James Gosling: Java, JVM, Emacs, and the Early Days of Computing

Lex Fridman (01:42.320)

It's the best way to support this podcast.

James Gosling (01:45.800)

This show sponsored by Public Goods, the one stop shop for affordable,

Lex Fridman (01:50.040)

sustainable, healthy household products.

James Gosling (01:53.520)

I take their fish oil and use their toothbrush, for example.

Lex Fridman (01:57.400)

Their products often have a minimalist black and white design that I find

James Gosling (02:00.880)

to be just beautiful.

Lex Fridman (02:02.480)

Some people ask why I wear this black suit and tie.

James Gosling (02:06.080)

There's a simplicity to it that to me focuses my mind on the most important

Lex Fridman (02:10.560)

bits of every moment of every day, pulling only at the thread of the

James Gosling (02:15.560)

essential in all that life has to throw at me.

Lex Fridman (02:18.360)

It's not about how I look, it's about how I feel.

James Gosling (02:21.000)

That's what design is to me, creating an inner conscious experience,

Lex Fridman (02:24.960)

not an external look.

James Gosling (02:27.240)

Anyway, Public Goods plants one tree for every order placed, which is kind of cool.

Lex Fridman (02:32.920)

Visit publicgoods.com slash Lex, or use code Lex at checkout to get 15 bucks

James Gosling (02:38.120)

off your first order.

Lex Fridman (02:40.440)

This show is also sponsored by BetterHelp, spelled H E L P, help.

James Gosling (02:45.080)

Check it out at betterhelp.com slash Lex.

Lex Fridman (02:48.160)

They figure out what you need and match you with a licensed professional

James Gosling (02:51.720)

therapist in under 48 hours.

Lex Fridman (02:54.080)

I chat with the person on there and enjoy it.

James Gosling (02:57.200)

Of course, I also regularly talk to David Goggins these days, who is definitely not

James Gosling (03:02.320)

a licensed professional therapist, but he does help me meet his and my demons and

James Gosling (03:08.840)

become comfortable to exist in their presence.

James Gosling (03:11.400)

Everyone is different, but for me, I think suffering is essential for creation, but

James Gosling (03:16.320)

you can suffer beautifully in a way that doesn't destroy you.

Lex Fridman (03:20.160)

I think therapy can help in whatever form that therapy takes.

Lex Fridman (03:24.000)

And I do think that BetterHelp is an option worth trying.

Lex Fridman (03:27.720)

They're easy, private, affordable, and available worldwide.

James Gosling (03:31.840)

You can communicate by text anytime and schedule weekly audio and video sessions.

Lex Fridman (03:36.760)

Check it out at betterhelp.com slash Lex.

James Gosling (03:39.440)

This show is also sponsored by Express CPM.

Lex Fridman (03:42.240)

You can use it to unlock movies and shows that are only

James Gosling (03:45.560)

available in other countries.

Lex Fridman (03:47.360)

I did this recently with Star Trek Discovery and UK Netflix, mostly because

James Gosling (03:52.200)

I wonder what it's like to live in London.

Lex Fridman (03:54.800)

I'm thinking of moving from Boston to a place where I can build the

James Gosling (03:57.720)

business I've always dreamed of building.

Lex Fridman (04:00.200)

London is probably not in the top three, but top 10 for sure.

James Gosling (04:04.680)

The number one show I've been to is the one that I'm most excited about.

Lex Fridman (04:08.000)

The number one choice currently is Austin.

James Gosling (04:11.640)

For many reasons that I'll probably speak to another time.

Lex Fridman (04:14.800)

San Francisco, unfortunately dropped out from the number one spot,

Lex Fridman (04:18.880)

but it's still in the running.

Lex Fridman (04:20.600)

If you have advice, let me know.

James Gosling (04:22.960)

Anyway, check out ExpressVPN.

Lex Fridman (04:24.920)

It lets you change your location to almost 100 countries and it's super fast.

James Gosling (04:29.600)

Go to expressvpn.com slash LexPod to get an extra three

Lex Fridman (04:33.520)

months of ExpressVPN for free.

James Gosling (04:35.560)

That's expressvpn.com slash LexPod.

Lex Fridman (04:40.240)

And now here's my conversation with James Gosling.

James Gosling (04:44.680)

I've read somewhere that the square root of two is your favorite irrational number.

Lex Fridman (04:49.080)

I have no idea where that got started.

Lex Fridman (04:53.440)

Is there any truth to it?

Lex Fridman (04:54.600)

Is there anything in mathematics or numbers that you find beautiful?

James Gosling (04:58.400)

Oh, well, there's lots of things in math that's really beautiful.

Lex Fridman (05:03.120)

I used to consider myself really good at math and these days I consider

James Gosling (05:07.960)

myself really bad at math.

James Gosling (05:11.040)

I never really had a thing for the square root of two, but when I was a teenager,

James Gosling (05:18.480)

there was this book called The Dictionary of Curious and Interesting Numbers,

Lex Fridman (05:23.360)

which for some reason I read through and damn near memorized the whole thing.

Lex Fridman (05:33.040)

And I started this weird habit of when I was like filling out checks or paying for things

Lex Fridman (05:43.120)

with credit cards, I would want to make the receipt add up to a new number.

Lex Fridman (05:49.840)

Is there some numbers that stuck with you that just kind of make you feel good?

James Gosling (05:54.320)

They all have a story and fortunately, I've actually mostly forgotten all of them.

James Gosling (06:02.800)

Are they, uh, so like 42, uh, well, yeah, I mean, that one 42 is pretty magical.

Lex Fridman (06:08.880)

And then the irrationals, I mean, but is there a square root of two story in there

Lex Fridman (06:12.880)

somewhere, how did that come about?

James Gosling (06:14.720)

It's, it's like the only number that has destroyed a religion in which way, well,

James Gosling (06:22.640)

the, the pathogorians, they, they believed that all numbers were perfect and you

Lex Fridman (06:28.160)

could re represent anything as, as a, as a rational number.

James Gosling (06:33.040)

And, um, in that, in that time period, um, the, this proof came out that there was no,

Lex Fridman (06:44.720)

you know, rational fraction whose value was equal to the square root of two.

Lex Fridman (06:50.720)

And that, that means nothing in this world, right?

Lex Fridman (06:54.320)

Yeah.

James Gosling (06:54.640)

So, so, so, so, so, so, so, so, so, so, so, so, so, so, so, so, so, so, so, so, so, so,

Lex Fridman (07:02.800)

in this world is perfect, not even mathematics.

James Gosling (07:05.600)

Well, it means that your definition of perfect was imperfect.

Lex Fridman (07:11.280)

Well then, then there's the Gaitl incompleteness theorems and the 20th

James Gosling (07:14.640)

century that ruined it once again for everybody.

Lex Fridman (07:17.360)

Yeah.

James Gosling (07:17.840)

Although, although, although Goerl's theorem, um, you know, the lesson I take

Lex Fridman (07:25.360)

from Goerl's theorem is not that, you know, there are things you can't know,

James Gosling (07:29.920)

which is fundamentally what it says.

Lex Fridman (07:35.920)

But people want black and white answers.

James Gosling (07:39.360)

They want true or false.

Lex Fridman (07:43.540)

But if you allow a three state logic

James Gosling (07:48.000)

that is true, false, or maybe,

Lex Fridman (07:51.860)

then life's good.

James Gosling (07:55.360)

I feel like there's a parallel to modern political discourse

Lex Fridman (07:58.660)

in there somewhere, but let me ask.

Lex Fridman (08:04.160)

So with your kind of early love or appreciation

Lex Fridman (08:10.100)

of the beauty of mathematics,

Lex Fridman (08:11.480)

do you see a parallel between that world

Lex Fridman (08:15.000)

and the world of programming?

James Gosling (08:17.920)

Programming is all about logical structure,

Lex Fridman (08:21.960)

understanding the patterns that come out of computation,

James Gosling (08:32.680)

understanding sort of, I mean, it's often like the path

Lex Fridman (08:38.880)

through the graph of possibilities to find a short route.

James Gosling (08:45.760)

Meaning like find a short program

Lex Fridman (08:48.320)

that gets the job done kind of thing.

Lex Fridman (08:50.960)

But so then on the topic of irrational numbers,

Lex Fridman (08:54.560)

do you see programming?

James Gosling (08:58.640)

You just painted it so cleanly.

Lex Fridman (09:01.260)

It's a little of this trajectory

James Gosling (09:02.880)

to find like a nice little program,

Lex Fridman (09:05.200)

but do you see it as fundamentally messy?

Lex Fridman (09:08.880)

Maybe unlike mathematics?

Lex Fridman (09:10.840)

I don't think of it as, I mean, you know,

James Gosling (09:13.740)

you watch somebody who's good at math do math

Lex Fridman (09:16.820)

and you know, often it's fairly messy.

James Gosling (09:22.480)

Sometimes it's kind of magical.

Lex Fridman (09:27.680)

When I was a grad student, one of the students,

James Gosling (09:32.920)

his name was Jim Sachs, was he had this reputation

Lex Fridman (09:37.920)

of being sort of a walking, talking human

James Gosling (09:46.920)

theorem proving machine.

Lex Fridman (09:49.280)

And if you were having a hard problem with something,

James Gosling (09:51.520)

you could just like accost him in the hall

Lex Fridman (09:54.260)

and say, Jim, and he would do this funny thing

James Gosling (09:59.740)

where he would stand up straight.

Lex Fridman (10:01.840)

His eyes would kind of defocus.

James Gosling (10:03.560)

He'd go, you know, just like something in today's movies.

Lex Fridman (10:10.880)

And then he'd straighten up and say n log n and walk away.

Lex Fridman (10:15.800)

And you'd go, well, okay, so n log n is the answer.

Lex Fridman (10:20.520)

How did he get there?

James Gosling (10:24.240)

By which time he's, you know, down the hallway somewhere.

Lex Fridman (10:27.120)

Yeah, he's just the oracle, the black box

James Gosling (10:30.280)

that just gives you the answer.

Lex Fridman (10:31.360)

Yeah, and then you have to figure out the path

James Gosling (10:33.360)

from the question to the answer.

Lex Fridman (10:36.440)

I think in one of the videos I watched,

James Gosling (10:38.160)

you mentioned Don Knuth, well, at least recommending his,

Lex Fridman (10:44.560)

you know, his book is something people should read.

James Gosling (10:47.800)

Oh, yeah.

Lex Fridman (10:48.640)

But in terms of, you know, theoretical computer science,

Lex Fridman (10:53.920)

do you see something beautiful that has been inspiring

Lex Fridman (10:58.400)

to you, speaking of n log n, in your work

James Gosling (11:01.360)

on programming languages, that's in that whole world

Lex Fridman (11:06.040)

of algorithms and complexity and, you know,

Lex Fridman (11:10.240)

these kinds of more formal mathematical things?

Lex Fridman (11:13.520)

Or did that not really stick with you

Lex Fridman (11:16.720)

in your programming life?

Lex Fridman (11:20.040)

It did stick pretty clearly for me,

James Gosling (11:24.080)

because one of the things that I care about

Lex Fridman (11:27.040)

is being able to sort of look at a piece of code

Lex Fridman (11:34.240)

and be able to prove to myself that it works.

Lex Fridman (11:41.360)

And, you know, so for example, I find that I'm at odds

James Gosling (11:51.440)

with many of the people around me over issues

Lex Fridman (11:55.440)

about like how you lay out a piece of software, right?

James Gosling (12:03.400)

You know, so software engineers get really cranky

Lex Fridman (12:07.600)

about how they format the documents that are the programs,

James Gosling (12:11.380)

you know, where they put new lines and where they put,

Lex Fridman (12:14.280)

you know.

James Gosling (12:15.120)

The braces.

Lex Fridman (12:15.960)

The braces and all the rest of that, right.

Lex Fridman (12:18.720)

And I tend to go for a style that's very dense.

Lex Fridman (12:27.600)

Minimize the white space.

James Gosling (12:30.080)

Yeah, well, to maximize the amount

Lex Fridman (12:34.360)

that I can see at once, right?

Lex Fridman (12:37.600)

So I like to be able to see a whole function

Lex Fridman (12:40.760)

and to understand what it does,

James Gosling (12:43.080)

rather than have to go scroll, scroll, scroll

Lex Fridman (12:44.960)

and remember, right?

James Gosling (12:46.160)

Yeah, I'm with you on that.

Lex Fridman (12:47.520)

Yeah, that's.

Lex Fridman (12:48.360)

And people don't like that.

Lex Fridman (12:52.640)

Yeah, I've had, you know, multiple times

James Gosling (12:56.920)

when engineering teams have staged

Lex Fridman (13:01.000)

what was effectively an intervention.

James Gosling (13:07.760)

You know, where they invite me to a meeting

Lex Fridman (13:10.120)

and everybody's arrived before me

Lex Fridman (13:12.100)

and they all look at me and say,

Lex Fridman (13:14.500)

James, about your coding style, I'm sort of an odd person

James Gosling (13:23.080)

to be programming because I don't think very well verbally.

Lex Fridman (13:31.560)

I am just naturally a slow reader.

James Gosling (13:37.760)

I'm what most people would call a visual thinker.

Lex Fridman (13:42.040)

So when you think about a program,

Lex Fridman (13:43.560)

what do you see?

Lex Fridman (13:45.080)

I see pictures, right?

Lex Fridman (13:47.320)

So when I look at a piece of code on a piece of paper,

Lex Fridman (13:52.120)

it very quickly gets transformed into a picture.

James Gosling (13:58.320)

And, you know, it's almost like a piece of machinery

Lex Fridman (14:01.880)

with, you know, this connected to that and.

James Gosling (14:05.440)

Like these gears and different sizes.

Lex Fridman (14:07.480)

Yeah, yeah, I see them more like that

James Gosling (14:11.320)

than I see the sort of verbal structure

Lex Fridman (14:15.960)

or the lexical structure of letters.

Lex Fridman (14:18.640)

So then when you look at the program,

Lex Fridman (14:19.960)

that's why you want to see it all in the same place,

James Gosling (14:21.920)

then you can just map it to something visual.

Lex Fridman (14:24.180)

Yeah, and it just kind of like,

James Gosling (14:25.600)

like it leaps off the page at me and.

Lex Fridman (14:28.240)

Yeah, what are the inputs, what are the outputs?

Lex Fridman (14:30.800)

What the heck is this thing doing?

Lex Fridman (14:32.160)

Yeah, yeah.

James Gosling (14:33.680)

Getting a whole vision of it.

