Steven Pinker: AI in the Age of Reason
AI 与机器学习心理与人性音乐与艺术生物与进化历史与文明
📋 章节目录
暂无章节信息
🔑 关键词
donhumanintelligenceartificialgoalgoingengineeringbooksaidknowledgeneuralthreatworrysystemscancernuclearbetternetworksweaponsfear
💬 精彩语录
"harm. Yep. So let me ask this question. So speaking of imaginability, you said it's important to think"
— Steven Pinker （30:08.080）
"feel. Whether an artificial system would experience that or not, I don't know and I don't think I can"
— Steven Pinker （04:49.840）
"purpose AI is not going to do anything conceivable in order to attain a goal. All engineering systems"
— Steven Pinker （17:50.240）
"C. to attaining not only knowledge but fulfillment more generally, that is life, health, stimulation,"
— Steven Pinker （00:51.200）
"much better than polyester. It's not that cotton has something magic in it. It's not that we couldn't"
— Steven Pinker （08:21.600）
🎙️ 完整对话（373 条）
Lex Fridman (00:00.000)
You've studied the human mind, cognition, language, vision, evolution, psychology,
Lex Fridman (00:05.280)
from child to adult, from the level of individual to the level of our entire civilization.
Lex Fridman (00:11.680)
So I feel like I can start with a simple multiple choice question.
Lex Fridman (00:16.240)
What is the meaning of life? Is it A. to attain knowledge as Plato said,
Lex Fridman (00:22.400)
B. to attain power as Nietzsche said, C. to escape death as Ernest Becker said,
Steven Pinker (00:27.920)
D. to propagate our genes as Darwin and others have said,
Lex Fridman (00:33.200)
E. there is no meaning as the nihilists have said,
Steven Pinker (00:37.520)
F. knowing the meaning of life is beyond our cognitive capabilities as Stephen Pinker said,
Lex Fridman (00:42.960)
based on my interpretation 20 years ago, and G. none of the above.
Steven Pinker (00:48.160)
I'd say A. comes closest, but I would amend that to
Steven Pinker (00:51.200)
C. to attaining not only knowledge but fulfillment more generally, that is life, health, stimulation,
Steven Pinker (01:00.640)
access to the living cultural and social world.
Steven Pinker (01:06.080)
Now this is our meaning of life. It's not the meaning of life if you were to ask our genes.
Steven Pinker (01:12.160)
Their meaning is to propagate copies of themselves, but that is distinct from the
Lex Fridman (01:17.680)
meaning that the brain that they lead to sets for itself.
Lex Fridman (01:22.400)
So to you knowledge is a small subset or a large subset?
Steven Pinker (01:27.840)
It's a large subset, but it's not the entirety of human striving because we also want to
Steven Pinker (01:35.200)
interact with people. We want to experience beauty. We want to experience the richness
Steven Pinker (01:39.600)
of the natural world, but understanding what makes the universe tick is way up there.
Steven Pinker (01:47.840)
For some of us more than others, certainly for me that's one of the top five.
Lex Fridman (01:54.560)
So is that a fundamental aspect? Are you just describing your own preference or is this a
Steven Pinker (02:00.080)
fundamental aspect of human nature is to seek knowledge? In your latest book you talk about
Steven Pinker (02:05.920)
the power, the usefulness of rationality and reason and so on. Is that a fundamental
Lex Fridman (02:11.760)
nature of human beings or is it something we should just strive for?
Steven Pinker (02:17.040)
Both. We're capable of striving for it because it is one of the things that make us what we are,
Steven Pinker (02:23.920)
homo sapiens, wise men. We are unusual among animals in the degree to which we acquire
Steven Pinker (02:32.320)
knowledge and use it to survive. We make tools. We strike agreements via language. We extract
Steven Pinker (02:41.600)
poisons. We predict the behavior of animals. We try to get at the workings of plants.
Lex Fridman (02:47.760)
And when I say we, I don't just mean we in the modern West, but we as a species everywhere,
Steven Pinker (02:52.640)
which is how we've managed to occupy every niche on the planet, how we've managed to drive other
Steven Pinker (02:58.160)
animals to extinction. And the refinement of reason in pursuit of human wellbeing, of health,
Steven Pinker (03:06.480)
happiness, social richness, cultural richness is our main challenge in the present. That is
Steven Pinker (03:14.480)
using our intellect, using our knowledge to figure out how the world works, how we work
Steven Pinker (03:19.280)
in order to make discoveries and strike agreements that make us all better off in the long run.
Steven Pinker (03:25.200)
Right. And you do that almost undeniably and in a data driven way in your recent book,
Lex Fridman (03:31.840)
but I'd like to focus on the artificial intelligence aspect of things and not just
Steven Pinker (03:36.400)
artificial intelligence, but natural intelligence too. So 20 years ago in a book you've written on
Lex Fridman (03:41.920)
how the mind works, you conjecture again, am I right to interpret things? You can correct me
Steven Pinker (03:49.520)
if I'm wrong, but you conjecture that human thought in the brain may be a result of a
Steven Pinker (03:54.400)
massive network of highly interconnected neurons. So from this interconnectivity emerges thought
Steven Pinker (04:01.280)
compared to artificial neural networks, which we use for machine learning today,
Steven Pinker (04:06.160)
is there something fundamentally more complex, mysterious, even magical about the biological
Steven Pinker (04:12.640)
neural networks versus the ones we've been starting to use over the past 60 years and
Steven Pinker (04:19.440)
have become to success in the past 10? There is something a little bit mysterious
Steven Pinker (04:24.720)
about the human neural networks, which is that each one of us who is a neural network knows that
Steven Pinker (04:31.600)
we ourselves are conscious. Conscious not in the sense of registering our surroundings or even
Steven Pinker (04:36.960)
registering our internal state, but in having subjective first person, present tense experience.
