Fridman (38:29.060)

and it feels like you spent all this time.

Lex Fridman (38:30.420)

But looking back, I think that like longterm optimal,

Kevin Systrom (38:34.740)

even with a team of four,

Lex Fridman (38:36.900)

it allowed us to move very, very quickly

Kevin Systrom (38:38.780)

because anyone could touch any part of the product

Lex Fridman (38:41.940)

and know that they weren't gonna bring down the site,

Kevin Systrom (38:44.440)

or at least in general.

Lex Fridman (38:45.900)

At which point do you know product market fit?

Lex Fridman (38:48.660)

How many users would you say?

Lex Fridman (38:50.540)

Is it all it takes is like 10 people?

Lex Fridman (38:52.540)

Or is it a thousand?

Lex Fridman (38:53.780)

Is it 50,000?

Kevin Systrom (38:55.820)

I don't think it is generally

Lex Fridman (38:58.220)

a question of absolute numbers.

Kevin Systrom (38:59.880)

I think it's a question of cohorts

Lex Fridman (39:01.860)

and I think it's a question of trends.

Kevin Systrom (39:03.900)

So, you know, it depends how big your business is trying

Lex Fridman (39:08.780)

to be, right?

Lex Fridman (39:09.620)

But if I were signing up a thousand people a week

Lex Fridman (39:12.260)

and they all retain,

Kevin Systrom (39:14.100)

like the retention curves for those cohorts looked good,

Lex Fridman (39:16.980)

healthy, and even like,

Kevin Systrom (39:19.740)

as you started getting more people on the service,

Lex Fridman (39:21.940)

maybe those earlier cohorts started curving up again

Kevin Systrom (39:24.660)

because now there are network effects

Lex Fridman (39:26.060)

and their friends are on the service

Kevin Systrom (39:27.380)

or totally depends what type of business you're in,

Lex Fridman (39:29.860)

but I'm talking purely social, right?

Kevin Systrom (39:34.060)

I don't think it's an absolute number.

Lex Fridman (39:36.580)

I think it is,

Kevin Systrom (39:37.900)

I guess you could call it a marginal number.

Lex Fridman (39:39.660)

So I spent a lot of time when I work with startups

Kevin Systrom (39:42.340)

asking them like, okay,

Lex Fridman (39:44.380)

have you looked at that cohort versus this cohort,

Kevin Systrom (39:47.260)

whether it's your clients

Lex Fridman (39:48.500)

or whether it's people signing up for the service?

Lex Fridman (39:52.300)

But a lot of people think you just have to hit some mark,

Lex Fridman (39:55.180)

like 10,000 people or 50,000 people,

Lex Fridman (39:57.780)

but really seven ish billion people in the world.

Lex Fridman (40:01.980)

Most people forever will not know about your product.

Kevin Systrom (40:05.300)

There are always more people out there to sign up.

Lex Fridman (40:07.860)

It's just a question of how you turn on the spigot, so.

Kevin Systrom (40:11.380)

At that stage, early stage yourself,

Lex Fridman (40:14.940)

but also by way of advice,

Lex Fridman (40:16.700)

should you worry about money at all?

Lex Fridman (40:18.220)

How this thing's gonna make money?

Kevin Systrom (40:20.340)

Or do you just try to find product market fit

Lex Fridman (40:24.300)

and get a lot of users to enjoy using your thing?

Kevin Systrom (40:28.660)

I think it totally depends.

Lex Fridman (40:30.020)

And that's an unsatisfying answer.

Kevin Systrom (40:32.620)

I was talking with a friend today who,

Lex Fridman (40:36.460)

he was one of our earlier investors and he was saying,

Lex Fridman (40:39.100)

hey, like, have you been doing any angel investing lately?

Lex Fridman (40:41.580)

I said, not really.

Kevin Systrom (40:42.540)

I'm just like focused on what I wanna do next.

Lex Fridman (40:44.780)

And he said, the number of financings have just gone bonkers.

Kevin Systrom (40:49.780)

Like people are throwing money everywhere right now.

Lex Fridman (40:56.340)

And I think the question is,

Lex Fridman (41:00.580)

do you have an inkling of how you're gonna make money?

Lex Fridman (41:04.140)

Or are you really just like waving your hands?

Kevin Systrom (41:07.420)

I would not like to be an entrepreneur in the position of,

Lex Fridman (41:11.460)

well, I have no idea how this will eventually make money.