Lex Fridman (14:35.720)

Can we go back into your memory?

James Gosling (14:39.080)

Memory, long term memory access.

Lex Fridman (14:42.060)

What's the first program you've ever written?

James Gosling (14:47.460)

Oh, I have no idea what the first one was.

Lex Fridman (14:52.100)

I mean, I know the first machine

James Gosling (14:54.780)

that I learned to program on.

Lex Fridman (14:58.520)

What is it?

James Gosling (14:59.360)

Was a PDP eight at the University of Calgary.

Lex Fridman (15:06.500)

Do you remember the specs?

James Gosling (15:08.100)

Oh yeah, so the thing had 4K of RAM.

Lex Fridman (15:12.700)

Nice.

James Gosling (15:13.700)

12 bit words.

Lex Fridman (15:16.340)

The clock rate was,

James Gosling (15:21.500)

it was about a third of a megahertz.

Lex Fridman (15:26.580)

Oh, so you didn't even get to the M, okay.

James Gosling (15:29.260)

Yeah, yeah, so we're like 10,000 times faster these days.

Lex Fridman (15:34.260)

10,000 times faster these days.

Lex Fridman (15:38.020)

And was this kind of like a super computer,

Lex Fridman (15:40.500)

like a serious computer for.

James Gosling (15:42.340)

No, the PDP eight I was the first thing

Lex Fridman (15:46.180)

that people were calling like mini computer.

James Gosling (15:49.700)

Got it.

Lex Fridman (15:50.660)

They were sort of inexpensive enough

James Gosling (15:53.180)

that a university lab could maybe afford to buy one.

Lex Fridman (15:59.340)

And was there time sharing, all that kind of stuff?

James Gosling (16:02.100)

There actually was a time sharing OS for that,

Lex Fridman (16:06.220)

but it wasn't used really widely.

James Gosling (16:10.500)

The machine that I learned on was one

Lex Fridman (16:12.980)

that was kind of hidden in the back corner

James Gosling (16:17.560)

of the computer center.

Lex Fridman (16:22.340)

And it was bought as part of a project

James Gosling (16:27.340)

to do computer networking,

Lex Fridman (16:34.180)

but they didn't actually use it very much.

James Gosling (16:38.740)

It was mostly just kind of sitting there

Lex Fridman (16:42.200)

and it was kind of sitting there

Lex Fridman (16:44.180)

and I noticed it was just kind of sitting there.

Lex Fridman (16:47.020)

And so I started fooling around with it

Lex Fridman (16:50.580)

and nobody seemed to mind.

Lex Fridman (16:52.900)

So I just kept doing that and.

Lex Fridman (16:56.020)

And it had a keyboard and like a monitor, are we?

Lex Fridman (16:58.980)

Oh, this is way before monitors were common.

Lex Fridman (17:02.580)

So it was literally a model 33 teletype

Lex Fridman (17:07.820)

with a paper tape reader.

James Gosling (17:12.060)

Okay, so the user interface wasn't very good.

Lex Fridman (17:14.580)

Yeah, it was the first computer ever built

James Gosling (17:20.140)

with integrated circuits, but by integrated circuits,

Lex Fridman (17:24.820)

I mean that they would have like 10 or 12 transistors

James Gosling (17:29.620)

on one piece of silicon,

Lex Fridman (17:31.940)

not the 10 or 12 billion that the machines have today.

Lex Fridman (17:38.740)

So what did that, I mean, feel like if you remember those?

Lex Fridman (17:43.740)

I mean, did you have kind of inklings of the magic

James Gosling (17:48.880)

of exponential kind of improvement of Moore's law

Lex Fridman (17:52.840)

of the potential of the future that,

Lex Fridman (17:55.220)

was that your fingertips kind of thing?

Lex Fridman (17:57.100)

Or was it just a cool?

James Gosling (17:58.480)

Yeah, it was just a toy.

Lex Fridman (18:01.140)

I had always liked building stuff,

Lex Fridman (18:04.080)

but one of the problems with building stuff

Lex Fridman (18:06.820)

is that you need to have parts.

James Gosling (18:08.740)

You need to have pieces of wood or wire or switches

Lex Fridman (18:13.580)

or stuff like that.

Lex Fridman (18:14.980)

And those all cost money.

Lex Fridman (18:16.820)

And here you could build.

James Gosling (18:17.940)

You could build arbitrarily complicated things

Lex Fridman (18:20.640)

and I didn't need any physical materials.

James Gosling (18:25.180)

It required no money.

Lex Fridman (18:27.220)

That's a good way to put programming.

James Gosling (18:29.060)

You're right, it's, if you love building things,

Lex Fridman (18:35.900)

completely accessible, you don't need anything.

James Gosling (18:38.780)

Anybody from anywhere could just build

Lex Fridman (18:40.660)

something really cool.

James Gosling (18:41.600)

Yeah, yeah.

Lex Fridman (18:43.060)

If you've got access to a computer,

James Gosling (18:44.760)

you can build all kinds of crazy stuff.

Lex Fridman (18:51.420)

And when you were somebody like me

James Gosling (18:58.220)

who had like really no money, and I mean,

Lex Fridman (19:05.340)

I remember just lusting after being

James Gosling (19:08.100)

able to buy like a transistor.

Lex Fridman (19:15.780)

And when I would do sort of electronics kind of projects,

James Gosling (19:20.260)

they were mostly made, done by like dumpster diving for trash.

Lex Fridman (19:27.420)

And one of my big hauls was discarded relay racks

James Gosling (19:33.580)

from the back of the phone company switching center.

Lex Fridman (19:37.340)

Oh, nice.

James Gosling (19:39.220)

That was the big memorable treasure.

Lex Fridman (19:41.340)

Oh, yeah.

James Gosling (19:42.020)

Yeah, that was a really.

Lex Fridman (19:43.260)

What do you use that for?

James Gosling (19:45.140)

I built a machine that played tic tac toe.

Lex Fridman (19:50.420)

Nice.

James Gosling (19:51.180)

Out of relays.

Lex Fridman (19:52.060)

Of course, the thing that was really hard

James Gosling (19:55.780)

was that all the relays required a specific voltage.

Lex Fridman (1:00:04.460)

must be open source, I get kind of weird about that

James Gosling (1:00:10.860)

because it's sort of like saying some versions of that

Lex Fridman (1:00:17.420)

end up saying that all software engineers

James Gosling (1:00:22.300)

must take a vow of poverty, right, as though.

Lex Fridman (1:00:28.100)

It's unethical to have money.

James Gosling (1:00:30.060)

Yeah.

Lex Fridman (1:00:31.540)

To build a company, right.

Lex Fridman (1:00:34.500)

And there's a slice of me that actually kind of buys into that

Lex Fridman (1:00:41.260)

because people who make billions of dollars off of a patent,

Lex Fridman (1:00:50.060)

and the patent came from literally a stroke

Lex Fridman (1:00:56.260)

of lightning that hits you as you lie half awake in bed.

James Gosling (1:01:02.500)

Yeah, that's lucky.

Lex Fridman (1:01:03.980)

Good for you.

James Gosling (1:01:06.060)

The way that that sometimes sort of explodes

Lex Fridman (1:01:09.700)

into something that looks to me a lot like exploitation,

James Gosling (1:01:14.420)

you see a lot of that in the drug industry.

Lex Fridman (1:01:18.540)

You know, when you've got medications that cost you

James Gosling (1:01:26.620)

like $100 a day, and it's like, no.

Lex Fridman (1:01:33.900)

Yeah, so the interesting thing about the sort of open source,

Lex Fridman (1:01:39.060)

what bothers me is when something is not open source,

Lex Fridman (1:01:44.220)

and because of that, it's a worse product.

James Gosling (1:01:48.700)

Yeah.

Lex Fridman (1:01:49.700)

So like, I mean, if I look at your just implementation

James Gosling (1:01:52.700)

of Emacs, like that could have been the dominant implementation.

Lex Fridman (1:01:56.620)

Like I use Emacs.

James Gosling (1:01:57.460)

That's my main ID.

Lex Fridman (1:01:58.700)

I apologize to the world, but I still love it.

Lex Fridman (1:02:02.540)

And I could have been using your implementation of Emacs.

Lex Fridman (1:02:09.020)

And why aren't I?

Lex Fridman (1:02:11.180)

So are you using the GNU Emacs?

Lex Fridman (1:02:14.180)

I guess the default on Linux is that GNU?

James Gosling (1:02:16.380)

Yeah.

Lex Fridman (1:02:16.900)

And that, through a strange passage,

James Gosling (1:02:20.420)

started out as the one that I wrote.

Lex Fridman (1:02:22.220)

Exactly.

Lex Fridman (1:02:22.740)

So it still has a, yeah.

Lex Fridman (1:02:26.540)

Well, and part of that was because in the last couple

James Gosling (1:02:32.860)

of years of grad school, it became really clear

Lex Fridman (1:02:38.980)

to me that I was either going to be Mr. Emacs forever

James Gosling (1:02:46.060)

or I was going to graduate.

Lex Fridman (1:02:50.060)

Got it.

James Gosling (1:02:50.580)

I couldn't actually do both.

Lex Fridman (1:02:54.340)

Was that a hard decision?

James Gosling (1:02:56.260)

That's so interesting to think about you as a,

Lex Fridman (1:02:58.700)

like it's a different trajectory that could have happened.

James Gosling (1:03:01.260)

Yeah.

Lex Fridman (1:03:01.900)

That's fascinating.

Lex Fridman (1:03:04.420)

And maybe I could be fabulously wealthy today

Lex Fridman (1:03:10.220)

if I had become Mr. Emacs, and Emacs

James Gosling (1:03:13.580)

had mushroomed into a series of text processing applications

Lex Fridman (1:03:19.220)

and all kinds of stuff.

Lex Fridman (1:03:20.340)

And I would have, you know.

Lex Fridman (1:03:24.340)

But I have a long history of financially suboptimal

Lex Fridman (1:03:30.620)

decisions because I didn't want that life, right?

Lex Fridman (1:03:38.300)

And I went to grad school because I wanted to graduate.

Lex Fridman (1:03:47.300)

And being Mr. Emacs for a while was kind of fun,

Lex Fridman (1:03:58.260)

and then it kind of became not fun.

Lex Fridman (1:04:03.420)

And when it was not fun, there was

Lex Fridman (1:04:10.540)

no way I could pay my rent, right?

Lex Fridman (1:04:17.060)

And I was like, OK, do I carry on as a grad student?

Lex Fridman (1:04:23.780)

I had a research assistantship, and I

James Gosling (1:04:25.620)

was sort of living off of that.

Lex Fridman (1:04:27.220)

And I was trying to do my, you know,

James Gosling (1:04:30.180)

I was doing all my RA work, all my RA,

Lex Fridman (1:04:33.380)

you know, being grad student work

Lex Fridman (1:04:35.220)

and being Mr. Emacs all at the same time.

Lex Fridman (1:04:40.660)

And I decided to pick one.

Lex Fridman (1:04:45.260)

And one of the things that I did at the time

Lex Fridman (1:04:48.340)

was I went around all the people I knew on the ARPANET who

James Gosling (1:04:53.660)

might be able to take over looking after Emacs.

Lex Fridman (1:04:58.980)

And pretty much everybody said, eh, I got a day job.

Lex Fridman (1:05:06.300)

So I actually found two folks and a couple

Lex Fridman (1:05:12.340)

of folks in a garage in New Jersey, complete with a dog,

James Gosling (1:05:20.540)

who were willing to take it over.

Lex Fridman (1:05:22.940)

But they were going to have to charge money.

Lex Fridman (1:05:26.140)

But my deal with them was that they would only,

Lex Fridman (1:05:31.180)

that they would make it free for universities and schools

Lex Fridman (1:05:33.700)

and stuff.

Lex Fridman (1:05:34.980)

And they said sure.

Lex Fridman (1:05:36.620)

And you know, that upset some people.

Lex Fridman (1:05:41.100)

So you have some, now I don't know the full history of this,

Lex Fridman (1:05:43.860)

but I think it's kind of interesting.

Lex Fridman (1:05:46.780)

You have some tension with Mr. Richard Stallman over the,

Lex Fridman (1:05:54.340)

and he kind of represents this kind of,

Lex Fridman (1:05:56.660)

like you mentioned, free software,

James Gosling (1:06:03.660)

sort of a dogmatic focus on, yeah.

Lex Fridman (1:06:07.260)

All information must be free.

James Gosling (1:06:10.140)

Must be free.

Lex Fridman (1:06:10.860)

So what, is there an interesting way

James Gosling (1:06:14.220)

to paint a picture of the disagreement

Lex Fridman (1:06:17.420)

you have with Richard through the years?

James Gosling (1:06:19.900)

My basic opposition is that when you say information

Lex Fridman (1:06:26.980)

must be free, to a really extreme form that

James Gosling (1:06:32.100)

turns into all people whose job is

Lex Fridman (1:06:38.780)

the production of everything from movies to software.

James Gosling (1:06:49.300)

They must all take a vow of poverty

Lex Fridman (1:06:52.580)

because information must be free.

Lex Fridman (1:06:56.740)

And that doesn't work for me.

Lex Fridman (1:06:59.980)

And I don't want to be wildly rich.

James Gosling (1:07:05.060)

I am not wildly rich.

Lex Fridman (1:07:08.780)

I do OK.

Lex Fridman (1:07:14.260)

But I do actually, I can feed my children.

Lex Fridman (1:07:20.620)

Yeah, I totally agree with you.

James Gosling (1:07:22.220)

It does just make me sad that sometimes

Lex Fridman (1:07:25.340)

the closing of the source, for some reason

James Gosling (1:07:28.540)

that people that, like a bureaucracy begins to build,

Lex Fridman (1:07:33.620)

and sometimes it doesn't, it hurts the product.

James Gosling (1:07:37.460)

Oh, absolutely.

Lex Fridman (1:07:38.660)

Absolutely.

James Gosling (1:07:39.380)

It's always sad.

Lex Fridman (1:07:40.420)

And there is a balance in there.

James Gosling (1:07:44.740)

There's a balance.

Lex Fridman (1:07:47.060)

And it's not hard over rapacious capitalism.

Lex Fridman (1:07:56.100)

And it's not hard over in the other direction.

Lex Fridman (1:08:02.660)

And a lot of the open source movement,

James Gosling (1:08:07.820)

they have been managing to find the path

Lex Fridman (1:08:11.780)

to actually making money.