Steven Pinker (04:42.720)
That is when I see red, it's not just different from green, but there's a redness to it that I
Steven Pinker (04:49.840)
feel. Whether an artificial system would experience that or not, I don't know and I don't think I can
Steven Pinker (04:54.960)
know. That's why it's mysterious. If we had a perfectly lifelike robot that was behaviorally
Steven Pinker (05:00.480)
indistinguishable from a human, would we attribute consciousness to it or ought we to attribute
Steven Pinker (05:06.800)
consciousness to it? And that's something that it's very hard to know. But putting that aside,
Steven Pinker (05:12.160)
putting aside that largely philosophical question, the question is, is there some difference between
Steven Pinker (05:19.040)
the human neural network and the ones that we're building in artificial intelligence will mean
Steven Pinker (05:23.760)
that we're on the current trajectory, not going to reach the point where we've got a lifelike
Steven Pinker (05:30.000)
robot indistinguishable from a human because the way their so called neural networks are organized
Steven Pinker (05:35.120)
are different from the way ours are organized. I think there's overlap, but I think there are
Steven Pinker (05:39.760)
some big differences that current neural networks, current so called deep learning systems are in
Steven Pinker (05:48.800)
reality not all that deep. That is, they are very good at extracting high order statistical
Steven Pinker (05:53.840)
regularities, but most of the systems don't have a semantic level, a level of actual understanding
Steven Pinker (06:00.640)
of who did what to whom, why, where, how things work, what causes what else. Do you think that
Steven Pinker (06:07.520)
kind of thing can emerge as it does? So artificial neural networks are much smaller, the number of
Steven Pinker (06:11.840)
connections and so on than the current human biological networks, but do you think sort of
Steven Pinker (06:18.320)
to go to consciousness or to go to this higher level semantic reasoning about things, do you
Lex Fridman (06:22.800)
think that can emerge with just a larger network with a more richly weirdly interconnected network?
Steven Pinker (06:29.760)
Separate it in consciousness because consciousness is even a matter of complexity.
Lex Fridman (06:33.280)
A really weird one.
Steven Pinker (06:34.320)
Yeah, you could sensibly ask the question of whether shrimp are conscious, for example,
Steven Pinker (06:38.400)
they're not terribly complex, but maybe they feel pain. So let's just put that part of it aside.
Lex Fridman (06:44.400)
But I think sheer size of a neural network is not enough to give it structure and knowledge,
Lex Fridman (06:52.480)
but if it's suitably engineered, then why not? That is, we're neural networks, natural selection
Steven Pinker (06:59.520)
did a kind of equivalent of engineering of our brains. So I don't think there's anything mysterious
Steven Pinker (07:04.880)
in the sense that no system made out of silicon could ever do what a human brain can do. I think
Steven Pinker (07:11.760)
it's possible in principle. Whether it'll ever happen depends not only on how clever we are
Steven Pinker (07:17.360)
in engineering these systems, but whether we even want to, whether that's even a sensible goal.
Lex Fridman (07:22.240)
That is, you can ask the question, is there any locomotion system that is as good as a human?
Steven Pinker (07:29.920)
Well, we kind of want to do better than a human ultimately in terms of legged locomotion.
Steven Pinker (07:35.520)
There's no reason that humans should be our benchmark. They're tools that might be better
Steven Pinker (07:39.760)
in some ways. It may be that we can't duplicate a natural system because at some point it's so much
Steven Pinker (07:49.760)
cheaper to use a natural system that we're not going to invest more brainpower and resources.
Lex Fridman (07:54.560)
So for example, we don't really have an exact substitute for wood. We still build houses out
Steven Pinker (08:00.480)
of wood. We still build furniture out of wood. We like the look. We like the feel. It has certain
Steven Pinker (08:05.040)
properties that synthetics don't. It's not that there's anything magical or mysterious about wood.