Kevin Systrom (41:14.420)

That's not fun.

Lex Fridman (41:16.980)

If you are in an area,

Lex Fridman (41:18.740)

like let's say you wanted to start a social network, right?

Lex Fridman (41:22.900)

Not saying this is a good idea, but if you did,

Kevin Systrom (41:25.900)

there are only a handful of ways they've made money

Lex Fridman (41:27.780)

and really only one way they've made money in the past

Lex Fridman (41:29.820)

and that's ads.

Lex Fridman (41:31.780)

So, if you have a service that's amenable to that

Lex Fridman (41:37.300)

and then I wouldn't worry too much about that

Lex Fridman (41:39.780)

because if you get to the scale,

Kevin Systrom (41:40.900)

you can hire some smart people and figure that out.

Lex Fridman (41:44.220)

I do think that is really healthy for a lot of startups

Kevin Systrom (41:48.620)

these days, especially the ones doing

Lex Fridman (41:50.260)

like enterprise software, slacks of the world, et cetera,

Kevin Systrom (41:54.820)

to be worried about money from the beginning,

Lex Fridman (41:56.660)

but mostly as a way of winning over clients

Lex Fridman (42:00.180)

and having stickiness.

Lex Fridman (42:05.500)

Of course you need to be worried about money,

Lex Fridman (42:06.940)

but I'm gonna also say this again,

Lex Fridman (42:08.620)

which is it's like longterm profitability.

Kevin Systrom (42:12.100)

If you have a roadmap to that, then that's great.

Lex Fridman (42:15.860)

But if you're just like, I don't know, maybe never,

Kevin Systrom (42:18.100)

like we're working on this metaverse thing,

Lex Fridman (42:19.860)

I think maybe someday, I don't know.

Kevin Systrom (42:22.900)

Like that seems harder to me.

Lex Fridman (42:24.700)

So you have to be as big as Facebook

Lex Fridman (42:26.740)

to like finance that bet, right?

Lex Fridman (42:29.420)

Do you think it's possible, you said,

Kevin Systrom (42:31.180)

you're not saying it's necessarily a good idea

Lex Fridman (42:33.380)

to launch a social network.

Lex Fridman (42:35.260)

Do you think it's possible today,

Lex Fridman (42:38.980)

maybe you can put yourself in those shoes,

Kevin Systrom (42:41.260)

to launch a social network that achieves the scale

Lex Fridman (42:45.420)

of a Facebook or a Twitter or an Instagram,

Lex Fridman (42:49.500)

and maybe even greater scale?

Lex Fridman (42:51.300)

Absolutely.

Lex Fridman (42:53.300)

How do you do it?

Lex Fridman (42:55.460)

Asking for a friend.

Kevin Systrom (42:56.380)

Yeah, if I knew, I'd probably be doing it right now

Lex Fridman (42:59.020)

and not sitting here.

Kevin Systrom (43:00.940)

So, I mean, there's a lot of ways to ask this question.

Lex Fridman (43:03.580)

One is create a totally new product market fit,

Kevin Systrom (43:07.220)

create a new market, create something like Instagram did,

Lex Fridman (43:10.300)

which is like create something kind of new,

Kevin Systrom (43:13.100)

or literally out compete Facebook at its own thing,

Lex Fridman (43:17.060)

or I'll compete Twitter at its own thing.

Kevin Systrom (43:19.820)

The only way to compete now,

Lex Fridman (43:21.500)

if you wanna build a large social network

Kevin Systrom (43:23.780)

is to look for the cracks, look for the openings.

Lex Fridman (43:28.980)

No one competed, I mean,

Kevin Systrom (43:31.340)

no one competed with the core business of Google.

Lex Fridman (43:33.420)

No one competed with the core business of Microsoft.

Kevin Systrom (43:36.860)

You don't go at the big guys

Lex Fridman (43:39.380)

doing exactly what they're doing.

Kevin Systrom (43:41.860)

Instagram didn't win, quote unquote,

Lex Fridman (43:43.900)

because it tried to be a visual Twitter.

Kevin Systrom (43:46.940)

Like we spotted things that either Twitter

Lex Fridman (43:49.980)

wasn't going to do or refuse to do,

Lex Fridman (43:52.820)

images and feed for the longest time, right?