Lex Fridman (1:08:15.220)

So doing things like service and support

Lex Fridman (1:08:18.900)

works for a lot of people.

Lex Fridman (1:08:20.860)

And there are some ways where it's kind of, some of them

Lex Fridman (1:08:33.100)

are a little perverse.

Lex Fridman (1:08:35.820)

So as a part of things like this Sarbanes–Oxley Act

Lex Fridman (1:08:43.180)

and various people's interpretations

James Gosling (1:08:45.860)

of all kinds of accounting principles.

Lex Fridman (1:08:49.860)

And this is kind of a worldwide thing.

Lex Fridman (1:08:51.860)

But if you've got a corporation that

Lex Fridman (1:08:55.300)

is depending on some piece of software,

James Gosling (1:09:01.220)

often various accounting and reporting standards

Lex Fridman (1:09:04.540)

say if you don't have a support contract on this thing

James Gosling (1:09:08.420)

that your business is depending on, then that's bad.

Lex Fridman (1:09:15.940)

So if you've got a database, you need to pay for support.

Lex Fridman (1:09:24.500)

But there's a difference between the sort of support contracts

Lex Fridman (1:09:30.100)

that the average open source database producer charges

Lex Fridman (1:09:37.060)

and what somebody who is truly rapacious like Oracle charges.

Lex Fridman (1:09:44.500)

Yeah, so it's a balance, like you said.

James Gosling (1:09:47.260)

It is absolutely a balance.

Lex Fridman (1:09:49.660)

And there are a lot of different ways

James Gosling (1:09:57.060)

to make the math work out for everybody.

Lex Fridman (1:10:04.900)

And the very unbalanced sort of like the winner

James Gosling (1:10:16.700)

takes all thing that happens in so much of modern commerce,

Lex Fridman (1:10:23.140)

that just doesn't work for me either.

James Gosling (1:10:26.180)

I know you've talked about this in quite a few places,

Lex Fridman (1:10:31.380)

but you have created one of the most popular programming

James Gosling (1:10:37.020)

languages in the world.

Lex Fridman (1:10:39.180)

This is a programming language that I first

James Gosling (1:10:41.740)

learned about object oriented programming with.

Lex Fridman (1:10:46.740)

I think it's a programming language

James Gosling (1:10:49.260)

that a lot of people use in a lot of different places

Lex Fridman (1:10:52.620)

and millions of devices today, Java.

Lex Fridman (1:10:55.740)

So the absurd question, but can you

Lex Fridman (1:11:00.420)

tell the origin story of Java?

Lex Fridman (1:11:03.060)

So a long time ago at Sun in about 1990,

Lex Fridman (1:11:07.340)

there was a group of us who were kind of worried

James Gosling (1:11:12.740)

that there was stuff going on in the universe of computing

Lex Fridman (1:11:18.620)

that the computing industry was missing out on.

Lex Fridman (1:11:21.020)

And so a few of us started this project

Lex Fridman (1:11:29.740)

at Sun that really got going.

James Gosling (1:11:32.380)

I mean, we started talking about it in 1990,

Lex Fridman (1:11:34.860)

and it really got going in 91.

Lex Fridman (1:11:39.020)

And it was all about what was happening

Lex Fridman (1:11:44.860)

in terms of computing hardware processors

Lex Fridman (1:11:48.820)

and networking and all of that that was outside

Lex Fridman (1:11:53.220)

of the computer industry.

Lex Fridman (1:11:54.340)

And that was everything from the sort

Lex Fridman (1:11:59.060)

of early glimmers of cell phones that were happening then

James Gosling (1:12:03.500)

to you look at elevators and locomotives

Lex Fridman (1:12:07.940)

and process control systems in factories

Lex Fridman (1:12:12.500)

and all kinds of audio equipment and video equipment.

Lex Fridman (1:12:20.820)

They all had processors in them, and they were all

James Gosling (1:12:22.740)

doing stuff with them.

Lex Fridman (1:12:24.820)

And it sort of felt like there was something going on there

James Gosling (1:12:30.860)

that we needed to understand.

Lex Fridman (1:12:34.420)

And

Lex Fridman (1:12:36.020)

So C and C++ was in the air already.

Lex Fridman (1:12:38.980)

Oh, no, C and C++ absolutely owned the universe

James Gosling (1:12:42.300)

at that time.

Lex Fridman (1:12:42.940)

Everything was written in C and C++.

Lex Fridman (1:12:45.140)

So where was the hunch that there

Lex Fridman (1:12:46.860)

was a need for a revolution?

James Gosling (1:12:48.780)

Well, so the need for a revolution

Lex Fridman (1:12:50.700)

was not about a language.

James Gosling (1:12:54.500)

It was just as simple and vague as there

Lex Fridman (1:12:59.900)

are things happening out there.

James Gosling (1:13:03.620)

We need to understand them.

Lex Fridman (1:13:04.980)

We need to understand them.

Lex Fridman (1:13:08.580)

And so a few of us went on several somewhat epic road

Lex Fridman (1:13:16.900)

trips.

Lex Fridman (1:13:19.420)

Literal road trips?

Lex Fridman (1:13:20.660)

Literal road trips.

James Gosling (1:13:21.860)

It's like get on an airplane, go to Japan,

Lex Fridman (1:13:24.740)

visit Toshiba and Sharp and Mitsubishi and Sony

Lex Fridman (1:13:31.820)

and all of these folks.

Lex Fridman (1:13:33.700)

And because we worked for Sun, we

James Gosling (1:13:36.620)

had folks who were willing to give us introductions.

Lex Fridman (1:13:42.300)

We visited Samsung and a bunch of Korean companies.

Lex Fridman (1:13:50.060)

And we went all over Europe.

Lex Fridman (1:13:51.340)

We went to places like Philips and Siemens and Thomson.

Lex Fridman (1:13:56.620)

And what did you see there?

Lex Fridman (1:14:00.020)

For me, one of the things that sort of leapt out

James Gosling (1:14:02.980)

was that they were doing all the usual computer things

Lex Fridman (1:14:07.300)

that people had been doing like 20 years before.

James Gosling (1:14:10.540)

The thing that really leapt out to me

Lex Fridman (1:14:13.100)

was that they were sort of reinventing

James Gosling (1:14:16.700)

computer networking.

Lex Fridman (1:14:18.820)

And they were making all the mistakes

James Gosling (1:14:24.340)

that people in the computer industry had made.

Lex Fridman (1:14:28.580)

And since I had been doing a lot of work in the networking

James Gosling (1:14:32.180)

area, we'd go and visit Company X.

Lex Fridman (1:14:37.140)

They'd describe this networking thing that they were doing.

Lex Fridman (1:14:40.540)

And just without any thought, I could tell them

Lex Fridman (1:14:43.260)

like the 25 things that were going

James Gosling (1:14:45.980)

to be complete disasters with that thing

Lex Fridman (1:14:49.020)

that they were doing.

Lex Fridman (1:14:52.220)

And I don't know whether that had any impact on any of them.

Lex Fridman (1:14:55.700)

But that particular story of repeating the disasters

James Gosling (1:15:03.180)

of the computer science industry was there.

Lex Fridman (1:15:07.940)

And one of the things we thought was, well,

James Gosling (1:15:11.300)

maybe we could do something useful here with bringing them

Lex Fridman (1:15:15.420)

forward somewhat.

Lex Fridman (1:15:16.380)

But also, at the same time, we learned a bunch of things

Lex Fridman (1:15:24.180)

from these mostly consumer electronics companies.

Lex Fridman (1:15:31.940)

And high on the list was that they

Lex Fridman (1:15:39.780)

viewed their relationship with the customer as sacred.

James Gosling (1:15:45.460)

They were never, ever willing to make tradeoffs

Lex Fridman (1:15:51.820)

between for safety.

Lex Fridman (1:15:56.540)

So one of the things that had always

Lex Fridman (1:16:00.300)

made me nervous in the computer industry

James Gosling (1:16:04.900)

was that people were willing to make tradeoffs in reliability

Lex Fridman (1:16:12.220)

to get performance.

James Gosling (1:16:17.540)

They want faster, faster.

Lex Fridman (1:16:18.740)

It breaks a little more often because it's fast.

James Gosling (1:16:21.460)

Maybe you run it a little hotter than you should.

Lex Fridman (1:16:23.940)

Or the one that always blew my mind

James Gosling (1:16:27.620)

was the way that the folks at Cray Supercomputers

Lex Fridman (1:16:34.100)

got their division to be really fast

James Gosling (1:16:38.780)

was that they did Newton Raphson approximations.

Lex Fridman (1:16:44.220)

And so the bottom several bits of A over B

James Gosling (1:16:51.420)

were essentially random numbers.

Lex Fridman (1:16:55.540)

What could possibly go wrong?

Lex Fridman (1:16:56.860)

What could go wrong?

Lex Fridman (1:16:59.700)

And just figuring out how to nail the bottom bit,

Lex Fridman (1:17:08.580)

how to make sure that if you put a piece of toast in a toaster,

Lex Fridman (1:17:14.700)

it's not going to kill the customer.

James Gosling (1:17:18.300)

It's not going to burst into flames and burn the house down.

Lex Fridman (1:17:22.540)

So I guess those are the principles that were inspiring.

Lex Fridman (1:17:26.860)

But how did, from the days of Java's called oak,

Lex Fridman (1:17:33.100)

because of a tree outside the window story

James Gosling (1:17:34.980)

that a lot of people know, how did it

Lex Fridman (1:17:37.100)

become this incredible, powerful language?

James Gosling (1:17:43.900)

Well, so it was a bunch of things.

Lex Fridman (1:17:46.780)

So after all that, the way that we

James Gosling (1:17:50.660)

decided that we could understand things better

Lex Fridman (1:17:54.500)

was by building a demo, building a prototype of something.

Lex Fridman (1:17:59.300)

So because it was easy and fun, we

Lex Fridman (1:18:02.740)

decided to build a control system for some home

James Gosling (1:18:07.220)

electronics, TV, VCR, that kind of stuff.

Lex Fridman (1:18:11.020)

And as we were building it, we discovered

James Gosling (1:18:16.580)

that there were some things about standard practice

Lex Fridman (1:18:19.660)

in C programming that were really getting in the way.

Lex Fridman (1:18:26.380)

And it wasn't exactly because we were writing all this C code

Lex Fridman (1:18:32.380)

and C++ code that we couldn't write it to do the right thing.

Lex Fridman (1:18:37.740)

But one of the things that was weird in the group

Lex Fridman (1:18:42.340)

was that we had a guy whose top level job was

James Gosling (1:18:50.860)

he was a business guy.

Lex Fridman (1:18:52.940)

He was an MBA kind of person, think about business plans

Lex Fridman (1:18:57.780)

and all of that.

Lex Fridman (1:18:58.660)

And there were a bunch of things that were kind of,

Lex Fridman (1:19:05.260)

and we would talk about things that were going wrong

Lex Fridman (1:19:07.540)

and things that were going wrong,

James Gosling (1:19:10.100)

things that were going right.

Lex Fridman (1:19:11.260)

And as we thought about things like the requirements

James Gosling (1:19:16.140)

for security and safety, some low level details

Lex Fridman (1:19:21.540)

and see like naked pointers.

James Gosling (1:19:24.060)

Yeah.

Lex Fridman (1:19:24.980)

And so back in the early 90s, it was well understood

James Gosling (1:19:35.300)

that the number one source of security vulnerabilities.

Lex Fridman (1:19:41.580)

Was pointers.

James Gosling (1:19:42.380)

Was just pointers, was just bugs.

Lex Fridman (1:19:45.980)

And it was like 50%, 60%, 70% of all security vulnerabilities

James Gosling (1:19:52.420)

were bugs.

Lex Fridman (1:19:53.300)

And the vast majority of them were like buffer overflows.

Lex Fridman (1:19:57.980)

So you're like, we have to fix this.

Lex Fridman (1:20:00.140)

We have to make sure that this cannot happen.

Lex Fridman (1:20:04.020)

And that was kind of the original thing for me

Lex Fridman (1:20:07.420)

was this cannot continue.

Lex Fridman (1:20:11.420)

And one of the things I find really entertaining this year

Lex Fridman (1:20:16.500)

was, I forget which Rag published it,

Lex Fridman (1:20:22.340)

but there was this article that came out

Lex Fridman (1:20:24.780)

that was sort of the result of an examination

James Gosling (1:20:31.220)

of all the security vulnerabilities in Chrome.

Lex Fridman (1:20:35.260)

And Chrome is like a giant piece of C++ code.

Lex Fridman (1:20:39.100)

And 60% or 70% of all the security vulnerabilities

Lex Fridman (1:20:44.100)

were stupid pointer tricks.

Lex Fridman (1:20:48.100)

And I thought, it's 30 years later and we're still there.

Lex Fridman (1:20:54.260)

Still there.

Lex Fridman (1:20:54.940)

And we're still there.

Lex Fridman (1:20:56.260)

And that's one of those slap your forehead

Lex Fridman (1:21:00.900)

and just want to cry moments.

Lex Fridman (1:21:05.220)

Would you attribute, or is that too much of a simplification,

Lex Fridman (1:21:09.060)

but would you attribute the creation of Java

Lex Fridman (1:21:11.300)

to C pointers?

James Gosling (1:21:15.460)

Obvious problem.

Lex Fridman (1:21:16.660)

Well, I mean, that was one of the trigger points.

James Gosling (1:21:21.820)

Concurrency you've mentioned.

Lex Fridman (1:21:23.420)

Concurrency was a big deal.

James Gosling (1:21:27.900)

Because when you're interacting with people,

Lex Fridman (1:21:31.100)

the last thing you ever want to see

James Gosling (1:21:32.740)

is the thing like waiting and issues

Lex Fridman (1:21:37.860)

about the software development process.

Lex Fridman (1:21:42.100)

When faults happen, can you recover from them?

Lex Fridman (1:21:45.820)

What can you do to make it easier

Lex Fridman (1:21:49.300)

to create and eliminate complex data structures?

Lex Fridman (1:21:54.700)

What can you do to fix one of the most common C problems,

Lex Fridman (1:21:59.940)

which is storage leaks?

Lex Fridman (1:22:03.620)

And it's evil twin, the freed but still being used

James Gosling (1:22:12.500)

piece of memory.

Lex Fridman (1:22:14.340)

You free something and then you keep using it.

James Gosling (1:22:17.780)

Oh, yeah.

Lex Fridman (1:22:19.660)

So when I was originally thinking about that,

James Gosling (1:22:21.580)

I was thinking about it in terms of safety and security issues.