Steven Pinker (08:11.120)
It's just that the extra steps of duplicating everything about wood is something we just haven't
Steven Pinker (08:17.600)
bothered because we have wood. Likewise, say cotton. I'm wearing cotton clothing now. It feels
Steven Pinker (08:21.600)
much better than polyester. It's not that cotton has something magic in it. It's not that we couldn't
Steven Pinker (08:29.760)
ever synthesize something exactly like cotton, but at some point it's just not worth it. We've got
Steven Pinker (08:35.920)
cotton. Likewise, in the case of human intelligence, the goal of making an artificial system that is
Steven Pinker (08:42.000)
exactly like the human brain is a goal that we probably know is going to pursue to the bitter
Steven Pinker (08:47.840)
end, I suspect, because if you want tools that do things better than humans, you're not going to
Steven Pinker (08:53.120)
care whether it does something like humans. So for example, diagnosing cancer or predicting the
Steven Pinker (08:58.240)
weather, why set humans as your benchmark? But in general, I suspect you also believe
Steven Pinker (09:05.840)
that even if the human should not be a benchmark and we don't want to imitate humans in their
Steven Pinker (09:10.480)
system, there's a lot to be learned about how to create an artificial intelligence system by
Steven Pinker (09:15.520)
studying the human. Yeah, I think that's right. In the same way that to build flying machines,
Steven Pinker (09:22.480)
we want to understand the laws of aerodynamics, including birds, but not mimic the birds,
Lex Fridman (09:27.760)
but they're the same laws. You have a view on AI, artificial intelligence, and safety
Steven Pinker (09:35.440)
that, from my perspective, is refreshingly rational or perhaps more importantly, has elements
Steven Pinker (09:47.040)
of positivity to it, which I think can be inspiring and empowering as opposed to paralyzing.
Steven Pinker (09:53.600)
For many people, including AI researchers, the eventual existential threat of AI is obvious,
Steven Pinker (09:59.840)
not only possible, but obvious. And for many others, including AI researchers, the threat
Steven Pinker (10:05.840)
is not obvious. So Elon Musk is famously in the highly concerned about AI camp, saying things like
Steven Pinker (10:14.640)
AI is far more dangerous than nuclear weapons, and that AI will likely destroy human civilization.
Steven Pinker (10:21.200)
Human civilization. So in February, he said that if Elon was really serious about AI, the threat
Steven Pinker (10:29.360)
of AI, he would stop building self driving cars that he's doing very successfully as part of Tesla.
Steven Pinker (10:35.760)
Then he said, wow, if even Pinker doesn't understand the difference between narrow AI,
Steven Pinker (10:40.800)
like a car and general AI, when the latter literally has a million times more compute power
Lex Fridman (10:47.280)
and an open ended utility function, humanity is in deep trouble. So first, what did you mean by
Lex Fridman (10:54.160)
the statement about Elon Musk should stop building self driving cars if he's deeply concerned?
Steven Pinker (11:00.080)
Not the last time that Elon Musk has fired off an intemperate tweet.
Lex Fridman (11:04.320)
Well, we live in a world where Twitter has power.
Steven Pinker (11:07.680)
Yes. Yeah, I think there are two kinds of existential threat that have been discussed
Steven Pinker (11:16.640)
in connection with artificial intelligence, and I think that they're both incoherent.
Steven Pinker (11:20.480)
One of them is a vague fear of AI takeover, that just as we subjugated animals and less technologically
Steven Pinker (11:29.520)
advanced peoples, so if we build something that's more advanced than us, it will inevitably turn us
Steven Pinker (11:34.640)
into pets or slaves or domesticated animal equivalents. I think this confuses intelligence
Steven Pinker (11:42.320)
with a will to power, that it so happens that in the intelligence system we are most familiar with,
Steven Pinker (11:49.200)
namely homo sapiens, we are products of natural selection, which is a competitive process,
Lex Fridman (11:54.160)
and so bundled together with our problem solving capacity are a number of nasty traits like
Steven Pinker (12:00.320)
dominance and exploitation and maximization of power and glory and resources and influence.
Steven Pinker (12:08.720)
There's no reason to think that sheer problem solving capability will set that as one of its
Steven Pinker (12:13.120)
goals. Its goals will be whatever we set its goals as, and as long as someone isn't building a
Steven Pinker (12:18.720)
megalomaniacal artificial intelligence, then there's no reason to think that it would naturally
Steven Pinker (12:24.320)
evolve in that direction. Now, you might say, well, what if we gave it the goal of maximizing
Steven Pinker (12:28.960)
its own power source? That's a pretty stupid goal to give an autonomous system. You don't give it
Steven Pinker (12:34.880)
that goal. I mean, that's just self evidently idiotic. So if you look at the history of the
Steven Pinker (12:40.720)
world, there's been a lot of opportunities where engineers could instill in a system
Steven Pinker (12:45.120)
destructive power and they choose not to because that's the natural process of engineering.
Steven Pinker (12:49.440)
Well, except for weapons. I mean, if you're building a weapon, its goal is to destroy people,
Lex Fridman (12:53.680)
and so I think there are good reasons to not build certain kinds of weapons. I think building
Steven Pinker (12:58.560)
nuclear weapons was a massive mistake. You do. So maybe pause on that because that is one of
Steven Pinker (13:06.480)
the serious threats. Do you think that it was a mistake in a sense that it should have been
Steven Pinker (13:12.800)
stopped early on? Or do you think it's just an unfortunate event of invention that this was
Steven Pinker (13:19.200)
invented? Do you think it's possible to stop? I guess is the question. It's hard to rewind the
Steven Pinker (13:23.280)
clock because of course it was invented in the context of World War II and the fear that the
Steven Pinker (13:28.320)
Nazis might develop one first. Then once it was initiated for that reason, it was hard to turn
Steven Pinker (13:35.120)
off, especially since winning the war against the Japanese and the Nazis was such an overwhelming
Steven Pinker (13:42.080)
goal of every responsible person that there's just nothing that people wouldn't have done then
Steven Pinker (13:47.120)
to ensure victory. It's quite possible if World War II hadn't happened that nuclear weapons
Steven Pinker (13:52.560)
wouldn't have been invented. We can't know, but I don't think it was by any means a necessity,
Steven Pinker (13:57.440)
any more than some of the other weapon systems that were envisioned but never implemented,
Steven Pinker (14:02.720)
like planes that would disperse poison gas over cities like crop dusters or systems to try to
Lex Fridman (14:10.560)
create earthquakes and tsunamis in enemy countries, to weaponize the weather,
Steven Pinker (14:16.000)
weaponize solar flares, all kinds of crazy schemes that we thought the better of.