Lex Fridman (43:55.980)

Or that Facebook wasn't doing or not paying attention to

Kevin Systrom (43:58.540)

because they were mostly desktop at the time

Lex Fridman (44:00.620)

and we were purely mobile, purely visual.

Kevin Systrom (44:05.060)

Often there are opportunities sitting there.

Lex Fridman (44:07.620)

You have to figure out like,

Kevin Systrom (44:12.020)

I think like there's a strategy book,

Lex Fridman (44:13.860)

I can't remember the name,

Lex Fridman (44:14.780)

but talk about moats and just like the best place to play

Lex Fridman (44:19.700)

is where your competitor literally can't pivot

Kevin Systrom (44:22.220)

because structurally they're set up not to be there.

Lex Fridman (44:26.060)

And that's where you win.

Lex Fridman (44:28.100)

And what's fascinating is like,

Lex Fridman (44:30.460)

do you know how many people were like,

Kevin Systrom (44:31.980)

images, Facebook does that, Twitter does that.

Lex Fridman (44:35.260)

I mean, how wrong were they, really wrong?

Lex Fridman (44:37.220)

And these are some of the smartest people

Lex Fridman (44:38.580)

in Silicon Valley, right?

Lex Fridman (44:40.420)

But now Instagram exists for a while.

Lex Fridman (44:42.740)

How is it that Snapchat could then exist?

Kevin Systrom (44:45.220)

It makes no sense.

Lex Fridman (44:47.300)

Like plenty of people would say,

Kevin Systrom (44:48.720)

well, there's Facebook, no images.

Lex Fridman (44:50.140)

Okay, okay, Instagram, I'll give you that one.

Lex Fridman (44:52.820)

But wait, now another image based social network

Lex Fridman (44:55.260)

is gonna get really big.

Lex Fridman (44:57.140)

And then TikTok comes along.

Lex Fridman (44:59.980)

Like the prior, so you asked me, is it possible?

Kevin Systrom (45:03.500)

The only reason I'm answering yes

Lex Fridman (45:05.700)

is because my prior is that it's happened once every,

Kevin Systrom (45:09.620)

I don't know, three, four or five years consistently.

Lex Fridman (45:13.020)

And I can't imagine there's anything structurally

Kevin Systrom (45:15.340)

that would change that.

Lex Fridman (45:17.620)

So that's why I answer that way.

Kevin Systrom (45:18.980)

Not because I know how, I just,

Lex Fridman (45:21.260)

when you see a pattern, you see a pattern

Lex Fridman (45:22.900)

and there's no reason to believe that's gonna stop.

Lex Fridman (45:25.580)

And it's subtle too, because like you said,

Kevin Systrom (45:27.780)

Snapchat and TikTok,

Lex Fridman (45:29.720)

they're all doing the same space of things,

Lex Fridman (45:32.660)

but there's something fundamentally different

Lex Fridman (45:34.820)

about like a three second video and a five second video

Lex Fridman (45:38.740)

and a 15 second video and a one minute video

Lex Fridman (45:41.020)

and a one hour video, like fundamentally different.

Kevin Systrom (45:43.900)

Fundamentally different.

Lex Fridman (45:45.020)

I mean, I think one of the reasons Snapchat exists

Kevin Systrom (45:48.260)

is because Instagram was so focused on posting great,

Lex Fridman (45:51.660)

beautiful manicured versions of yourself throughout time.

Lex Fridman (45:57.180)

And there was this enormous demand of like,

Lex Fridman (45:58.780)

hey, I really like this behavior.

Kevin Systrom (46:00.740)

I love using Instagram, but man,

Lex Fridman (46:03.400)

I just like wish I could share something going on in my day.

Lex Fridman (46:07.940)

Do I really have to put it on my profile?

Lex Fridman (46:10.000)

Do I really have to make it last forever?

Kevin Systrom (46:11.900)

Do I really, and that opened up a door,

Lex Fridman (46:14.680)

it created a market, right?

Lex Fridman (46:16.660)

And then what's fascinating is Instagram had an explore page

Lex Fridman (46:20.460)

for the longest time and it was image driven, right?

Lex Fridman (46:23.740)

But there's absolutely a behavior where you open up Instagram

Lex Fridman (46:26.780)

and you sit on the explore page all day.

Kevin Systrom (46:28.340)

That is effectively TikTok,

Lex Fridman (46:29.900)

but obviously focused on videos.