Lex Fridman (1:22:26.700)

And one of the things I came to understand

James Gosling (1:22:29.900)

was that it wasn't just about safety and security,

Lex Fridman (1:22:33.260)

but it was about developer velocity.

Lex Fridman (1:22:39.180)

So and I got really religious about this

Lex Fridman (1:22:43.340)

because at that point, I had spent an ungodly amount

James Gosling (1:22:46.620)

of my life hunting down mystery pointer bugs.

Lex Fridman (1:22:54.900)

And two thirds of my time as a software developer

James Gosling (1:23:01.300)

was because the mystery pointer bugs tend

Lex Fridman (1:23:03.900)

to be the hardest to find because they tend

James Gosling (1:23:07.900)

to be very, very statistical.

Lex Fridman (1:23:11.820)

The ones that hurt, they're like a one in a million chance.

Lex Fridman (1:23:20.780)

But nevertheless, create an infinite amount of suffering.

Lex Fridman (1:23:23.420)

Right.

James Gosling (1:23:24.460)

Because when you're doing a billion operations a second,

Lex Fridman (1:23:28.740)

one in a million chance means it's going to happen.

Lex Fridman (1:23:35.260)

And so I got really religious about this thing,

Lex Fridman (1:23:38.420)

about making it so that if something fails,

James Gosling (1:23:41.180)

it fails immediately and visibly.

Lex Fridman (1:23:44.900)

And one of the things that was a real attraction of Java

James Gosling (1:23:52.780)

to lots of development shops was that we get our code up

Lex Fridman (1:23:57.220)

and running twice as fast.

James Gosling (1:24:00.900)

You mean like the entirety of the development process,

Lex Fridman (1:24:03.540)

debugging, all that kind of stuff?

James Gosling (1:24:04.900)

Yeah, so if you measure time from you first touch fingers

Lex Fridman (1:24:12.500)

to keyboard until you get your first demo out,

James Gosling (1:24:19.580)

not much different.

Lex Fridman (1:24:21.140)

But if you look from fingers touching keyboard

James Gosling (1:24:23.620)

to solid piece of software that you could release

Lex Fridman (1:24:27.900)

in production, it would be way faster.

Lex Fridman (1:24:32.180)

And I think what people don't often realize is, yeah,

Lex Fridman (1:24:34.740)

there's things that really slow you down,

James Gosling (1:24:36.780)

like the hard to catch bugs probably

Lex Fridman (1:24:40.260)

is the thing that really slows down the entire time.

James Gosling (1:24:43.580)

It really slows things down.

Lex Fridman (1:24:45.740)

But also, one of the things that you get out

James Gosling (1:24:48.540)

of object oriented programming is a strict methodology

Lex Fridman (1:24:52.940)

about what are the interfaces between things

Lex Fridman (1:24:55.860)

and being really clear about how parts relate to each other.

Lex Fridman (1:25:02.180)

And what that helps with is so many times

Lex Fridman (1:25:06.940)

what people do is they kind of like sneak around the side.

Lex Fridman (1:25:12.540)

So if you've built something and people are using it

Lex Fridman (1:25:16.980)

and you say, well, OK, I built this thing.

Lex Fridman (1:25:21.420)

You use it this way.

Lex Fridman (1:25:23.740)

And then you change it in such a way

Lex Fridman (1:25:26.180)

that it still does what you said it does.

James Gosling (1:25:28.540)

It just does it a little bit different.

Lex Fridman (1:25:30.260)

Then you find out that somebody out there

James Gosling (1:25:33.260)

was sneaking around the side.

Lex Fridman (1:25:35.660)

They sort of tunneled in a back door.

Lex Fridman (1:25:38.180)

And this person, their code broke.

Lex Fridman (1:25:41.860)

And because they were sneaking through a side door.

Lex Fridman (1:25:47.020)

And normally, the attitude is, dummy.

Lex Fridman (1:25:56.380)

But a lot of times, you can't just slap their hand

Lex Fridman (1:26:04.740)

and tell them to not do that.

Lex Fridman (1:26:06.740)

Because it's some bank's account reconciliation system

James Gosling (1:26:17.940)

that some developer decided, oh, I'm lazy.

Lex Fridman (1:26:22.780)

I'll just sneak through the back door.

James Gosling (1:26:24.900)

Because the language allows it.

Lex Fridman (1:26:26.420)

I mean, you can't even mad at them.

Lex Fridman (1:26:28.060)

And so one of the things I did that, on the one hand,

Lex Fridman (1:26:32.020)

upset a bunch of people was I made it

Lex Fridman (1:26:34.540)

so that you really couldn't go through back doors.

Lex Fridman (1:26:38.740)

So the whole point of that was to say,

James Gosling (1:26:44.340)

if the interface here isn't right,

Lex Fridman (1:26:47.140)

the wrong way to deal with that is to go through a back door.

James Gosling (1:26:51.380)

The right way to deal with it is to walk up

Lex Fridman (1:26:53.340)

to the developer of this thing and say, fix it.

Lex Fridman (1:26:58.860)

And so it was kind of like a social engineering thing.

Lex Fridman (1:27:02.300)

And people ended up discovering that that really

James Gosling (1:27:08.580)

made a difference in terms of.

Lex Fridman (1:27:12.340)

And a bunch of this stuff, if you're just screwing around

James Gosling (1:27:16.060)

writing your own class project scale stuff,

Lex Fridman (1:27:20.900)

a lot of this stuff isn't quite so important

James Gosling (1:27:25.460)

because you're both sides of the interface.

Lex Fridman (1:27:31.420)

But when you're building larger, more complex pieces of software

James Gosling (1:27:37.740)

that have a lot of people working on them,

Lex Fridman (1:27:39.940)

and especially when they span organizations,

James Gosling (1:27:46.380)

having clarity about how that stuff gets structured

Lex Fridman (1:27:52.860)

saves your life.

Lex Fridman (1:27:56.380)

And especially, there's so much software

Lex Fridman (1:28:00.180)

that is fundamentally untestable until you do the real thing.

James Gosling (1:28:08.140)

It's better to write good code in the beginning

Lex Fridman (1:28:11.980)

as opposed to writing crappy code

Lex Fridman (1:28:13.580)

and then trying to fix it and trying to scramble and figure

Lex Fridman (1:28:17.260)

out, and through testing, figure out where the bugs are.

James Gosling (1:28:20.300)

Yeah, it's like, which shortcut caused that rocket

Lex Fridman (1:28:27.980)

to not get where it was needed to go?

Lex Fridman (1:28:31.580)

So I think one of the most beautiful ideas philosophically

Lex Fridman (1:28:38.260)

and technically is of a virtual machine, a Java virtual machine.

James Gosling (1:28:45.220)

Again, I apologize to romanticize things,

Lex Fridman (1:28:47.380)

but how did the idea of the JVM come to be?

Lex Fridman (1:28:53.300)

How to you radical of an idea it is?

Lex Fridman (1:28:56.380)

Because it seems to me to be just a really interesting idea

James Gosling (1:29:01.460)

in the history of programming.

Lex Fridman (1:29:04.660)

And what is it?

Lex Fridman (1:29:05.660)

So the Java virtual machine, you can think of it

Lex Fridman (1:29:10.060)

in different ways because it was carefully designed

James Gosling (1:29:17.060)

to have different ways of viewing it.

Lex Fridman (1:29:19.820)

So one view of it that most people don't really realize

James Gosling (1:29:23.820)

is there is that you can view it as sort

Lex Fridman (1:29:30.580)

of an encoding of the abstract syntax tree

James Gosling (1:29:34.820)

in reverse Polish notation.

Lex Fridman (1:29:39.180)

I don't know if that makes any sense at all.

James Gosling (1:29:41.380)

I could explain it, and that would blow all of our time.

Lex Fridman (1:29:44.820)

But the other way to think of it and the way

James Gosling (1:29:47.620)

that it ends up being explained is

Lex Fridman (1:29:50.380)

that it's like the instruction set of an abstract machine

James Gosling (1:29:57.180)

that's designed such that you can translate

Lex Fridman (1:30:00.380)

that abstract machine to a physical machine.

Lex Fridman (1:30:03.820)

And the reason that that's important,

Lex Fridman (1:30:07.860)

so if you wind back to the early 90s

James Gosling (1:30:10.460)

when we were talking to all of these companies doing

Lex Fridman (1:30:15.260)

consumer electronics, and you talk to the purchasing people,

James Gosling (1:30:22.460)

there were interesting conversations with purchasing.

Lex Fridman (1:30:27.900)

So if you look at how these devices come together,

James Gosling (1:30:31.940)

they're sheet metal and gears and circuit boards

Lex Fridman (1:30:36.100)

and capacitors and resistors and stuff.

Lex Fridman (1:30:40.620)

And everything you buy has multiple sources.

Lex Fridman (1:30:46.180)

So you can buy a capacitor from here.

James Gosling (1:30:50.420)

You can buy a capacitor from there.

Lex Fridman (1:30:52.780)

And you've got kind of a market so that you can actually

James Gosling (1:30:57.700)

get a decent price for a capacitor.

Lex Fridman (1:31:03.340)

But CPUs, and particularly in the early 90s,

James Gosling (1:31:12.980)

CPUs were all different and all proprietary.

Lex Fridman (1:31:17.700)

So if you use the chip from Intel,

James Gosling (1:31:22.620)

you had to be an Intel customer till the end of time.

Lex Fridman (1:31:28.060)

Because if you wrote a bunch of software,

James Gosling (1:31:31.500)

when you wrote software using whatever technique you wanted,

Lex Fridman (1:31:36.780)

and C was particularly bad about this

James Gosling (1:31:39.580)

because there was a lot of properties

Lex Fridman (1:31:43.780)

of the underlying machine that came through.

Lex Fridman (1:31:48.340)

So you were stuck.

Lex Fridman (1:31:49.220)

So the code you wrote, you were stuck

James Gosling (1:31:50.660)

to that particular machine.

Lex Fridman (1:31:51.780)

You were stuck to that particular machine,

James Gosling (1:31:54.300)

which meant that they couldn't decide,

Lex Fridman (1:31:56.980)

you know, Intel is screwing us.

James Gosling (1:32:01.500)

I'll start buying chips from Bob's Better Chips.

Lex Fridman (1:32:08.220)

This drove the purchasing people absolutely insane

James Gosling (1:32:15.860)

that they were welded into this decision.

Lex Fridman (1:32:20.540)

And they would have to make this decision

James Gosling (1:32:22.980)

before the first line of software was written.

Lex Fridman (1:32:25.940)

That's funny that you're talking about the purchasing people.

Lex Fridman (1:32:28.140)

So there's one perspective, right?

Lex Fridman (1:32:31.300)

There's a lot of other perspectives

James Gosling (1:32:32.740)

that all probably hated this idea.

Lex Fridman (1:32:35.260)

Right.

Lex Fridman (1:32:36.060)

But from a technical aspect,

Lex Fridman (1:32:37.540)

just like the creation of an abstraction layer

James Gosling (1:32:41.740)

that's agnostic to the underlying machine

Lex Fridman (1:32:46.300)

from the perspective of the developer,

James Gosling (1:32:48.340)

I mean, that's brilliant.

Lex Fridman (1:32:50.100)

Right.

James Gosling (1:32:50.780)

Well, and so that's like across the spectrum

Lex Fridman (1:32:56.300)

of providers of chips.

Lex Fridman (1:32:58.500)

But then there's also the time thing

Lex Fridman (1:33:00.700)

because, you know, as you went from one generation

James Gosling (1:33:04.820)

to the next generation to the next generation,

Lex Fridman (1:33:06.820)

they were all different.

Lex Fridman (1:33:07.980)

And you would often have to rewrite your software.

Lex Fridman (1:33:10.180)

Oh, you mean generations of machines of different kinds?

James Gosling (1:33:14.740)

Yeah.

Lex Fridman (1:33:15.620)

So like one of the things that sucked about a year out

James Gosling (1:33:19.820)

of my life was when Sun went from the Motorola 68010

Lex Fridman (1:33:27.660)

processor to the 68020 processor.

James Gosling (1:33:31.660)

Then they had a number of differences.

Lex Fridman (1:33:33.820)

And one of them hit us really hard.

Lex Fridman (1:33:36.540)

And I ended up being the point guy

Lex Fridman (1:33:41.820)

on the worst case of where the new instruction cache

James Gosling (1:33:46.340)

architecture hurt us.

Lex Fridman (1:33:49.060)

Well, OK, so I mean, so one of this idea, I mean, OK.

Lex Fridman (1:33:53.300)

So yeah, you articulate a really clear fundamental problem

Lex Fridman (1:33:57.580)

in all of computing.

Lex Fridman (1:33:59.100)

But where do you get the guts to think

Lex Fridman (1:34:03.180)

we can actually solve this?

James Gosling (1:34:05.180)

You know, in our conversations with all of these vendors,

Lex Fridman (1:34:09.580)

these problems started to show up.

Lex Fridman (1:34:14.100)

And I kind of had this epiphany because it reminded me

Lex Fridman (1:34:23.500)

of a summer job that I had had in grad school.

Lex Fridman (1:34:31.220)

So back in grad school, my thesis advisor,

Lex Fridman (1:34:37.660)

well, I had two thesis advisors for bizarre reasons.

James Gosling (1:34:42.900)

One of them was a guy named Raj Reddy.

Lex Fridman (1:34:44.900)

The other one was Bob Sproul.

Lex Fridman (1:34:48.340)

And Raj, I love Raj.

Lex Fridman (1:34:53.020)

I really love both of them.

Lex Fridman (1:34:56.140)

So the department had bought a bunch of early workstations

Lex Fridman (1:35:06.300)

from a company called Three Rivers Computer Company.

Lex Fridman (1:35:11.220)

And Three Rivers Computer Company

Lex Fridman (1:35:13.340)

was a bunch of electrical engineers

James Gosling (1:35:15.260)

who wanted to do as little software as possible.

Lex Fridman (1:35:19.060)

So they knew that they'd need to have compilers and an OS

Lex Fridman (1:35:23.580)

and stuff like that.

Lex Fridman (1:35:24.380)

And they didn't want to do any of that.

Lex Fridman (1:35:26.660)

And they wanted to do that for as close to zero money

Lex Fridman (1:35:29.460)

as possible.

Lex Fridman (1:35:31.260)

So what they did was they built a machine whose instruction set

Lex Fridman (1:35:39.380)

was literally the byte code for UCSD Pascal, the P code.