Steven Pinker (14:21.120)
I think analogies between nuclear weapons and artificial intelligence are fundamentally
Steven Pinker (14:25.840)
misguided because the whole point of nuclear weapons is to destroy things. The point of
Lex Fridman (14:30.560)
artificial intelligence is not to destroy things. So the analogy is misleading.
Lex Fridman (14:36.080)
So there's two artificial intelligence you mentioned. The first one I guess is highly
Lex Fridman (14:39.920)
intelligent or power hungry.
Steven Pinker (14:42.080)
Yeah, it's a system that we design ourselves where we give it the goals. Goals are external to
Steven Pinker (14:46.800)
the means to attain the goals. If we don't design an artificially intelligent system to
Steven Pinker (14:55.200)
maximize dominance, then it won't maximize dominance. It's just that we're so familiar
Steven Pinker (15:00.800)
with homo sapiens where these two traits come bundled together, particularly in men,
Steven Pinker (15:06.320)
that we are apt to confuse high intelligence with a will to power, but that's just an error.
Steven Pinker (15:15.520)
The other fear is that will be collateral damage that will give artificial intelligence a goal
Steven Pinker (15:21.440)
like make paper clips and it will pursue that goal so brilliantly that before we can stop it,
Steven Pinker (15:27.440)
it turns us into paper clips. We'll give it the goal of curing cancer and it will turn us into
Steven Pinker (15:32.800)
guinea pigs for lethal experiments or give it the goal of world peace and its conception of world
Steven Pinker (15:38.720)
peace is no people, therefore no fighting and so it will kill us all. Now I think these are utterly
Steven Pinker (15:43.680)
fanciful. In fact, I think they're actually self defeating. They first of all assume that we're
Steven Pinker (15:49.040)
going to be so brilliant that we can design an artificial intelligence that can cure cancer,
Lex Fridman (15:53.600)
but so stupid that we don't specify what we mean by curing cancer in enough detail that it won't
Steven Pinker (15:59.600)
kill us in the process and it assumes that the system will be so smart that it can cure cancer,
Lex Fridman (16:06.240)
but so idiotic that it can't figure out that what we mean by curing cancer is not killing everyone.
Steven Pinker (16:12.880)
I think that the collateral damage scenario, the value alignment problem is also based on
Steven Pinker (16:18.320)
a misconception. So one of the challenges, of course, we don't know how to build either system
Steven Pinker (16:23.200)
currently or are we even close to knowing? Of course, those things can change overnight,
Lex Fridman (16:27.440)
but at this time, theorizing about it is very challenging in either direction. So that's
Steven Pinker (16:33.840)
probably at the core of the problem is without that ability to reason about the real engineering
Steven Pinker (16:39.600)
things here at hand is your imagination runs away with things. Exactly. But let me sort of ask,
Lex Fridman (16:45.120)
what do you think was the motivation, the thought process of Elon Musk? I build autonomous vehicles,
Steven Pinker (16:52.320)
I study autonomous vehicles, I study Tesla autopilot. I think it is one of the greatest
Steven Pinker (16:57.680)
currently large scale application of artificial intelligence in the world. It has potentially a
Steven Pinker (17:04.400)
very positive impact on society. So how does a person who's creating this very good quote unquote
Steven Pinker (17:10.880)
narrow AI system also seem to be so concerned about this other general AI? What do you think
Steven Pinker (17:19.280)
is the motivation there? What do you think is the thing? Well, you probably have to ask him,
Lex Fridman (17:23.040)
but there, and he is notoriously flamboyant, impulsive to the, as we have just seen,
Steven Pinker (17:31.520)
to the detriment of his own goals of the health of the company. So I don't know what's going on
Steven Pinker (17:37.360)
in his mind. You probably have to ask him, but I don't think the, and I don't think the distinction
Steven Pinker (17:42.560)
between special purpose AI and so called general AI is relevant that in the same way that special
Steven Pinker (17:50.240)
purpose AI is not going to do anything conceivable in order to attain a goal. All engineering systems
Steven Pinker (17:57.760)
are designed to trade off across multiple goals. When we build cars in the first place,
Steven Pinker (18:02.800)
we didn't forget to install brakes because the goal of a car is to go fast. It occurred to people,
Steven Pinker (18:08.880)
yes, you want it to go fast, but not always. So you would build in brakes too. Likewise,
Steven Pinker (18:13.920)
if a car is going to be autonomous and program it to take the shortest route to the airport,
Steven Pinker (18:20.320)
it's not going to take the diagonal and mow down people and trees and fences because that's the
Steven Pinker (18:24.560)
shortest route. That's not what we mean by the shortest route when we program it. And that's just
Lex Fridman (18:29.120)
what an intelligence system is by definition. It takes into account multiple constraints.