Lex Fridman (46:32.540)

And it's not like you could just put the explore page

Lex Fridman (46:35.420)

in TikTok form and it works.

Kevin Systrom (46:37.140)

It had to be video, it had to have music.

Lex Fridman (46:39.740)

These are the hard parts about product development

Kevin Systrom (46:42.380)

that are very hard to predict,

Lex Fridman (46:44.180)

but they're all versions of the same thing

Kevin Systrom (46:47.940)

with varying, like if you line them up

Lex Fridman (46:50.980)

in a bunch of dimensions, they're just like kind of on,

Kevin Systrom (46:55.180)

they're different values of the same dimensions,

Lex Fridman (46:56.980)

which is like, I guess, easy to say in retrospect.

Lex Fridman (46:59.680)

But like, if I were an entrepreneur going after that area,

Lex Fridman (47:02.580)

I'd ask myself like, where's the opening?

Lex Fridman (47:05.140)

What needs to exist because TikTok exists now?

Lex Fridman (47:08.500)

So I wonder how much things that don't yet exist

Lex Fridman (47:13.060)

and can exist is in the space of algorithms,

Lex Fridman (47:15.880)

in the space of recommender systems.

Lex Fridman (47:18.020)

So in the space of how the feed is generated.

Lex Fridman (47:21.680)

So we kind of talk about the actual elements

Kevin Systrom (47:24.180)

of the content, that's what we've been talking,

Lex Fridman (47:27.900)

the difference between photos,

Kevin Systrom (47:29.840)

between short videos, longer videos.

Lex Fridman (47:32.440)

I wonder how much disruption is possible

Kevin Systrom (47:35.140)

in the way the algorithms work.

Lex Fridman (47:37.260)

Because a lot of the criticism towards social media

Kevin Systrom (47:39.540)

is in the way the algorithms work currently.

Lex Fridman (47:41.900)

And it feels like, first of all,

Kevin Systrom (47:44.620)

talking about product market fit,

Lex Fridman (47:47.340)

there's certainly a hunger for social media algorithms

Kevin Systrom (47:54.420)

that do something different.

Lex Fridman (47:56.860)

I don't think anyone, everyone said complaining,

Kevin Systrom (47:59.580)

this is hurting me and this is hurting society,

Lex Fridman (48:04.180)

but I keep doing it because I'm addicted to it.

Lex Fridman (48:07.380)

And they say, we want something different,

Lex Fridman (48:09.940)

but we don't know what.

Kevin Systrom (48:11.380)

It feels like just different.

Lex Fridman (48:15.300)

It feels like there's a hunger for that,

Lex Fridman (48:17.740)

but that's in the space of algorithms.

Lex Fridman (48:19.300)

I wonder if it's possible to disrupt in that space.

Kevin Systrom (48:22.020)

Absolutely, I have this thesis that the worst part

Lex Fridman (48:27.860)

about social networks is that they're, is the people.

Lex Fridman (48:33.740)

It's a line that sounds funny, right?

Lex Fridman (48:36.660)

Because like, that's why you call it a social network.

Lex Fridman (48:39.660)

But what does social networks actually do for you?

Lex Fridman (48:42.420)

Like just think, like imagine you were an alien

Lex Fridman (48:45.740)

and you landed and someone says,

Lex Fridman (48:47.460)

hey, there's this site, it's a social network.

Kevin Systrom (48:49.460)

We're not gonna tell you what it is,

Lex Fridman (48:50.400)

but just what does it do?

Lex Fridman (48:51.860)

And you have to explain it to them.

Lex Fridman (48:53.140)

It does two things.

Kevin Systrom (48:54.340)

One is that people you know and have social ties with

Lex Fridman (49:00.220)

distribute updates through whether it's photos or videos

Kevin Systrom (49:05.940)

about their lives so that you don't have to physically

Lex Fridman (49:07.820)

be with them, but you can keep in touch with them.

Kevin Systrom (49:09.860)

That's one, that's like a big part of Instagram.

Lex Fridman (49:12.260)

That's a big part of Snap.

Kevin Systrom (49:14.020)

It is not part of TikTok at all.

Lex Fridman (49:16.300)

So there's another big part, which is there's all this

Kevin Systrom (49:19.700)

content out in the world that's entertaining,

Lex Fridman (49:22.420)

whether you wanna watch it or you wanna read it.