Lex Fridman (1:35:50.420)

And so we had a bunch of software

Lex Fridman (1:35:56.860)

that was written for this machine.

Lex Fridman (1:36:02.900)

And for various reasons, the company

Lex Fridman (1:36:06.700)

wasn't doing terrifically well.

James Gosling (1:36:08.340)

We had all this software on these machines.

Lex Fridman (1:36:10.100)

And we wanted it to run on other machines, principally

James Gosling (1:36:13.260)

the VAX.

Lex Fridman (1:36:15.460)

And so Raj asked me if I could come up

James Gosling (1:36:23.140)

with a way to port all of this software from the PERC machines

Lex Fridman (1:36:32.580)

to VAXs.

Lex Fridman (1:36:35.260)

And I think what he had in mind was something

Lex Fridman (1:36:41.460)

that would translate from Pascal to C or Pascal to, actually,

James Gosling (1:36:50.740)

at those times, pretty much it was

Lex Fridman (1:36:52.460)

you could translate to C or C. And if you

James Gosling (1:36:55.780)

didn't like translate to C, you could translate to C.

Lex Fridman (1:37:00.820)

There was, it's like the Henry Ford, any color you want,

James Gosling (1:37:05.820)

just as long as it's black.

Lex Fridman (1:37:09.300)

And I went, that's really hard.

James Gosling (1:37:15.140)

That's a.

Lex Fridman (1:37:16.500)

And I noticed that, and I was looking at stuff.

Lex Fridman (1:37:20.660)

And I went, oh, I bet I could rewrite the P code

Lex Fridman (1:37:25.820)

into VAX assembly code.

Lex Fridman (1:37:29.580)

And then I started to realize that there

Lex Fridman (1:37:33.940)

were some properties of P code that

James Gosling (1:37:36.060)

made that really easy, some properties that

Lex Fridman (1:37:39.500)

made it really hard.

Lex Fridman (1:37:40.820)

So I ended up writing this thing that

Lex Fridman (1:37:42.940)

translated from P code on the Three Rivers PERCs

James Gosling (1:37:49.620)

into assembly code on the VAX.

Lex Fridman (1:37:53.820)

And I actually got higher quality code than the C compiler.

Lex Fridman (1:38:00.180)

And so everything just got really fast.

Lex Fridman (1:38:03.260)

It was really easy.

James Gosling (1:38:04.340)

It was like, wow, I thought that was a sleazy hack

Lex Fridman (1:38:08.660)

because I was lazy.

Lex Fridman (1:38:10.660)

And in actual fact, it worked really well.

Lex Fridman (1:38:13.980)

And I tried to convince people that that was maybe

James Gosling (1:38:17.780)

a good thesis topic.

Lex Fridman (1:38:18.780)

And nobody was, it was like, nah.

Lex Fridman (1:38:23.460)

Really?

Lex Fridman (1:38:23.980)

That's, I mean, it's kind of a brilliant idea, right?

James Gosling (1:38:29.580)

Or maybe you didn't have the, you

Lex Fridman (1:38:31.820)

weren't able to articulate the big picture of it.

James Gosling (1:38:34.460)

Yeah.

Lex Fridman (1:38:34.980)

And I think that was a key part.

Lex Fridman (1:38:39.940)

But so then clock comes forward a few years.

Lex Fridman (1:38:44.100)

And it's like, we've got to be able to,

James Gosling (1:38:48.980)

if they want to be able to switch

Lex Fridman (1:38:50.500)

from this weird microprocessor to that weird and totally

Lex Fridman (1:38:55.300)

different microprocessor, how do you do that?

Lex Fridman (1:38:58.460)

And I kind of went, oh, maybe by doing something kind of

James Gosling (1:39:04.540)

in the space of Pascal P code, I could do multiple translations

Lex Fridman (1:39:12.260)

to multiple translators.

Lex Fridman (1:39:15.460)

And I spent some time thinking about that

Lex Fridman (1:39:17.260)

and thinking about what worked and what didn't work

James Gosling (1:39:19.780)

when I did the P code to Vax translator.

Lex Fridman (1:39:25.620)

And I talked to some of the folks

James Gosling (1:39:29.460)

who were involved in Smalltalk because Smalltalk also

Lex Fridman (1:39:31.980)

did a bytecode.

Lex Fridman (1:39:34.780)

And then I kind of went, yeah, I want to do that.

Lex Fridman (1:39:41.300)

Because that actually, and it had the other advantage

James Gosling (1:39:44.940)

that you could either interpret it or compile it.

Lex Fridman (1:39:49.900)

And interpreters are usually easier to do,

Lex Fridman (1:39:55.940)

but not as fast as a compiler.

Lex Fridman (1:40:00.060)

So I figured, good, I can be lazy again.

James Gosling (1:40:06.820)

Sometimes I think that most of my good ideas

Lex Fridman (1:40:08.940)

are driven by laziness.

Lex Fridman (1:40:12.340)

And often I find that some of people's stupidest ideas

Lex Fridman (1:40:15.860)

are because they're insufficiently lazy.

James Gosling (1:40:21.660)

They just want to build something really complicated.

Lex Fridman (1:40:23.900)

And it's like, it doesn't need to be that complicated.

James Gosling (1:40:26.740)

Yeah, and so that's how that came out.

Lex Fridman (1:40:30.380)

But that also turned into almost a religious position

James Gosling (1:40:39.140)

on my part, which got me in several other fights.

Lex Fridman (1:40:44.300)

So one of the things that was a real difference

James Gosling (1:40:49.020)

was the way that arithmetic worked.

Lex Fridman (1:40:51.020)

And once upon a time, it wasn't always just

James Gosling (1:40:59.620)

two's complement arithmetic.

Lex Fridman (1:41:01.540)

There were some machines that had one's complement

James Gosling (1:41:03.580)

arithmetic, which was like almost anything built by CDC.

Lex Fridman (1:41:08.820)

And occasionally, there were machines

James Gosling (1:41:10.660)

that were decimal arithmetic.

Lex Fridman (1:41:13.100)

And I was like, this is crazy.

James Gosling (1:41:17.860)

Pretty much two's complement integer arithmetic has one.

Lex Fridman (1:41:22.580)

So just do that.

James Gosling (1:41:28.180)

One of the other places where there was a lot of variability

Lex Fridman (1:41:31.180)

was in the way that floating point behaved.

Lex Fridman (1:41:36.180)

And that was causing people throughout the software

Lex Fridman (1:41:40.780)

industry much pain because you couldn't

James Gosling (1:41:44.500)

do a numerical computing library that

Lex Fridman (1:41:47.380)

would work on CDC and then have it work on an IBM machine

Lex Fridman (1:41:50.860)

and work on a DEC machine.

Lex Fridman (1:41:54.700)

And as a part of that whole struggle,

James Gosling (1:41:57.700)

there had been this big body of work on floating point

Lex Fridman (1:42:03.900)

standards.

Lex Fridman (1:42:05.300)

And this thing emerged that came to be called IEEE 754,

Lex Fridman (1:42:11.660)

which is the floating point standard that pretty much

James Gosling (1:42:14.580)

has taken over the entire universe.

Lex Fridman (1:42:20.060)

And at the time I was doing Java,

James Gosling (1:42:21.860)

it had pretty much completed taking over the universe.

Lex Fridman (1:42:25.740)

There were still a few pockets of holdouts,

Lex Fridman (1:42:28.180)

but I was like, it's important to be

Lex Fridman (1:42:32.100)

able to say what two plus two means.

Lex Fridman (1:42:37.500)

And so I went that.

Lex Fridman (1:42:42.380)

And one of the ways that I got into fights with people

James Gosling (1:42:46.180)

was that there were a few machines that did not

Lex Fridman (1:42:50.380)

implement IEEE 754 correctly.

James Gosling (1:42:55.300)

Well, of course, that's all short term kind of fights.

Lex Fridman (1:42:58.500)

I think in the long term, I think this vision is one out.

James Gosling (1:43:03.020)

Yeah, and I think it's worked out over time.

Lex Fridman (1:43:06.180)

I mean, the biggest fights were with Intel

James Gosling (1:43:10.700)

because they had done some strange things with rounding.

Lex Fridman (1:43:15.380)

They'd done some strange things with their transcendental

James Gosling (1:43:17.980)

functions, which turned into a mushroom cloud of weirdness.

Lex Fridman (1:43:24.860)

And in the name of optimization, but from the perspective

James Gosling (1:43:29.100)

of the developer, that's not good.

Lex Fridman (1:43:32.500)

Well, their issues with transcendental functions

James Gosling (1:43:34.580)

were just stupid.

Lex Fridman (1:43:35.740)

OK, so that's not even a trade off.

James Gosling (1:43:39.340)

That's just absolutely.

Lex Fridman (1:43:40.740)

Yeah, they were doing range reduction for sine and cosine

James Gosling (1:43:45.620)

using a slightly wrong value for pi.

Lex Fridman (1:43:48.500)

Got it.

James Gosling (1:43:49.580)

We've got 10 minutes.

Lex Fridman (1:43:50.580)

So in the interest of time, two questions.

Lex Fridman (1:43:53.460)

So one about Android and one about life.

Lex Fridman (1:43:58.100)

So one, I mean, we could talk for many more hours.

James Gosling (1:44:02.300)

I hope eventually we might talk again.

Lex Fridman (1:44:05.140)

But I got to ask you about Android and the use of Java

James Gosling (1:44:09.220)

there because it's one of the many places where Java just

Lex Fridman (1:44:14.220)

has a huge impact on this world.

James Gosling (1:44:16.380)

Just on your opinion, is there things

Lex Fridman (1:44:19.780)

that make you happy about the way Java

Lex Fridman (1:44:24.300)

is used in the Android world?

Lex Fridman (1:44:25.860)

And are there things that you wish were different?

James Gosling (1:44:29.580)

I don't know how to do a short answer to that.

Lex Fridman (1:44:32.660)

But I have to do a short answer to that.

Lex Fridman (1:44:34.500)

So I'm happy that they did it.

Lex Fridman (1:44:38.980)

Java had been running on cell phones

James Gosling (1:44:40.940)

at that time for quite a few years.

Lex Fridman (1:44:42.540)

And it worked really, really well.

James Gosling (1:44:46.020)

There were things about how they did it.

Lex Fridman (1:44:49.260)

And in particular, various ways that they kind of violated

James Gosling (1:44:59.420)

all kinds of contracts.

Lex Fridman (1:45:00.540)

The guy who led it, Andy Rubin, he crossed a lot of lines.

James Gosling (1:45:06.380)

There's some lines crossed.

Lex Fridman (1:45:07.700)

Yeah, lines were crossed that have since mushroomed

James Gosling (1:45:11.820)

into giant court cases.

Lex Fridman (1:45:16.620)

And they didn't need to do that.

Lex Fridman (1:45:19.740)

And in fact, it would have been so much cheaper for them

Lex Fridman (1:45:23.220)

to not cross lines.

James Gosling (1:45:25.260)

I mean, I suppose they didn't anticipate

Lex Fridman (1:45:28.260)

the success of this whole endeavor.

James Gosling (1:45:34.540)

Or do you think at that time it was already clear

Lex Fridman (1:45:36.500)

that this is going to blow up?

James Gosling (1:45:38.820)

I guess I sort of came to believe

Lex Fridman (1:45:41.900)

that it didn't matter what Andy did,

James Gosling (1:45:44.420)

it was going to blow up.

Lex Fridman (1:45:50.500)

I kind of started to think of him as a manufacturer of bombs.

James Gosling (1:45:58.380)

Yeah, some of the best things in this world

Lex Fridman (1:46:01.500)

come about through a little bit of explosive.

James Gosling (1:46:05.820)

Well, and some of the worst.

Lex Fridman (1:46:07.180)

And some of the worst, beautifully put.

Lex Fridman (1:46:11.420)

And like you said, I mean, does that

Lex Fridman (1:46:13.260)

make you proud that Java is in millions?

James Gosling (1:46:19.140)

I mean, it could be billions of devices.

Lex Fridman (1:46:21.660)

Yeah, well, I mean, it was in billions of phones

James Gosling (1:46:24.940)

before Android came along.

Lex Fridman (1:46:26.300)

And I'm just as proud of the way that the smart card standards

James Gosling (1:46:37.380)

adopted Java.

Lex Fridman (1:46:39.020)

And everybody involved in that did a really good job.

Lex Fridman (1:46:43.100)

And that's billions and billions.

Lex Fridman (1:46:48.300)

That's crazy.

James Gosling (1:46:49.300)

The SIM cards, the SIM cards in your pocket.

Lex Fridman (1:46:54.220)

I've been outside of that world for a decade.

Lex Fridman (1:46:56.860)

So I don't know how that has evolved.

Lex Fridman (1:46:59.980)

But it's just been crazy.

Lex Fridman (1:47:04.340)

So on that topic, let me ask, again,

Lex Fridman (1:47:07.940)

there's a million technical things we could talk about.

Lex Fridman (1:47:12.260)

But let me ask the absurd, the old philosophical question

Lex Fridman (1:47:16.620)

about life.

Lex Fridman (1:47:20.060)

What do you hope when you look back at your life

Lex Fridman (1:47:23.060)

and the people talk about you, write about you 500 years

Lex Fridman (1:47:28.140)

from now, what do you hope your legacy is?

Lex Fridman (1:47:34.460)

People not being afraid to take a leap of faith.

James Gosling (1:47:39.820)

I mean, I've got this kind of weird history

Lex Fridman (1:47:44.340)

of doing weird stuff.

Lex Fridman (1:47:45.580)

And it worked out pretty damn well.

Lex Fridman (1:47:49.220)

It worked out.

Lex Fridman (1:47:50.820)

And I think some of the weirder stuff that I've done

Lex Fridman (1:47:55.820)

has been the coolest.

Lex Fridman (1:47:56.940)

And some of it crashed and burned.

Lex Fridman (1:47:59.940)

And I think well over half of the stuff that I've done

James Gosling (1:48:05.860)

has crashed and burned, which has occasionally

Lex Fridman (1:48:09.540)

been really annoying.

Lex Fridman (1:48:12.060)

But still, you kept doing it.

Lex Fridman (1:48:13.580)

But yeah.

James Gosling (1:48:15.060)

Yeah.

Lex Fridman (1:48:15.540)

Yeah.

Lex Fridman (1:48:16.020)

And even when things crash and burn,

Lex Fridman (1:48:19.340)

you at least learn something from it.