Steven Pinker (18:36.000)
The same is true, in fact, even more true of so called general intelligence. That is,
Steven Pinker (18:41.280)
if it's genuinely intelligent, it's not going to pursue some goal singlemindedly, omitting every
Steven Pinker (18:48.080)
other consideration and collateral effect. That's not artificial and general intelligence. That's
Steven Pinker (18:54.400)
artificial stupidity. I agree with you, by the way, on the promise of autonomous vehicles for
Steven Pinker (19:01.120)
improving human welfare. I think it's spectacular. And I'm surprised at how little press coverage
Steven Pinker (19:06.160)
notes that in the United States alone, something like 40,000 people die every year on the highways,
Steven Pinker (19:11.680)
vastly more than are killed by terrorists. And we spent a trillion dollars on a war to combat
Steven Pinker (19:18.240)
deaths by terrorism, about half a dozen a year. Whereas year in, year out, 40,000 people are
Steven Pinker (19:24.160)
massacred on the highways, which could be brought down to very close to zero. So I'm with you on
Steven Pinker (19:29.920)
the humanitarian benefit. Let me just mention that as a person who's building these cars,
Steven Pinker (19:34.800)
it is a little bit offensive to me to say that engineers would be clueless enough not to engineer
Steven Pinker (19:39.360)
safety into systems. I often stay up at night thinking about those 40,000 people that are dying.
Lex Fridman (19:45.280)
And everything I tried to engineer is to save those people's lives. So every new invention that
Steven Pinker (19:50.800)
I'm super excited about, in all the deep learning literature and CVPR conferences and NIPS, everything
Steven Pinker (19:59.280)
I'm super excited about is all grounded in making it safe and help people. So I just don't see how
Steven Pinker (20:08.240)
that trajectory can all of a sudden slip into a situation where intelligence will be highly
Steven Pinker (20:13.280)
negative. You and I certainly agree on that. And I think that's only the beginning of the
Steven Pinker (20:17.920)
potential humanitarian benefits of artificial intelligence. There's been enormous attention to
Lex Fridman (20:24.320)
what are we going to do with the people whose jobs are made obsolete by artificial intelligence,
Lex Fridman (20:28.720)
but very little attention given to the fact that the jobs that are going to be made obsolete are
Steven Pinker (20:32.560)
horrible jobs. The fact that people aren't going to be picking crops and making beds and driving
Steven Pinker (20:38.960)
trucks and mining coal, these are soul deadening jobs. And we have a whole literature sympathizing
Steven Pinker (20:45.520)
with the people stuck in these menial, mind deadening, dangerous jobs. If we can eliminate
Steven Pinker (20:53.040)
them, this is a fantastic boon to humanity. Now granted, you solve one problem and there's another
Steven Pinker (20:58.480)
one, namely, how do we get these people a decent income? But if we're smart enough to invent machines
Steven Pinker (21:05.840)
that can make beds and put away dishes and handle hospital patients, I think we're smart enough to
Steven Pinker (21:12.400)
figure out how to redistribute income to apportion some of the vast economic savings to the human
Steven Pinker (21:19.200)
beings who will no longer be needed to make beds. Okay. Sam Harris says that it's obvious that
Steven Pinker (21:26.000)
eventually AI will be an existential risk. He's one of the people who says it's obvious.
Steven Pinker (21:31.760)
We don't know when the claim goes, but eventually it's obvious. And because we don't know when,
Steven Pinker (21:38.640)
we should worry about it now. This is a very interesting argument in my eyes. So how do we
Steven Pinker (21:45.680)
think about timescale? How do we think about existential threats when we don't really, we know
Lex Fridman (21:51.120)
so little about the threat, unlike nuclear weapons perhaps, about this particular threat, that it
Steven Pinker (21:58.160)
could happen tomorrow, right? So, but very likely it won't. Very likely it'd be a hundred years away.
Lex Fridman (22:04.560)
So how do we ignore it? How do we talk about it? Do we worry about it? How do we think about those?
Lex Fridman (22:12.480)
What is it?
Lex Fridman (22:13.840)
A threat that we can imagine. It's within the limits of our imagination,
Lex Fridman (22:18.560)
but not within our limits of understanding to accurately predict it.
Lex Fridman (22:24.320)
But what is the it that we're afraid of?
Steven Pinker (22:26.880)
Sorry. AI being the existential threat.
Lex Fridman (22:30.320)
AI. How? Like enslaving us or turning us into paperclips?
Steven Pinker (22:35.120)
I think the most compelling from the Sam Harris perspective would be the paperclip situation.
Steven Pinker (22:39.520)
Yeah. I mean, I just think it's totally fanciful. I mean, that is don't build a system.
Steven Pinker (22:43.440)
Don't give a, don't, first of all, the code of engineering is you don't implement a system with
Steven Pinker (22:50.080)
massive control before testing it. Now, perhaps the culture of engineering will radically change.