Lex Fridman (49:26.860)

And matchmaking between content that exists in the world

Lex Fridman (49:30.340)

and people that want that content

Lex Fridman (49:33.460)

turns out to be like a really big business, right?

Lex Fridman (49:35.980)

Search and discovery, which you?

Kevin Systrom (49:37.260)

Search and discovery, but my point is it could be video,

Lex Fridman (49:39.940)

it could be text, it could be websites, it could be,

Lex Fridman (49:42.020)

I mean, think back to like dig, right?

Lex Fridman (49:46.580)

Or stumble upon or, right?

Kevin Systrom (49:49.260)

Nice, yeah.

Lex Fridman (49:50.620)

But like, what did those do?

Kevin Systrom (49:51.820)

Like they basically distributed interesting content

Lex Fridman (49:54.340)

to you, right?

Kevin Systrom (49:57.500)

I think the most interesting part or the future

Lex Fridman (50:00.740)

of social networks is going to be making them less social

Kevin Systrom (50:03.540)

because I think people are part of the root cause

Lex Fridman (50:06.340)

of the problem.

Lex Fridman (50:07.180)

So for instance, often in recommender systems,

Lex Fridman (50:10.340)

we talk about two stages.

Kevin Systrom (50:11.500)

There's a candidate generation step, which is just like

Lex Fridman (50:14.500)

of our vast trove of stuff that you might wanna see,

Lex Fridman (50:18.380)

what small subset should we pick for you, okay?

Lex Fridman (50:24.020)

Typically that is grabbed from things

Lex Fridman (50:25.900)

your friends have shared, right?

Lex Fridman (50:28.380)

Then there's a ranking step which says, okay,

Kevin Systrom (50:30.820)

now given these 100, 200 things depends on the network,

Lex Fridman (50:33.580)

right?

Kevin Systrom (50:34.420)

Let's like be really good about ranking them

Lex Fridman (50:36.460)

and generally rank the things up higher

Lex Fridman (50:38.940)

that get the most engagement, right?

Lex Fridman (50:40.820)

So what's the problem with that?

Kevin Systrom (50:42.540)

Step one is we've limited everything you could possibly see

Lex Fridman (50:46.860)

to things that your friends have chosen to share

Kevin Systrom (50:49.540)

or maybe not friends, but influencers.

Lex Fridman (50:52.300)

What things do people generally want to share?

Kevin Systrom (50:54.580)

They wanna share things that are gonna get likes,

Lex Fridman (50:56.340)

that are gonna show up broadly.

Lex Fridman (50:58.660)

So they tend to be more emotionally driven.

Lex Fridman (51:00.740)

They tend to be more risque or whatever.

Lex Fridman (51:03.820)

So why do we have this problem?

Lex Fridman (51:05.140)

It's because we show people things people have decided

Kevin Systrom (51:08.980)

to share and those things self select to being the things

Lex Fridman (51:11.660)

that are most divisive.

Lex Fridman (51:14.060)

So how do you fix that?

Lex Fridman (51:15.540)

Well, what if you just imagine for a second

Kevin Systrom (51:20.140)

that why do you have to grab things

Lex Fridman (51:21.620)

from things your friends have shared?

Lex Fridman (51:23.140)

Why not just like grab things?

Lex Fridman (51:25.420)

That's really fascinating to me.

Lex Fridman (51:27.420)

And that's something I've been thinking a lot about.

Lex Fridman (51:29.100)

And just like, why is it that when you log onto Twitter,

Kevin Systrom (51:34.940)

you're just sitting there looking at things from accounts

Lex Fridman (51:38.940)

that you've followed for whatever reason?

Lex Fridman (51:42.300)

And TikTok I think has done a wonderful job here,

Lex Fridman (51:44.460)

which is like, you can literally be anyone.

Lex Fridman (51:47.180)

And if you produce something fascinating, it'll go viral.

Lex Fridman (51:51.700)

But like, you don't have to be someone that anyone knows.

Kevin Systrom (51:54.780)

You don't have to have built up a giant following.

Lex Fridman (51:57.300)

You don't have to have paid for followers.

Kevin Systrom (52:00.060)

You don't have to try to maintain those followers.

Lex Fridman (52:01.860)

You literally just have to produce something interesting.

Kevin Systrom (52:04.620)

That is I think the future of social networking.

Lex Fridman (52:07.260)

That's the direction things will head.