James Gosling (1:48:22.340)

By way of advice, people, developers, engineers,

Lex Fridman (1:48:27.180)

scientists, or just people who are young, to look up to you,

Lex Fridman (1:48:32.100)

what advice would you give them how to approach their life?

Lex Fridman (1:48:37.660)

Don't be afraid of risk.

James Gosling (1:48:39.620)

It's OK to do stupid things once.

Lex Fridman (1:48:41.900)

Maybe a couple of times you get a pass on the first time

James Gosling (1:48:50.580)

or two that you do something stupid.

Lex Fridman (1:48:53.260)

The third or fourth time, yeah, not so much.

Lex Fridman (1:48:59.660)

But also, I don't know why, but really early on,

Lex Fridman (1:49:07.420)

I started to think about ethical choices in my life.

Lex Fridman (1:49:14.220)

And because I'm a big science fiction fan,

Lex Fridman (1:49:20.220)

I got to thinking about just about every technical decision

James Gosling (1:49:25.700)

I make in terms of, are you building

Lex Fridman (1:49:30.340)

Blade Runner or Star Trek?

Lex Fridman (1:49:33.700)

Which one's better?

Lex Fridman (1:49:34.540)

Which future would you rather live in?

Lex Fridman (1:49:38.820)

So what's the answer to that?

Lex Fridman (1:49:40.380)

Well, I would sure rather live in the universe of Star Trek.

James Gosling (1:49:44.500)

Star Trek, yeah.

Lex Fridman (1:49:46.100)

That opens up a whole topic about AI,

Lex Fridman (1:49:48.180)

but that's a really interesting idea.

Lex Fridman (1:49:53.140)

So your favorite AI system would be data from Star Trek.

Lex Fridman (1:49:58.140)

And my least favorite would easily be Skynet.

Lex Fridman (1:50:00.700)

Yeah.

James Gosling (1:50:02.340)

Beautifully put.

Lex Fridman (1:50:03.060)

I don't think there's a better way to end it, James.

James Gosling (1:50:05.700)

I can't say enough how much of an honor

Lex Fridman (1:50:08.100)

it is to meet you, to talk to you.

James Gosling (1:50:09.540)

Thanks so much for wasting your time with me today.

Lex Fridman (1:50:12.620)

Not a waste at all.

James Gosling (1:50:13.940)

Thanks, James.

Lex Fridman (1:50:14.660)

All right, thanks.

James Gosling (1:50:16.580)

Thanks for listening to this conversation with James

Lex Fridman (1:50:18.940)

Gosling, and thank you to our sponsors, Public Goods,

James Gosling (1:50:21.780)

BetterHelp, and ExpressVPN.

Lex Fridman (1:50:24.300)

Please check out these sponsors in the description

James Gosling (1:50:26.580)

to get a discount and to support this podcast.

Lex Fridman (1:50:30.460)

If you enjoy this thing, subscribe on YouTube,

James Gosling (1:50:32.780)

review it with five stars on Apple Podcast,

Lex Fridman (1:50:34.940)

follow on Spotify, support on Patreon,

James Gosling (1:50:37.660)

or connect with me on Twitter at Lex Friedman.

Lex Fridman (1:50:40.780)

And now, let me leave you with some words from James Gosling.

James Gosling (1:50:44.460)

One of the toughest things about life is making choices.

Lex Fridman (1:50:49.540)

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

Lex Fridman (20:00.140)

But getting a power supply that would do that voltage

Lex Fridman (20:04.540)

was pretty hard.

Lex Fridman (20:06.260)

And since I had a bunch of trashed television sets,

Lex Fridman (20:10.180)

I had to sort of cobble together something

James Gosling (20:16.140)

that was wrong but worked.

Lex Fridman (20:21.380)

So I was actually running these relays at 300 volts.

Lex Fridman (20:26.740)

And none of the electrical connections

Lex Fridman (20:30.460)

were like properly sealed off.

James Gosling (20:33.460)

Surprised you survived that period of your life.

Lex Fridman (20:36.380)

Oh, for so many reasons.

James Gosling (20:38.940)

For so many reasons.

Lex Fridman (20:40.420)

I mean, it's pretty common for teenage geeks

James Gosling (20:45.380)

to discover, oh, thermite.

Lex Fridman (20:47.380)

That's real easy to make.

James Gosling (20:50.460)

Yeah.

Lex Fridman (20:51.220)

Well, I'm glad you did.

Lex Fridman (20:52.740)

But do you remember what program in Calgary

Lex Fridman (20:58.820)

that you wrote, anything that stands out?

Lex Fridman (21:01.820)

And what language?

Lex Fridman (21:03.700)

Well, so mostly anything of any size was assembly code.

Lex Fridman (21:14.780)

And actually, before I learned assembly code,

Lex Fridman (21:17.420)

there was this programming language

James Gosling (21:19.100)

on the PDP 8 called Focal 5.

Lex Fridman (21:22.660)

And Focal 5 was kind of like a really stripped down Fortran.

Lex Fridman (21:28.540)

And I remember building programs that did things

Lex Fridman (21:33.580)

like play blackjack or solitaire.

James Gosling (21:40.940)

Or for some reason or other, the things that I really liked

Lex Fridman (21:45.180)

were ones where they were just like plotting graphs.

Lex Fridman (21:50.300)

So something with like a function or data,

Lex Fridman (21:53.940)

and then you plot it.

James Gosling (21:55.420)

Yeah.

Lex Fridman (21:56.900)

Yeah, I did a bunches of those things

Lex Fridman (21:59.300)

and went, ooh, pretty pictures.

Lex Fridman (22:03.100)

And so this would like print out, again, no monitors.

James Gosling (22:07.340)

Right, so it was like on a teletype.

Lex Fridman (22:11.820)

Yeah.

Lex Fridman (22:13.260)

So it's using something that's kind of like a typewriter.

Lex Fridman (22:18.620)

And then using those to plot functions.

Lex Fridman (22:23.140)

So when, I apologize to romanticize things,

Lex Fridman (22:26.100)

but when did you first fall in love with programming?

Lex Fridman (22:32.300)

What was the first programming language?

Lex Fridman (22:34.180)

Like as a serious, maybe, software engineer,

Lex Fridman (22:36.220)

where you thought this is a beautiful thing?

Lex Fridman (22:40.460)

I guess I never really thought of any particular language

James Gosling (22:43.300)

as being beautiful, because it was never really

Lex Fridman (22:46.620)

about the language for me.

James Gosling (22:47.820)

It was about what you could do with it.

Lex Fridman (22:49.540)

And even today, people try to get me

James Gosling (22:55.660)

into arguments about particular forms of syntax

Lex Fridman (22:59.740)

for this or that.

Lex Fridman (23:00.460)

And I'm like, who cares?

Lex Fridman (23:03.380)

It's about what you can do, not how you spell the word.

Lex Fridman (23:10.340)

And so back in those days, I learned like PL1 and Fortran

Lex Fridman (23:16.660)

and COBOL, and by the time that people were willing to hire me

James Gosling (23:23.260)

to do stuff, it was mostly assembly code and PDP assembly

Lex Fridman (23:29.500)

code and Fortran code and control data assembly

James Gosling (23:35.820)

code for the CDC 6400, which was an early, I guess,

Lex Fridman (23:40.620)

supercomputer.

James Gosling (23:43.100)

Even though that supercomputer has less compute power

Lex Fridman (23:46.340)

than my phone by a lot.

Lex Fridman (23:50.540)

And that was mostly, like you said, Fortran world.

Lex Fridman (23:54.820)

That said, you've also showed appreciation

James Gosling (23:57.380)

for the greatest language ever that I

Lex Fridman (24:01.580)

think everyone agrees is Lisp.

James Gosling (24:05.020)

Well, Lisp is definitely on my list of the greatest ones

Lex Fridman (24:09.100)

that have existed.

Lex Fridman (24:11.700)

Is it at number one?

Lex Fridman (24:12.740)

Or I mean, are you, I mean?

James Gosling (24:16.140)

You know, the thing is that it's, you know,

Lex Fridman (24:21.060)

I wouldn't put it number one, no.

Lex Fridman (24:23.380)

Is it the parentheses?

Lex Fridman (24:24.620)

What do you not love about Lisp?

James Gosling (24:34.180)

Well, I guess the number one thing

Lex Fridman (24:35.780)

to not love about it is so freaking many parentheses.

James Gosling (24:41.020)

On the love thing is, you know, out of those tons

Lex Fridman (24:46.460)

of parentheses, you actually get an interesting language

James Gosling (24:50.180)

structure.

Lex Fridman (24:51.820)

And I've always thought that there was a friendlier version

James Gosling (24:54.300)

of Lisp hiding out there somewhere.

Lex Fridman (24:57.940)

But I've never really spent much time thinking about it.

James Gosling (25:03.140)

But, you know, so like up the food chain for me from Lisp

Lex Fridman (25:10.460)

is Simula, which a very small number of people

James Gosling (25:14.420)

have ever used.

Lex Fridman (25:16.140)

But a lot of people, I think you had a huge influence, right,

James Gosling (25:19.460)

on the programming.

Lex Fridman (25:20.620)

But in the Simula, I apologize if I'm wrong on this,

Lex Fridman (25:24.900)

but is that one of the first functional languages?

Lex Fridman (25:27.780)

Or no?

James Gosling (25:28.300)

No, it was the first object oriented programming language.

Lex Fridman (25:32.780)

It's really where object oriented and languages sort

James Gosling (25:36.540)

of came together.

Lex Fridman (25:39.700)

And it was also the language where coroutines first showed

James Gosling (25:46.060)

up as a part of the language.

Lex Fridman (25:48.620)

So you could have a programming style that was,

James Gosling (25:53.900)

you could think of it as sort of multi threaded

Lex Fridman (25:57.740)

with a lot of parallelism.

Lex Fridman (26:00.900)

Really?

Lex Fridman (26:01.620)

There's ideas of parallelism in there?

James Gosling (26:03.580)

Yeah.

Lex Fridman (26:05.340)

Yeah, so that was back, you know,

Lex Fridman (26:08.220)

so the first Simula spec was Simula 67.

Lex Fridman (26:12.780)

Like 1967?

James Gosling (26:14.740)

Yeah.

Lex Fridman (26:15.780)

Wow.

Lex Fridman (26:16.420)

So it had coroutines, which are almost threads.

Lex Fridman (26:21.980)

The thing about coroutines is that they

James Gosling (26:24.180)

don't have true concurrency.

Lex Fridman (26:26.220)

So you can get away without really complex locking.

James Gosling (26:32.220)

You can't useably do coroutines on the multi core machine.

Lex Fridman (26:39.780)

Or if you try to do coroutines on the multi core machine,

James Gosling (26:44.100)

you don't actually get to use the multiple cores.

Lex Fridman (26:48.460)

Either that or you, because you start then

James Gosling (26:52.340)

having to get into the universe of semaphores and locks

Lex Fridman (26:56.940)

and things like that.

Lex Fridman (27:00.060)

But in terms of the style of programming,

Lex Fridman (27:04.700)

you could write code and think of it as being multi threaded.

James Gosling (27:11.460)

The mental model was very much a multi threaded one.

Lex Fridman (27:15.860)

And all kinds of problems you could

James Gosling (27:18.820)

approach very differently.

Lex Fridman (27:22.420)

To return to the world of Lisp for a brief moment,

James Gosling (27:27.940)

at CMU you wrote a version of Emacs

Lex Fridman (27:32.940)

that I think was very impactful on the history of Emacs.

Lex Fridman (27:37.660)

What was your motivation for doing so?

Lex Fridman (27:42.220)

At that time, so that was in like 85 or 86.

James Gosling (27:52.820)

I had been using Unix for a few years.

Lex Fridman (27:57.500)

And most of the editing was this tool

James Gosling (28:02.580)

called ED, which was sort of an ancestor of VI.

Lex Fridman (28:10.940)

Is it a pretty good editor, not a good editor?

James Gosling (28:14.060)

Well, if what you're using, if your input device

Lex Fridman (28:19.220)

is a teletype, it's pretty good.

James Gosling (28:22.620)

It's certainly more humane than TECO,

Lex Fridman (28:25.740)

which was kind of the common thing

James Gosling (28:28.180)

in a lot of the DEC universe at the time.

Lex Fridman (28:32.580)

TECO is spelled TK, is that the?

James Gosling (28:35.060)

No, TECO, T E C O, the text editor and corrector.

Lex Fridman (28:39.620)

Corrector, wow, so many features.

Lex Fridman (28:44.100)

And the original Emacs came out as,

Lex Fridman (28:48.740)

so Emacs stands for editor macros.

Lex Fridman (28:52.060)

And TECO had a way of writing macros.

Lex Fridman (28:56.500)

And so the original Emacs from MIT

James Gosling (29:02.780)

started out as a collection of macros for TECO.

Lex Fridman (29:07.060)

But then the Emacs style got popular originally at MIT.

Lex Fridman (29:15.060)

And then people did a few other implementations

Lex Fridman (29:20.300)

of Emacs that were, the code base was entirely different,

Lex Fridman (29:26.180)

but it was sort of the philosophical style

Lex Fridman (29:28.780)

of the original Emacs.

Lex Fridman (29:30.700)

What was the philosophy of Emacs?

Lex Fridman (29:33.260)

And by the way, were all the implementations always in C?

Lex Fridman (29:36.580)

And then how does Lisp fit into the picture?

Lex Fridman (29:39.700)

No, so the very first Emacs was written

James Gosling (29:43.300)

as a bunch of macros for the TECO text editor.

Lex Fridman (29:46.420)

Wow, that's so interesting.

Lex Fridman (29:47.780)

And the macro language for TECO was probably

Lex Fridman (29:54.220)

the most ridiculously obscure format.

James Gosling (29:59.100)

If you just look at a TECO program on a page,

Lex Fridman (30:02.420)

you think it was just random characters.

James Gosling (30:05.900)

It really looks like just line noise.

Lex Fridman (30:09.580)

So it's kind of like LaTeX or something.

James Gosling (30:11.380)

Oh, way worse than LaTeX.

Lex Fridman (30:14.860)

Way, way worse than LaTeX.

Lex Fridman (30:18.100)

But if you use TECO a lot, which I did,

Lex Fridman (30:23.700)

TECO was completely optimized for touch typing at high speed.

Lex Fridman (30:31.180)

So there were no two character commands.

Lex Fridman (30:35.300)

Well, there were a few, but mostly they

James Gosling (30:37.980)

were just one character.

Lex Fridman (30:39.220)

So every character on the keyboard was a separate command.