Steven Pinker (22:55.040)
Then I would worry, but I don't see any signs that engineers will suddenly do idiotic things,
Steven Pinker (23:00.320)
like put a electric power plant in control of a system that they haven't tested first.
Steven Pinker (23:07.360)
Or all of these scenarios, not only imagine almost a magically powered intelligence,
Steven Pinker (23:15.600)
including things like cure cancer, which is probably an incoherent goal because there's
Lex Fridman (23:20.240)
so many different kinds of cancer or bring about world peace. I mean, how do you even specify that
Steven Pinker (23:25.600)
as a goal? But the scenarios also imagine some degree of control of every molecule in the
Steven Pinker (23:31.360)
universe, which not only is itself unlikely, but we would not start to connect these systems to
Lex Fridman (23:39.120)
infrastructure without testing as we would any kind of engineering system.
Steven Pinker (23:45.680)
Now, maybe some engineers will be irresponsible and we need legal and regulatory and legal
Steven Pinker (23:53.920)
responsibility implemented so that engineers don't do things that are stupid by their own standards.
Lex Fridman (24:00.640)
But the, I've never seen enough of a plausible scenario of existential threat to devote large
Steven Pinker (24:08.560)
amounts of brain power to, to forestall it. So you believe in the sort of the power on
Steven Pinker (24:14.240)
mass of the engineering of reason, as you argue in your latest book of Reason and Science, to sort of
Steven Pinker (24:20.400)
be the very thing that guides the development of new technology so it's safe and also keeps us safe.
Steven Pinker (24:28.000)
You know, granted the same culture of safety that currently is part of the engineering mindset for
Steven Pinker (24:34.560)
airplanes, for example. So yeah, I don't think that that should be thrown out the window and
Steven Pinker (24:40.480)
that untested all powerful systems should be suddenly implemented, but there's no reason to
Steven Pinker (24:45.520)
think they are. And in fact, if you look at the progress of artificial intelligence, it's been,
Steven Pinker (24:50.400)
you know, it's been impressive, especially in the last 10 years or so, but the idea that suddenly
Steven Pinker (24:54.160)
there'll be a step function that all of a sudden before we know it, it will be all powerful,
Steven Pinker (25:00.080)
that there'll be some kind of recursive self improvement, some kind of fume is also fanciful.
Steven Pinker (25:06.800)
We, certainly by the technology that we, that we're now impresses us, such as deep learning,
Steven Pinker (25:13.200)
where you train something on hundreds of thousands or millions of examples,
Steven Pinker (25:18.320)
they're not hundreds of thousands of problems of which curing cancer is a typical example.
Lex Fridman (25:26.000)
And so the kind of techniques that have allowed AI to increase in the last five years are not the
Steven Pinker (25:31.520)
kind that are going to lead to this fantasy of exponential sudden self improvement. I think it's
Steven Pinker (25:40.320)
kind of a magical thinking. It's not based on our understanding of how AI actually works.
Steven Pinker (25:45.440)
Now give me a chance here. So you said fanciful, magical thinking. In his TED talk,
Steven Pinker (25:51.040)
Sam Harris says that thinking about AI killing all human civilization is somehow fun,
Lex Fridman (25:55.760)
intellectually. Now I have to say as a scientist engineer, I don't find it fun,
Lex Fridman (26:00.400)
but when I'm having beer with my non AI friends, there is indeed something fun and appealing about
Steven Pinker (26:08.560)
it. Like talking about an episode of Black Mirror, considering if a large meteor is headed towards
Steven Pinker (26:14.640)
Earth, we were just told a large meteor is headed towards Earth, something like this. And can you
Lex Fridman (26:20.640)
relate to this sense of fun? And do you understand the psychology of it?
Steven Pinker (26:24.560)
Yes. Good question. I personally don't find it fun. I find it kind of actually a waste of time
Steven Pinker (26:32.800)
because there are genuine threats that we ought to be thinking about like pandemics, like cyber
Steven Pinker (26:39.760)
security vulnerabilities, like the possibility of nuclear war and certainly climate change.