Lex Fridman (52:10.020)

And I think what you'll find is it's far less

Lex Fridman (52:12.540)

about people manipulating distribution

Lex Fridman (52:15.860)

and far more about what is like, is this content good?

Lex Fridman (52:19.660)

And good is obviously a vague definition

Kevin Systrom (52:22.420)

that we could spend hours on.

Lex Fridman (52:23.700)

But different networks I think will decide

Kevin Systrom (52:27.140)

different value functions to decide what is good

Lex Fridman (52:29.300)

and what isn't good.

Lex Fridman (52:30.140)

And I think that's a fascinating direction.

Lex Fridman (52:32.100)

So that's almost like creating an internet.

Kevin Systrom (52:33.620)

I mean, that's what Google did for web pages

Lex Fridman (52:36.180)

that did page rank search.

Lex Fridman (52:39.380)

So it's discovery, you don't follow anybody on Google

Lex Fridman (52:42.660)

when you use a search engine.

Kevin Systrom (52:44.380)

You just discover web pages.

Lex Fridman (52:46.260)

And so what TikTok does is saying,

Kevin Systrom (52:49.780)

let's start from scratch.

Lex Fridman (52:51.740)

Let's like start a new internet

Lex Fridman (52:54.540)

and have people discover stuff on that new internet

Lex Fridman (52:56.820)

within a particular kind of pool of people.

Lex Fridman (52:59.820)

But what's so fascinating about this

Lex Fridman (53:01.300)

is like the field of information retrieval.

Kevin Systrom (53:05.140)

Like I always talked about as I was studying this stuff,

Lex Fridman (53:08.460)

it was used the word query and document.

Lex Fridman (53:10.500)

So I was like, why are they saying query and documents?

Lex Fridman (53:12.860)

Like they're literally imagine,

Kevin Systrom (53:14.660)

like if you just stop thinking about query

Lex Fridman (53:17.060)

as like literally a search query

Lex Fridman (53:18.900)

and a query could be a person.

Lex Fridman (53:20.460)

I mean, a lot of the way,

Kevin Systrom (53:22.100)

I'm not gonna claim to know how Instagram or Facebook

Lex Fridman (53:24.220)

machine learning works today,

Lex Fridman (53:26.140)

but if you want to find a match for a query,

Lex Fridman (53:30.300)

the query is actually the attributes of the person,

Kevin Systrom (53:33.260)

their age, their gender, where they're from,

Lex Fridman (53:36.740)

maybe some kind of summarization of their interests.

Lex Fridman (53:39.740)

And that's a query, right?

Lex Fridman (53:41.980)

And that matches against documents.

Lex Fridman (53:43.460)

And by the way, documents don't have to be texts.

Lex Fridman (53:45.100)

They can be videos, however long.

Kevin Systrom (53:48.180)

I don't know what the limit is on TikTok these days.

Lex Fridman (53:50.140)

They keep changing it.

Kevin Systrom (53:51.340)

My point is just, you've got a query,

Lex Fridman (53:53.380)

which is someone in search of something

Kevin Systrom (53:55.260)

that they want to match and you've got the document

Lex Fridman (53:58.260)

and it doesn't have to be text.

Kevin Systrom (53:59.460)

It could be anything.

Lex Fridman (54:00.700)

And how do you match make?

Lex Fridman (54:02.060)

And that's one of these like,

Lex Fridman (54:03.220)

I mean, I've spent a lot of time thinking about this

Lex Fridman (54:06.220)

and I don't claim to have mastered it at all,

Lex Fridman (54:08.820)

but I think it's so fascinating about where that will go

Kevin Systrom (54:11.700)

with new social networks.

Lex Fridman (54:13.300)

See, what I'm also fascinated by is metrics

Kevin Systrom (54:16.060)

that are different than engagement.

Lex Fridman (54:18.140)

So the other thing from an alien perspective,

Lex Fridman (54:20.900)

what social networks are doing is they,

Lex Fridman (54:25.420)

they in the short term,

Kevin Systrom (54:26.780)

bring out different aspects of each human being.

Lex Fridman (54:30.100)

So first, let me say that an algorithm or a social network

Kevin Systrom (54:38.580)

for each individual can bring out the best of that person

Lex Fridman (54:41.780)

or the worst of that person,

Kevin Systrom (54:43.380)

or there's a bunch of different parts to us,

Lex Fridman (54:45.620)

parts we're proud of that we are,

Kevin Systrom (54:48.780)

parts we're not so proud of.