Lex Fridman (30:43.780)

And actually, every character on the keyboard

Lex Fridman (30:45.860)

was usually two or three commands

James Gosling (30:48.180)

because you could hit Shift and Control and all

Lex Fridman (30:51.900)

of those things.

James Gosling (30:52.700)

It's just a way of very tightly encoding it.

Lex Fridman (30:56.420)

And mostly what Emacs did was it made that visual.

Lex Fridman (31:03.660)

So one way to think of TECO is use Emacs with your eyes

Lex Fridman (31:10.140)

closed, where you have to maintain

James Gosling (31:15.500)

a mental model of a mental image of your document.

Lex Fridman (31:20.660)

You have to go, OK, so the cursor is between the A and the E.

Lex Fridman (31:27.780)

And I want to exchange those, so I do these things.

Lex Fridman (31:31.860)

So it is almost exactly the Emacs command set.

James Gosling (31:38.780)

Well, it's roughly the same as Emacs command set,

Lex Fridman (31:42.820)

but using Emacs with your eyes closed.

Lex Fridman (31:47.780)

So part of what Emacs added to the whole thing

Lex Fridman (31:52.940)

was being able to visually see what

James Gosling (31:56.700)

you were editing in a form that matched your document.

Lex Fridman (32:02.180)

And a lot of things changed in the command set because it

James Gosling (32:12.420)

was programmable.

Lex Fridman (32:13.820)

It was really flexible.

James Gosling (32:15.980)

You could add new commands for all kinds of things.

Lex Fridman (32:18.500)

And then people rewrote Emacs multiple times in Lisp.

James Gosling (32:24.700)

There was one done at MIT for the Lisp machine.

Lex Fridman (32:28.180)

There was one done for Multics.

Lex Fridman (32:31.420)

And one summer, I got a summer job

Lex Fridman (32:35.100)

to work on the Pascal compiler for Multics.

Lex Fridman (32:40.220)

And that was actually the first time I used Emacs.

Lex Fridman (32:47.540)

To write the compilers.

James Gosling (32:48.900)

You've worked on compilers, too.

Lex Fridman (32:50.460)

That's fascinating.

James Gosling (32:52.180)

Yeah, so I did a lot of work.

Lex Fridman (32:55.940)

I spent a really intense three months

James Gosling (32:59.580)

working on this Pascal compiler, basically living in Emacs.

Lex Fridman (33:05.300)

And it was the one written in Mac Lisp by Bernie Greenberg.

Lex Fridman (33:11.260)

And I thought, wow, this is just a way better way

Lex Fridman (33:15.420)

to do editing.

Lex Fridman (33:19.060)

And then I got back to CMU, where

Lex Fridman (33:22.980)

we had one of everything and two of a bunch of things

Lex Fridman (33:28.700)

and four of a few things.

Lex Fridman (33:30.460)

And since I mostly worked in the Unix universe,

Lex Fridman (33:36.140)

and Unix didn't have an Emacs, I decided

Lex Fridman (33:39.340)

that I needed to fix that problem.

Lex Fridman (33:43.420)

So I wrote this implementation of Emacs in C,

Lex Fridman (33:47.620)

because at the time, C was really the only language that

James Gosling (33:51.140)

worked on Unix.

Lex Fridman (33:55.820)

And you were comfortable with C as well at that point?

James Gosling (33:58.660)

Yeah, at that time, I had done a lot of C coding.

Lex Fridman (34:02.380)

This was in, like, 86.

Lex Fridman (34:06.820)

And it was running well enough for me

Lex Fridman (34:14.100)

to use it to edit itself within a month or two.

Lex Fridman (34:17.380)

And then it kind of took over the university.

Lex Fridman (34:23.580)

And it spread outside.

James Gosling (34:25.380)

Yeah, and then it went outside.

Lex Fridman (34:28.100)

And largely because Unix kind of took over the research

James Gosling (34:33.060)

community on the ARPANET, and Emacs

Lex Fridman (34:39.700)

was kind of the best editor out there.

James Gosling (34:43.220)

It kind of took over.

Lex Fridman (34:44.300)

And there was actually a brief period

James Gosling (34:47.940)

where I actually had login IDs on every nonmilitary host

Lex Fridman (34:55.300)

on the ARPANET.

Lex Fridman (34:59.140)

Because people would say, oh, can we install this?

Lex Fridman (35:01.420)

And I'd like, well, yeah, but you'll need some help.

James Gosling (35:09.700)

The days when security wasn't.

Lex Fridman (35:11.860)

When nobody cared.

James Gosling (35:12.820)

Nobody cared.

Lex Fridman (35:15.500)

I can ask briefly, what were those early days of ARPANET

Lex Fridman (35:20.180)

and the internet like?

Lex Fridman (35:25.780)

Did you, again, sorry for the silly question,

Lex Fridman (35:28.820)

but could you have possibly imagined

Lex Fridman (35:31.940)

that the internet would look like what it is today?

James Gosling (35:36.940)

Some of it is remarkably unchanged.

Lex Fridman (35:42.420)

So one of the things that I noticed really early on

James Gosling (35:49.420)

when I was at Carnegie Mellon was

Lex Fridman (35:51.980)

that a lot of social life became centered around the ARPANET.

Lex Fridman (36:00.740)

So things like between email and text messaging.

Lex Fridman (36:06.380)

Because text messaging was a part of the ARPANET

James Gosling (36:09.620)

really early on.

Lex Fridman (36:11.660)

There were no cell phones, but you're sitting at a terminal

Lex Fridman (36:15.100)

and you're typing stuff.

Lex Fridman (36:16.820)

So essentially email, or what is text messaging?

James Gosling (36:19.860)

Well, just like a one line message.

Lex Fridman (36:23.740)

Oh, cool.

Lex Fridman (36:24.260)

So like chat.

Lex Fridman (36:25.540)

Like chat.

Lex Fridman (36:27.660)

So it's like sending a one line message to somebody.

Lex Fridman (36:31.820)

And so pretty much everything from arranging lunch

James Gosling (36:40.020)

to going out on dates was all like driven by social media.

Lex Fridman (36:48.100)

Social media.

James Gosling (36:49.700)

Right, in the 80s.

Lex Fridman (36:52.860)

Easier than phone calls, yeah.

Lex Fridman (36:55.420)

And my life had gotten to where I

Lex Fridman (37:00.660)

was living on social media from the early mid 80s.

Lex Fridman (37:11.180)

And so when it sort of transformed into the internet

Lex Fridman (37:16.620)

and social media explodes, I was kind of like,

Lex Fridman (37:19.820)

what's the big deal?

Lex Fridman (37:21.900)

It's just a scale thing.

James Gosling (37:24.140)

Right, the scale thing is just astonishing.

Lex Fridman (37:29.060)

But the fundamentals in some ways remain the same.

James Gosling (37:32.380)

The fundamentals have hardly changed.

Lex Fridman (37:36.020)

And the technologies behind the networking

James Gosling (37:40.860)

have changed significantly.

Lex Fridman (37:42.460)

The watershed moment of going from the ARPANET

James Gosling (37:49.460)

to the internet.

Lex Fridman (37:52.580)

And then people starting to just scale and scale and scale.

James Gosling (37:57.300)

I mean, the scaling that happened in the early 90s

Lex Fridman (38:03.660)

and the way that so many vested interests fought the internet.

James Gosling (38:12.620)

Oh, interesting.

Lex Fridman (38:14.220)

What was the, oh, because you can't really

James Gosling (38:16.380)

control the internet.

Lex Fridman (38:17.980)

Yeah, so who fought the internet?

Lex Fridman (38:19.900)

So fundamentally, the cable TV companies

Lex Fridman (38:25.900)

and broadcasters and phone companies,

James Gosling (38:32.940)

at the deepest fibers of their being, they hated the internet.

Lex Fridman (38:38.300)

But it was often kind of a funny thing because, so think

James Gosling (38:50.580)

of a cable company.

Lex Fridman (38:53.900)

Most of the employees of a cable company,

James Gosling (38:57.220)

their job is getting TV shows, movies, whatever,

Lex Fridman (39:04.460)

out to their customers.

James Gosling (39:06.980)

They view their business as serving their customers.

Lex Fridman (39:13.940)

But as you climb up the hierarchy in the cable companies,

James Gosling (39:20.780)

that view shifts because really the business of the cable

Lex Fridman (39:29.780)

companies had always been selling eyeballs

James Gosling (39:34.460)

to advertisers.

Lex Fridman (39:36.580)

Right.

Lex Fridman (39:39.820)

And that view of a cable company

Lex Fridman (39:45.580)

didn't really dawn on most people who

James Gosling (39:49.060)

worked at the cable companies.

Lex Fridman (39:50.900)

But I had various dust ups with various cable companies

James Gosling (39:57.340)

where you could see in the stratified layers

Lex Fridman (40:00.220)

of the corporation that this view of the reason

James Gosling (40:06.260)

that you have cable TV is to capture eyeballs.

Lex Fridman (40:13.140)

So they didn't see it that way.

James Gosling (40:15.060)

Well, so most of the people who worked at the phone company

Lex Fridman (40:19.820)

or at the cable companies, their view

James Gosling (40:22.460)

was that their job was getting delightful content out

Lex Fridman (40:28.340)

to their customers.

Lex Fridman (40:29.540)

And their customers would pay for that.

Lex Fridman (40:33.140)

Higher up, they viewed this as a way of attracting eyeballs

James Gosling (40:40.100)

to them.

Lex Fridman (40:41.940)

And then what they were really doing

James Gosling (40:45.620)

was selling the eyeballs that were glued to their content

Lex Fridman (40:51.580)

to the advertisers.

James Gosling (40:52.380)

To the advertisers, yeah.

Lex Fridman (40:54.420)

And so the internet was a competition in that sense.

James Gosling (40:57.260)

Right.

Lex Fridman (40:58.980)

They were right.

James Gosling (41:00.300)

Well, yeah.

Lex Fridman (41:02.620)

I mean, there was one proposal that we sent,

James Gosling (41:08.620)

one detailed proposal that we wrote up back

Lex Fridman (41:14.460)

at Sun in the early 90s that was essentially like, look,

James Gosling (41:19.700)

anybody with internet technologies,

Lex Fridman (41:22.260)

anybody can become provider of content.

Lex Fridman (41:27.100)

So you could be distributing home movies

Lex Fridman (41:32.740)

to your parents or your cousins or who are anywhere else.

Lex Fridman (41:39.420)

So anybody can become a publisher.

Lex Fridman (41:41.380)

Wow, you were thinking about that already.

James Gosling (41:43.300)

Netflix, Netflix, YouTube.

Lex Fridman (41:45.660)

Yeah, that was like in the early 90s.

Lex Fridman (41:49.780)

And we thought, this would be great.

Lex Fridman (41:54.060)

And the kind of content we were thinking about at the time

James Gosling (41:57.380)

was like home movies, kids essays,

Lex Fridman (42:05.100)

stuff from grocery stores or a restaurant

James Gosling (42:12.380)

that they could actually start sending information about.

Lex Fridman (42:20.340)

That's brilliant.

Lex Fridman (42:21.220)

And the reaction of the cable companies

Lex Fridman (42:25.060)

was like, fuck no.

James Gosling (42:29.940)

Because then we're out of business.

Lex Fridman (42:34.940)

What is it about companies that, because they could have just,

James Gosling (42:38.420)

they could have been ahead of that wave.

Lex Fridman (42:40.060)

They could have listened to that.

Lex Fridman (42:41.780)

And they could have.

Lex Fridman (42:42.660)

They didn't see a path to revenue.

James Gosling (42:46.940)

Somewhere in there, there's a lesson for big companies,

Lex Fridman (42:50.420)

like to listen, to try to anticipate the renegade,

James Gosling (42:56.060)

the out there, out of the box, people like yourself

Lex Fridman (42:59.420)

in the early days writing proposals

James Gosling (43:01.460)

about what this could possibly be.

Lex Fridman (43:03.780)

Well, and that wasn't.

James Gosling (43:06.940)

If you're in a position where you're

Lex Fridman (43:09.020)

making truckloads of money off of a particular business model,

James Gosling (43:19.780)

the whole thought of leaping the chasm,

Lex Fridman (43:28.060)

you can see, oh, new models that are more effective

James Gosling (43:33.060)

are emerging, so like digital cameras versus film cameras.

Lex Fridman (43:44.060)

Why take the leap?

Lex Fridman (43:45.820)

Why take the leap?

Lex Fridman (43:46.580)

Because you're making so much money off of film.

Lex Fridman (43:50.980)

And in my past at Sun, one of our big customers was Kodak.

Lex Fridman (43:59.500)

And I ended up interacting with folks from Kodak quite a lot.

Lex Fridman (44:03.220)

And they actually had a big digital camera research

Lex Fridman (44:09.940)

and digital imaging business, or development group.

Lex Fridman (44:15.860)

And they knew that you just look at the trend lines

Lex Fridman (44:24.340)

and you look at the emerging quality of these digital

James Gosling (44:32.340)

cameras.

Lex Fridman (44:33.900)

And you can just plot it on the graph.

Lex Fridman (44:37.940)

And it's like, sure, film is better today.

Lex Fridman (44:44.100)

But digital is improving like this.

James Gosling (44:51.340)

The lines are going to cross.

Lex Fridman (44:52.860)

And the point at which the lines cross

James Gosling (44:56.820)

is going to be a collapse in their business.

Lex Fridman (45:00.260)

And they could see that.

James Gosling (45:05.100)

They absolutely knew that.

Lex Fridman (45:07.740)

The problem is that up to the point where they hit the wall,

James Gosling (45:12.860)

they were making truckloads of money.

Lex Fridman (45:16.300)

And when they did the math, it never

James Gosling (45:24.300)

started to make sense for them to kind of lead the charge.

Lex Fridman (45:29.300)

And part of the issues for a lot of companies

James Gosling (45:32.460)

for this kind of stuff is that if you're

Lex Fridman (45:37.020)

going to leap over a chasm like that,

James Gosling (45:39.260)

like with Kodak going from film to digital,

Lex Fridman (45:45.820)

that's a transition that's going to take a while.

James Gosling (45:49.900)

We had fights like this with people over smart cards.

Lex Fridman (45:53.500)

The smart cards fights were just ludicrous.

Lex Fridman (45:57.820)

But that's where visionary leadership comes in, right?

Lex Fridman (46:00.180)

Somebody needs to roll in and say, then take the leap.

James Gosling (46:04.700)

Well, it's partly take the leap,

Lex Fridman (46:07.300)

but it's also partly take the hit.