Steven Pinker (26:46.160)
You know, this is enough to fill many conversations. And I think Sam did put his
Steven Pinker (26:54.320)
finger on something, namely that there is a community, sometimes called the rationality
Steven Pinker (27:00.240)
community, that delights in using its brainpower to come up with scenarios that would not occur
Steven Pinker (27:07.280)
to mere mortals, to less cerebral people. So there is a kind of intellectual thrill in finding new
Steven Pinker (27:14.560)
things to worry about that no one has worried about yet. I actually think, though, that it's
Steven Pinker (27:19.840)
not only is it a kind of fun that doesn't give me particular pleasure, but I think there can be a
Steven Pinker (27:25.440)
pernicious side to it, namely that you overcome people with such dread, such fatalism, that there
Steven Pinker (27:32.400)
are so many ways to die, to annihilate our civilization, that we may as well enjoy life
Steven Pinker (27:39.200)
while we can. There's nothing we can do about it. If climate change doesn't do us in, then runaway
Steven Pinker (27:42.880)
robots will. So let's enjoy ourselves now. We've got to prioritize. We have to look at threats that
Steven Pinker (27:52.480)
are close to certainty, such as climate change, and distinguish those from ones that are merely
Steven Pinker (27:58.160)
imaginable but with infinitesimal probabilities. And we have to take into account people's worry
Steven Pinker (28:05.280)
budget. You can't worry about everything. And if you sow dread and fear and terror and fatalism,
Steven Pinker (28:12.480)
it can lead to a kind of numbness. Well, these problems are overwhelming, and the engineers are
Steven Pinker (28:17.280)
just going to kill us all. So let's either destroy the entire infrastructure of science, technology,
Steven Pinker (28:26.640)
or let's just enjoy life while we can. So there's a certain line of worry, which I'm worried about
Steven Pinker (28:32.560)
a lot of things in engineering. There's a certain line of worry when you cross, you're allowed to
Steven Pinker (28:36.800)
cross, that it becomes paralyzing fear as opposed to productive fear. And that's kind of what
Steven Pinker (28:44.240)
you're highlighting. Exactly right. And we've seen some, we know that human effort is not
Steven Pinker (28:50.160)
well calibrated against risk in that because a basic tenet of cognitive psychology is that
Steven Pinker (28:58.160)
perception of risk and hence perception of fear is driven by imaginability, not by data. And so we
Steven Pinker (29:05.680)
misallocate vast amounts of resources to avoiding terrorism, which kills on average about six
Steven Pinker (29:11.360)
Americans a year with one exception of 9 11. We invade countries, we invent entire new departments
Steven Pinker (29:18.640)
of government with massive, massive expenditure of resources and lives to defend ourselves against
Steven Pinker (29:25.120)
a trivial risk. Whereas guaranteed risks, one of them you mentioned traffic fatalities and even
Steven Pinker (29:34.720)
risks that are not here, but are plausible enough to worry about like pandemics, like nuclear war,
Steven Pinker (29:45.920)
receive far too little attention. In presidential debates, there's no discussion of how to minimize
Steven Pinker (29:51.440)
the risk of nuclear war. Lots of discussion of terrorism, for example. And so I think it's
Steven Pinker (29:58.240)
essential to calibrate our budget of fear, worry, concern, planning to the actual probability of
Steven Pinker (30:08.080)
harm. Yep. So let me ask this question. So speaking of imaginability, you said it's important to think
Steven Pinker (30:15.840)
about reason and one of my favorite people who likes to dip into the outskirts of reason through
Steven Pinker (30:23.840)
fascinating exploration of his imagination is Joe Rogan. Oh yes. So who has through reason used to
Steven Pinker (30:32.000)
believe a lot of conspiracies and through reason has stripped away a lot of his beliefs in that
Steven Pinker (30:37.280)
way. So it's fascinating actually to watch him through rationality kind of throw away the ideas
Steven Pinker (30:43.120)
of Bigfoot and 9 11. I'm not sure exactly. Kim Trails. I don't know what he believes in. Yes.
Steven Pinker (30:50.320)
Okay. But he no longer believed in. No, that's right. No, he's become a real force for good.
Steven Pinker (30:55.520)
Yep. So you were on the Joe Rogan podcast in February and had a fascinating conversation,
Lex Fridman (31:00.240)
but as far as I remember, didn't talk much about artificial intelligence. I will be on his podcast
Steven Pinker (31:05.920)
in a couple of weeks. Joe is very much concerned about existential threat of AI. I'm not sure if
Steven Pinker (31:11.520)
you're, this is why I was hoping that you would get into that topic. And in this way,
Steven Pinker (31:17.040)
he represents quite a lot of people who look at the topic of AI from 10,000 foot level.
Lex Fridman (31:22.480)
So as an exercise of communication, you said it's important to be rational and reason
Steven Pinker (31:29.040)
about these things. Let me ask, if you were to coach me as an AI researcher about how to speak
Steven Pinker (31:34.080)
to Joe and the general public about AI, what would you advise? Well, the short answer would be to
Steven Pinker (31:40.640)
read the sections that I wrote in enlightenment now about AI, but a longer reason would be I
Steven Pinker (31:45.200)
think to emphasize, and I think you're very well positioned as an engineer to remind people about
Steven Pinker (31:50.480)
the culture of engineering, that it really is safety oriented, that another discussion in
Steven Pinker (31:57.040)
enlightenment now, I plot rates of accidental death from various causes, plane crashes, car
Steven Pinker (32:04.560)
crashes, occupational accidents, even death by lightning strikes. And they all plummet because
Steven Pinker (32:12.640)
the culture of engineering is how do you squeeze out the lethal risks, death by fire, death by
Steven Pinker (32:18.320)
drowning, death by asphyxiation, all of them drastically declined because of advances in
Steven Pinker (32:24.320)
engineering that I got to say, I did not appreciate until I saw those graphs. And it is because
Steven Pinker (32:29.840)
exactly, people like you who stay up at night thinking, oh my God, is what I'm inventing likely
Steven Pinker (32:37.520)
to hurt people and to deploy ingenuity to prevent that from happening. Now, I'm not an engineer,
Steven Pinker (32:43.760)
although I spent 22 years at MIT, so I know something about the culture of engineering.