Lex Fridman (54:50.780)

When we look at the big picture of our lives,

Kevin Systrom (54:53.100)

when we look back 30 days from now,

Lex Fridman (54:55.140)

am I proud that I said those things or not?

Lex Fridman (54:57.580)

Am I proud that I felt those things?

Lex Fridman (54:59.260)

Am I proud that I experienced or read those things

Lex Fridman (55:02.820)

or thought about those things?

Lex Fridman (55:04.780)

Just in that kind of self reflected kind of way.

Lex Fridman (55:08.060)

And so coupled with that,

Lex Fridman (55:10.180)

I wonder if it's possible to have different metrics

Kevin Systrom (55:12.220)

that are not just about engagement,

Lex Fridman (55:14.100)

but are about long term happiness,

Kevin Systrom (55:18.860)

growth of a human being,

Lex Fridman (55:20.260)

where they look back and say,

Kevin Systrom (55:22.220)

I am a better human being

Lex Fridman (55:23.700)

for having spent 100 hours on that app.

Lex Fridman (55:26.860)

And that feels like it's actually strongly correlated

Lex Fridman (55:30.580)

with engagement in the long term.

Kevin Systrom (55:33.300)

In the short term, it may not be,

Lex Fridman (55:34.740)

but in long term, it's like the same kind of thing

Kevin Systrom (55:37.820)

where you really fall in love with the product.

Lex Fridman (55:40.580)

You fall in love with an iPhone,

Kevin Systrom (55:41.980)

you fall in love with a car.

Lex Fridman (55:43.780)

That's what makes you fall in love

Kevin Systrom (55:45.060)

is like really being proud

Lex Fridman (55:49.260)

and just in a self reflected way,

Kevin Systrom (55:51.580)

understanding that you're a better human being

Lex Fridman (55:53.620)

for having used the thing.

Lex Fridman (55:55.020)

And that's what great relationships are made from.

Lex Fridman (55:58.780)

It's not just like you're hot

Lex Fridman (56:01.660)

and we like being together or something like that.

Lex Fridman (56:04.340)

It's more like I'm a better human being

Kevin Systrom (56:06.100)

because I'm with you.

Lex Fridman (56:07.620)

And that feels like a metric

Kevin Systrom (56:09.300)

that could be optimized for by the algorithms.

Lex Fridman (56:13.060)

But anytime I kind of talk about this with anybody,

Kevin Systrom (56:17.140)

they seem to say, yeah, okay,

Lex Fridman (56:18.620)

that's going to get out competed immediately

Kevin Systrom (56:21.060)

by the engagement if it's ad driven especially.

Lex Fridman (56:24.700)

I just don't think so.

Kevin Systrom (56:26.540)

I don't, I mean, a lot of it's just implementation.

Lex Fridman (56:30.700)

I'll say a couple of things.

Kevin Systrom (56:31.660)

One is to pull back the curtain on daily meetings

Lex Fridman (56:36.900)

inside of these large social media companies.

Kevin Systrom (56:40.380)

A lot of what management,

Lex Fridman (56:42.900)

or at least the people that are tweaking these algorithms

Kevin Systrom (56:45.700)

spend their time on are trade offs.

Lex Fridman (56:47.900)

And there's these things called value functions,

Kevin Systrom (56:50.100)

which are like, okay,

Lex Fridman (56:51.940)

we can predict the probability that you'll click

Kevin Systrom (56:54.780)

on this thing or the probability that you'll share it,

Lex Fridman (56:58.260)

or the probability that you will leave a comment on it

Kevin Systrom (57:01.180)

or the probability you'll dwell on it.

Lex Fridman (57:04.660)

Individual actions, right?

Lex Fridman (57:07.300)

And you've got this neural network

Lex Fridman (57:09.460)

that basically has a bunch of heads at the end

Lex Fridman (57:11.420)

and all of them are between zero and one and great.

Lex Fridman (57:14.900)

They all have values, right?

Kevin Systrom (57:15.980)

Or they all have probabilities.

Lex Fridman (57:19.780)

And then in these meetings, what they will do is say,

Kevin Systrom (57:21.580)

well, how much do we value a comment versus a click

Lex Fridman (57:26.780)

versus a share versus a,

Lex Fridman (57:29.180)

and then maybe even some downstream thing, right?