James Gosling (46:09.500)

Take the hit in the short term.

Lex Fridman (46:12.540)

So you can draw the graphs you want that show that if we leap

James Gosling (46:17.380)

from here, on our present trajectory, we're doing this

Lex Fridman (46:21.540)

and there's a cliff.

James Gosling (46:22.940)

If we force ourselves into a transition

Lex Fridman (46:27.580)

and we proactively do that, we can be on the next wave.

Lex Fridman (46:33.660)

But there will be a period when we're in a trough.

Lex Fridman (46:39.020)

And pretty much always there ends up being a trough

James Gosling (46:43.940)

as you leap the chasm.

Lex Fridman (46:46.100)

But the way that public companies work on this planet,

James Gosling (46:52.620)

they're reporting every quarter.

Lex Fridman (46:55.700)

And the one thing that a CEO must never do is take a big hit.

James Gosling (47:02.700)

Take a big hit.

Lex Fridman (47:04.700)

Over some quarter.

Lex Fridman (47:06.180)

And many of these transitions involve a big hit

Lex Fridman (47:11.180)

for a period of time, one, two, three quarters.

Lex Fridman (47:17.340)

And so you get some companies and like Tesla and Amazon

Lex Fridman (47:26.060)

are really good examples of companies that take huge hits.

Lex Fridman (47:32.380)

But they have the luxury of being

Lex Fridman (47:34.140)

able to ignore the stock market for a little while.

Lex Fridman (47:38.060)

And that's not so true today, really.

Lex Fridman (47:42.100)

But in the early days of both of those companies,

James Gosling (47:50.580)

they both did this thing of, I don't care

Lex Fridman (47:56.220)

about the quarterly reports.

James Gosling (47:57.620)

I care about how many happy customers we have.

Lex Fridman (48:02.260)

And having as many happy customers as possible

James Gosling (48:05.900)

can often be an enemy of the bottom line.

Lex Fridman (48:11.340)

Yeah, so how do they make that work?

James Gosling (48:12.780)

I mean, Amazon operated in the negative for a long time.

Lex Fridman (48:15.460)

It's like investing into the future.

James Gosling (48:17.260)

Right.

Lex Fridman (48:18.060)

But so Amazon and Google and Tesla and Facebook, a lot

James Gosling (48:23.900)

of those had what amounted to patient money,

Lex Fridman (48:30.340)

often because there's like a charismatic central figure who

James Gosling (48:36.540)

has a really large block of stock.

Lex Fridman (48:41.020)

And they can just make it so.

Lex Fridman (48:45.620)

So on that topic, just maybe it's a small tangent,

Lex Fridman (48:49.380)

but you've gotten the chance to work

James Gosling (48:51.420)

with some pretty big leaders.

Lex Fridman (48:53.220)

What are your thoughts about on the Tesla side, Elon Musk

James Gosling (48:56.460)

leadership, on the Amazon side, Jeff Bezos,

Lex Fridman (49:00.060)

all of these folks with large amounts of stock and vision

Lex Fridman (49:03.900)

in their company?

Lex Fridman (49:04.940)

I mean, they're founders, either complete founders

James Gosling (49:09.260)

or early on folks.

Lex Fridman (49:11.300)

And Amazon have taken a lot of leaps.

Lex Fridman (49:17.300)

And that probably at the time, people

Lex Fridman (49:20.660)

would criticize as like, what is this bookstore thing?

James Gosling (49:26.340)

Yeah.

Lex Fridman (49:26.860)

And Bezos had a vision.

Lex Fridman (49:31.220)

And he had the ability to just follow it.

Lex Fridman (49:36.460)

Lots of people have visions.

Lex Fridman (49:37.820)

And the average vision is completely idiotic,

Lex Fridman (49:41.500)

and you crash and burn.

James Gosling (49:45.020)

The Silicon Valley crash and burn rate is pretty high.

Lex Fridman (49:52.500)

And they don't necessarily crash and burn

James Gosling (49:54.540)

because they were dumb ideas.

Lex Fridman (49:55.900)

But often, it's just timing and luck.

Lex Fridman (50:00.780)

And you take companies like Tesla,

Lex Fridman (50:07.860)

and really, the original Tesla sort of pre Elon

James Gosling (50:19.220)

was kind of doing sort of OK.

Lex Fridman (50:22.140)

But he just drove them.

Lex Fridman (50:26.300)

And because he had a really strong vision,

Lex Fridman (50:32.180)

he would make calls that were always mostly pretty good.

James Gosling (50:40.220)

I mean, the Model X was kind of a goofball thing to do.

Lex Fridman (50:44.580)

But he did it boldly anyway.

James Gosling (50:46.620)

There's so many people that just said,

Lex Fridman (50:50.100)

there's so many people that oppose them on the door.

James Gosling (50:54.300)

From the engineering perspective,

Lex Fridman (50:55.740)

those doors are ridiculous.

James Gosling (50:58.820)

They are a complete travesty.

Lex Fridman (51:02.060)

But they're exactly the symbol of what great leadership is,

James Gosling (51:05.700)

which is like, you have a vision, and you just go.

Lex Fridman (51:08.980)

If you're going to do something stupid, make it really stupid.

Lex Fridman (51:12.220)

And go all in.

Lex Fridman (51:14.380)

Yeah.

Lex Fridman (51:16.100)

And to Musk's credit, he's a really sharp guy.

Lex Fridman (51:22.500)

So going back in time a little bit to Steve Jobs,

James Gosling (51:26.860)

Steve Jobs was a similar sort of character

Lex Fridman (51:29.580)

who had a strong vision and was really, really smart.

Lex Fridman (51:34.260)

And he wasn't smart about the technology parts of things.

Lex Fridman (51:39.140)

But he was really sharp about the sort of human relationship

James Gosling (51:47.900)

between the relationship between humans and objects.

Lex Fridman (51:54.580)

But he was a jerk.

Lex Fridman (52:01.060)

Can we just linger on that a little bit?

Lex Fridman (52:02.660)

People say he's a jerk.

Lex Fridman (52:06.100)

Is that a feature or a bug?

Lex Fridman (52:08.060)

Well, that's the question, right?

Lex Fridman (52:11.220)

So you take people like Steve, who was really hard on people.

Lex Fridman (52:19.500)

And so the question is, was he needlessly hard on people?

Lex Fridman (52:25.380)

Or was he just making people reach to meet his vision?

Lex Fridman (52:34.420)

And you could kind of spin it either way.

James Gosling (52:41.540)

Well, the results tell a story.

Lex Fridman (52:45.020)

He, through whatever jerk ways he had,

James Gosling (52:47.820)

he made people often do the best work of their life.

Lex Fridman (52:51.420)

Yeah.

James Gosling (52:52.220)

Yeah, and that was absolutely true.

Lex Fridman (52:54.420)

And I interviewed with him several times.

James Gosling (52:59.980)

I did various negotiations with him.

Lex Fridman (53:04.380)

And even though kind of personally I liked him,

James Gosling (53:15.820)

I could never work for him.

Lex Fridman (53:18.540)

Why do you think that?

Lex Fridman (53:22.580)

Can you put into words the kind of tension

Lex Fridman (53:25.500)

that you feel would be destructive as opposed

Lex Fridman (53:29.580)

to constructive?

Lex Fridman (53:32.980)

Oh, he'd yell at people.

James Gosling (53:35.780)

He'd call them names.

Lex Fridman (53:39.020)

And you don't like that?

James Gosling (53:40.140)

No.

Lex Fridman (53:41.780)

No, I don't think you need to do that.

James Gosling (53:43.980)

Yeah.

Lex Fridman (53:46.660)

And I think there's pushing people to excel.

Lex Fridman (53:58.340)

And then there's too far.

Lex Fridman (54:01.580)

And I think he was on the wrong side of the line.

Lex Fridman (54:04.620)

And I've never worked for Musk.

Lex Fridman (54:07.460)

I know a number of people who have, many of them have said,

Lex Fridman (54:11.820)

and it shows up in the press a lot,

Lex Fridman (54:13.820)

that Musk is kind of that way.

Lex Fridman (54:18.260)

And one of the things that I sort of loathe

Lex Fridman (54:21.860)

about Silicon Valley these days is

James Gosling (54:24.700)

that a lot of the high flying successes

Lex Fridman (54:29.460)

are run by people who are complete jerks.

Lex Fridman (54:33.380)

But it seems like there's come this sort of mythology out

Lex Fridman (54:39.700)

of Steve Jobs that the reason that he succeeded

James Gosling (54:44.420)

was because he was super hard on people.

Lex Fridman (54:50.180)

And in a number of corners, people start going,

James Gosling (54:57.260)

oh, if I want to succeed, I need to be a real jerk.

Lex Fridman (55:02.060)

And that, for me, just does not compute.

James Gosling (55:05.860)

I know a lot of successful people

Lex Fridman (55:07.660)

who are not jerks, who are perfectly fine people.

Lex Fridman (55:11.780)

And they tend to not be in the public eye.

Lex Fridman (55:20.660)

The general public somehow lifts the jerks up

James Gosling (55:24.900)

into the hero status.

Lex Fridman (55:28.580)

Right.

James Gosling (55:29.340)

Well, because they do things that get them in the press.

Lex Fridman (55:32.900)

And the people who don't do the kind of things

James Gosling (55:44.180)

that spill into the press.

Lex Fridman (55:47.340)

Yeah, I just talked to Chris Ladner for the second time.

James Gosling (55:54.660)

He's a super nice guy.

Lex Fridman (55:56.700)

Just an example of this kind of individual

James Gosling (55:58.860)

that's in the background.

Lex Fridman (56:00.300)

I feel like he's behind a million technologies.

Lex Fridman (56:03.060)

But he also talked about the jerkiness of some of the folks.

Lex Fridman (56:06.860)

Yeah.

James Gosling (56:07.660)

Yeah, and the fact that being a jerk

Lex Fridman (56:10.660)

has become a required style.

Lex Fridman (56:13.140)

But one thing I maybe want to ask on that

Lex Fridman (56:15.820)

is maybe to push back a little bit.

Lex Fridman (56:17.780)

So there's the jerk side.

Lex Fridman (56:19.460)

But there's also, if I were to criticize

Lex Fridman (56:22.420)

what I've seen in Silicon Valley, which is almost

Lex Fridman (56:25.220)

the resistance to working hard.

Lex Fridman (56:28.380)

So on the jerkiness side, it's so Posty Jobs and Elon kind

Lex Fridman (56:37.660)

of push people to work really hard to do.

Lex Fridman (56:41.820)

And there's a question whether it's

Lex Fridman (56:43.820)

possible to do that nicely.

Lex Fridman (56:45.860)

But one of the things that bothers me,

Lex Fridman (56:47.540)

maybe I'm just Russian and just kind of romanticize

James Gosling (56:51.820)

the whole suffering thing.

Lex Fridman (56:53.300)

But I think working hard is essential for accomplishing

James Gosling (56:57.380)

anything interesting, like really hard.

Lex Fridman (57:00.140)

And in the parlance of Silicon Valley,

James Gosling (57:02.940)

it's probably too hard.

Lex Fridman (57:04.820)

This idea of that you should work smart, not hard often

James Gosling (57:10.340)

to me sounds like you should be lazy.

Lex Fridman (57:12.700)

Because of course you want to be to work smart.

James Gosling (57:15.220)

Of course you would be maximally efficient.

Lex Fridman (57:18.220)

But in order to discover the efficient path

James Gosling (57:20.140)

like we're talking about with the short programs, you have to.

Lex Fridman (57:23.340)

Well, the smart, hard thing isn't an either or.

James Gosling (57:28.620)

It's an and.

Lex Fridman (57:29.540)

It's an and, yeah.

James Gosling (57:30.620)

Right.

Lex Fridman (57:31.500)

And the people who say you should work smart, not hard,

James Gosling (57:42.460)

they pretty much always fail.

Lex Fridman (57:44.300)

Yeah.

James Gosling (57:45.020)

Thank you.

Lex Fridman (57:46.180)

Right.

James Gosling (57:46.660)

I mean, that's just a recipe for disaster.

Lex Fridman (57:49.700)

I mean, there are counterexamples,

Lex Fridman (57:53.460)

but they're more people who benefited from luck.

Lex Fridman (57:58.420)

And you're saying, yeah, exactly.

James Gosling (57:59.940)

Luck and timing, like you said, is often an essential thing.

Lex Fridman (58:04.620)

But you're saying you can push people

James Gosling (58:07.460)

to work hard and do incredible work without being nasty.

Lex Fridman (58:13.580)

Yeah, without being nasty.

James Gosling (58:14.900)

I think Google is a good example of the leadership of Google

Lex Fridman (58:22.100)

throughout its history has been a pretty good example of not

James Gosling (58:26.420)

being nasty and being kind.

Lex Fridman (58:28.140)

Yeah.

James Gosling (58:28.620)

I mean, the twins, Larry and Sergey,

Lex Fridman (58:35.820)

are both pretty nice people.

James Gosling (58:37.620)

Sandra Pichai is very nice.

Lex Fridman (58:39.660)

Yeah.

James Gosling (58:40.740)

Yeah.

Lex Fridman (58:41.260)

And it's a culture of people who work really, really hard.

James Gosling (58:49.860)

Let me ask maybe a little bit of a tense question.

Lex Fridman (58:54.220)

We're talking about Emacs.

James Gosling (58:56.500)

It seems like you've done some incredible work,

Lex Fridman (58:58.900)

so outside of Java, you've done some incredible work that

James Gosling (59:02.020)

didn't become as popular as it could have because of licensing

Lex Fridman (59:06.460)

issues and open source issues.

Lex Fridman (59:11.700)

Is it, what are your thoughts about that entire mess?

Lex Fridman (59:18.740)

Like what's about open source now in retrospect looking back?

James Gosling (59:24.340)

About licensing, about open sourcing,

Lex Fridman (59:27.060)

do you think open source is a good thing, a bad thing?

Lex Fridman (59:33.820)

Do you have regrets?

Lex Fridman (59:35.660)

Do you have wisdom that you've learned

Lex Fridman (59:38.020)

from that whole experience?

Lex Fridman (59:39.900)

So in general, I'm a big fan of open source.

James Gosling (59:45.380)

The way that it can be used to build communities and promote

Lex Fridman (59:49.780)

the development of things and promote collaboration and all

James Gosling (59:54.260)

of that is really pretty grand.

Lex Fridman (59:59.100)

When open source turns into a religion that says all things