Steven Pinker (32:48.240)
My understanding is that this is the way you think if you're an engineer. And it's essential
Steven Pinker (32:53.680)
that that culture not be suddenly switched off when it comes to artificial intelligence. So,
Lex Fridman (32:59.200)
I mean, that could be a problem, but is there any reason to think it would be switched off?
Lex Fridman (33:02.560)
I don't think so. And one, there's not enough engineers speaking up for this
Steven Pinker (33:06.960)
way, for the excitement, for the positive view of human nature, what you're trying to create
Steven Pinker (33:13.440)
is positivity. Like everything we try to invent is trying to do good for the world.
Lex Fridman (33:18.240)
But let me ask you about the psychology of negativity. It seems just objectively,
Steven Pinker (33:23.600)
not considering the topic, it seems that being negative about the future makes you sound smarter
Steven Pinker (33:28.480)
than being positive about the future, irregardless of topic. Am I correct in this observation? And
Lex Fridman (33:34.320)
if so, why do you think that is? Yeah, I think there is that phenomenon that,
Steven Pinker (33:40.080)
as Tom Lehrer, the satirist said, always predict the worst and you'll be hailed as a prophet.
Steven Pinker (33:45.360)
It may be part of our overall negativity bias. We are as a species more attuned to the negative
Steven Pinker (33:52.240)
than the positive. We dread losses more than we enjoy gains. And that might open up a space for
Steven Pinker (34:02.480)
prophets to remind us of harms and risks and losses that we may have overlooked.
Lex Fridman (34:07.680)
So I think there is that asymmetry. So you've written some of my favorite books
Steven Pinker (34:16.080)
all over the place. So starting from Enlightenment Now to The Better Ages of Our Nature,
Steven Pinker (34:21.600)
Blank Slate, How the Mind Works, the one about language, Language Instinct. Bill Gates,
Steven Pinker (34:29.200)
big fan too, said of your most recent book that it's my new favorite book of all time.
Lex Fridman (34:37.440)
So for you as an author, what was a book early on in your life that had a profound impact on the
Steven Pinker (34:43.840)
way you saw the world? Certainly this book, Enlightenment Now, was influenced by David
Steven Pinker (34:49.040)
Deutsch's The Beginning of Infinity, a rather deep reflection on knowledge and the power of
Steven Pinker (34:55.920)
knowledge to improve the human condition. And with bits of wisdom such as that problems are
Steven Pinker (35:02.320)
inevitable but problems are solvable given the right knowledge and that solutions create new
Steven Pinker (35:07.440)
problems that have to be solved in their turn. That's I think a kind of wisdom about the human
Steven Pinker (35:11.920)
condition that influenced the writing of this book. There are some books that are excellent
Lex Fridman (35:16.080)
but obscure, some of which I have on a page on my website. I read a book called The History of Force,
Steven Pinker (35:22.800)
self published by a political scientist named James Payne on the historical decline of violence
Lex Fridman (35:27.840)
and that was one of the inspirations for The Better Angels of Our Nature.
Lex Fridman (35:33.600)
What about early on? If you look back when you were maybe a teenager?
Steven Pinker (35:38.160)
I loved a book called One, Two, Three, Infinity. When I was a young adult I read that book by
Steven Pinker (35:43.040)
George Gamow, the physicist, which had very accessible and humorous explanations of
Steven Pinker (35:48.960)
relativity, of number theory, of dimensionality, high multiple dimensional spaces in a way that I
Steven Pinker (35:59.360)
think is still delightful 70 years after it was published. I like the Time Life Science series.
Steven Pinker (36:06.080)
These are books that would arrive every month that my mother subscribed to, each one on a different
Steven Pinker (36:11.280)
topic. One would be on electricity, one would be on forests, one would be on evolution and then one
Steven Pinker (36:17.280)
was on the mind. I was just intrigued that there could be a science of mind and that book I would
Steven Pinker (36:24.240)
cite as an influence as well. Then later on... That's when you fell in love with the idea of
Steven Pinker (36:28.480)
studying the mind? Was that the thing that grabbed you? It was one of the things I would say. I read
Steven Pinker (36:35.040)
as a college student the book Reflections on Language by Noam Chomsky. I spent most of his
Steven Pinker (36:41.360)
career here at MIT. Richard Dawkins, two books, The Blind Watchmaker and The Selfish Gene,
Steven Pinker (36:47.440)
were enormously influential, mainly for the content but also for the writing style, the
Steven Pinker (36:55.040)
ability to explain abstract concepts in lively prose. Stephen Jay Gould's first collection,
Steven Pinker (37:02.480)
Ever Since Darwin, also an excellent example of lively writing. George Miller, a psychologist that
Steven Pinker (37:10.240)
most psychologists are familiar with, came up with the idea that human memory has a capacity of
Steven Pinker (37:16.160)
seven plus or minus two chunks. That's probably his biggest claim to fame. But he wrote a couple
Steven Pinker (37:20.640)
of books on language and communication that I read as an undergraduate. Again, beautifully written
Lex Fridman (37:25.840)
and intellectually deep. Wonderful. Stephen, thank you so much for taking the time today.
Lex Fridman (37:31.920)
My pleasure. Thanks a lot, Lex.
🔗 相关节目