Lex Fridman (57:31.940)

That has nothing to do with the item there,

Lex Fridman (57:34.100)

but like driving follows or something.

Lex Fridman (57:37.540)

And what typically happens is they will say,

Lex Fridman (57:40.180)

well, what are our goals for this quarter at the company?

Lex Fridman (57:42.740)

Oh, we wanna drive sharing up, okay.

Kevin Systrom (57:44.740)

Well, let's turn down these metrics

Lex Fridman (57:46.780)

and turn up these metrics.

Lex Fridman (57:48.660)

And they blend them right into a single scalar

Lex Fridman (57:52.380)

which they're trying to optimize.

Kevin Systrom (57:55.260)

That is really hard because invariably

Lex Fridman (57:58.140)

you think you're solving for, I don't know,

Lex Fridman (57:59.860)

something called meaningful interactions, right?

Lex Fridman (58:02.780)

This was the big Facebook pivot.

Lex Fridman (58:04.180)

And I don't actually have any internal knowledge.

Lex Fridman (58:06.700)

Like I wasn't in those meetings,

Lex Fridman (58:09.060)

but at least from what we've seen over the last month or so,

Lex Fridman (58:12.900)

it seems by actually trying to optimize

Kevin Systrom (58:16.460)

for meaningful interactions,

Lex Fridman (58:18.140)

it had all these side effects of optimizing

Kevin Systrom (58:20.100)

for these other things.

Lex Fridman (58:22.100)

And I don't claim to fully understand them,

Lex Fridman (58:24.580)

but what I will say is that trade offs abound.

Lex Fridman (58:28.780)

And as much as you'd like to solve for one thing,

Kevin Systrom (58:31.740)

if you have a network of over a billion people,

Lex Fridman (58:34.700)

you're gonna have unintended consequences either way.

Lex Fridman (58:36.700)

And it gets really hard.

Lex Fridman (58:38.860)

So what you're describing is effectively a value model

Kevin Systrom (58:41.420)

that says like, can we capture,

Lex Fridman (58:43.420)

this is the thing that I spent a lot of time thinking about,

Kevin Systrom (58:45.620)

like, can you capture utility

Lex Fridman (58:49.940)

in a way that like actually measures someone's happiness

Lex Fridman (58:53.580)

that isn't just a, what do they call it?

Lex Fridman (58:56.660)

A surrogate problem where you say, well,

Kevin Systrom (58:59.100)

kind of think like the more you use the product,

Lex Fridman (59:01.900)

the happier you are.

Kevin Systrom (59:03.060)

That was always the argument at Facebook, by the way.

Lex Fridman (59:04.780)

It was like, well, people use it more,

Lex Fridman (59:07.500)

so they must be more happy.

Lex Fridman (59:09.300)

Turns out there are like a lot of things you use more

Kevin Systrom (59:11.140)

that make you less happy in the world.

Lex Fridman (59:12.900)

Not talking about Facebook,

Kevin Systrom (59:14.100)

just let's think about whether it's gambling or whatever,

Lex Fridman (59:17.660)

like that you can do more of,

Lex Fridman (59:19.060)

but doesn't necessarily make you happier.

Lex Fridman (59:20.740)

So the idea that time equals happiness,

Kevin Systrom (59:22.780)

obviously you can't map utility and time together easily.

Lex Fridman (59:27.060)

There are a lot of edge cases.

Lex Fridman (59:28.820)

So when you look around the world and you say,

Lex Fridman (59:30.460)

well, what are all the ways we can model utility?

Kevin Systrom (59:32.700)

That is like one of the,

Lex Fridman (59:34.380)

please, if you know someone smart doing this,

Kevin Systrom (59:36.260)

introduce me because I'm fascinated by it.

Lex Fridman (59:38.500)

And it seems really tough.

Lex Fridman (59:40.580)

But the idea that reinforcement learning,

Lex Fridman (59:42.740)

like everyone interesting I know in machine learning,

Kevin Systrom (59:46.460)

like I was really interested in recommender systems

Lex Fridman (59:48.820)

and supervised learning.

Lex Fridman (59:49.820)

And the more I dug into it, I was like,

Lex Fridman (59:52.700)

oh, literally everyone smart

Kevin Systrom (59:54.460)

is working on reinforcement learning.

Lex Fridman (59:56.460)

Like literally everyone.

Kevin Systrom (59:57.700)

You just made people at OpenAI and DeepMind very happy, yes.