episode
stringlengths
45
100
text
stringlengths
1
528
timestamp_link
stringlengths
56
56
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
So for example, we don't really have an exact substitute for wood. We still build houses out
https://karpathy.ai/lexicap/0003-large.html#00:07:54.560
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
of wood. We still build furniture out of wood. We like the look. We like the feel. It has certain
https://karpathy.ai/lexicap/0003-large.html#00:08:00.480
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
properties that synthetics don't. It's not that there's anything magical or mysterious about wood.
https://karpathy.ai/lexicap/0003-large.html#00:08:05.040
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
It's just that the extra steps of duplicating everything about wood is something we just haven't
https://karpathy.ai/lexicap/0003-large.html#00:08:11.120
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
bothered because we have wood. Likewise, say cotton. I'm wearing cotton clothing now. It feels
https://karpathy.ai/lexicap/0003-large.html#00:08:17.600
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
much better than polyester. It's not that cotton has something magic in it. It's not that we couldn't
https://karpathy.ai/lexicap/0003-large.html#00:08:21.600
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
ever synthesize something exactly like cotton, but at some point it's just not worth it. We've got
https://karpathy.ai/lexicap/0003-large.html#00:08:29.760
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
cotton. Likewise, in the case of human intelligence, the goal of making an artificial system that is
https://karpathy.ai/lexicap/0003-large.html#00:08:35.920
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
exactly like the human brain is a goal that we probably know is going to pursue to the bitter
https://karpathy.ai/lexicap/0003-large.html#00:08:42.000
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
end, I suspect, because if you want tools that do things better than humans, you're not going to
https://karpathy.ai/lexicap/0003-large.html#00:08:47.840
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
care whether it does something like humans. So for example, diagnosing cancer or predicting the
https://karpathy.ai/lexicap/0003-large.html#00:08:53.120
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
weather, why set humans as your benchmark? But in general, I suspect you also believe
https://karpathy.ai/lexicap/0003-large.html#00:08:58.240
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
that even if the human should not be a benchmark and we don't want to imitate humans in their
https://karpathy.ai/lexicap/0003-large.html#00:09:05.840
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
system, there's a lot to be learned about how to create an artificial intelligence system by
https://karpathy.ai/lexicap/0003-large.html#00:09:10.480
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
studying the human. Yeah, I think that's right. In the same way that to build flying machines,
https://karpathy.ai/lexicap/0003-large.html#00:09:15.520
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
we want to understand the laws of aerodynamics, including birds, but not mimic the birds,
https://karpathy.ai/lexicap/0003-large.html#00:09:22.480
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
but they're the same laws. You have a view on AI, artificial intelligence, and safety
https://karpathy.ai/lexicap/0003-large.html#00:09:27.760
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
that, from my perspective, is refreshingly rational or perhaps more importantly, has elements
https://karpathy.ai/lexicap/0003-large.html#00:09:35.440
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
of positivity to it, which I think can be inspiring and empowering as opposed to paralyzing.
https://karpathy.ai/lexicap/0003-large.html#00:09:47.040
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
For many people, including AI researchers, the eventual existential threat of AI is obvious,
https://karpathy.ai/lexicap/0003-large.html#00:09:53.600
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
not only possible, but obvious. And for many others, including AI researchers, the threat
https://karpathy.ai/lexicap/0003-large.html#00:09:59.840
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
is not obvious. So Elon Musk is famously in the highly concerned about AI camp, saying things like
https://karpathy.ai/lexicap/0003-large.html#00:10:05.840
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
AI is far more dangerous than nuclear weapons, and that AI will likely destroy human civilization.
https://karpathy.ai/lexicap/0003-large.html#00:10:14.640
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
Human civilization. So in February, he said that if Elon was really serious about AI, the threat
https://karpathy.ai/lexicap/0003-large.html#00:10:21.200
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
of AI, he would stop building self driving cars that he's doing very successfully as part of Tesla.
https://karpathy.ai/lexicap/0003-large.html#00:10:29.360
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
Then he said, wow, if even Pinker doesn't understand the difference between narrow AI,
https://karpathy.ai/lexicap/0003-large.html#00:10:35.760
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
like a car and general AI, when the latter literally has a million times more compute power
https://karpathy.ai/lexicap/0003-large.html#00:10:40.800
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
and an open ended utility function, humanity is in deep trouble. So first, what did you mean by
https://karpathy.ai/lexicap/0003-large.html#00:10:47.280
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
the statement about Elon Musk should stop building self driving cars if he's deeply concerned?
https://karpathy.ai/lexicap/0003-large.html#00:10:54.160
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
Not the last time that Elon Musk has fired off an intemperate tweet.
https://karpathy.ai/lexicap/0003-large.html#00:11:00.080
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
Well, we live in a world where Twitter has power.
https://karpathy.ai/lexicap/0003-large.html#00:11:04.320
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
Yes. Yeah, I think there are two kinds of existential threat that have been discussed
https://karpathy.ai/lexicap/0003-large.html#00:11:07.680
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
in connection with artificial intelligence, and I think that they're both incoherent.
https://karpathy.ai/lexicap/0003-large.html#00:11:16.640
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
One of them is a vague fear of AI takeover, that just as we subjugated animals and less technologically
https://karpathy.ai/lexicap/0003-large.html#00:11:20.480
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
advanced peoples, so if we build something that's more advanced than us, it will inevitably turn us
https://karpathy.ai/lexicap/0003-large.html#00:11:29.520
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
into pets or slaves or domesticated animal equivalents. I think this confuses intelligence
https://karpathy.ai/lexicap/0003-large.html#00:11:34.640
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
with a will to power, that it so happens that in the intelligence system we are most familiar with,
https://karpathy.ai/lexicap/0003-large.html#00:11:42.320
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
namely homo sapiens, we are products of natural selection, which is a competitive process,
https://karpathy.ai/lexicap/0003-large.html#00:11:49.200
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
and so bundled together with our problem solving capacity are a number of nasty traits like
https://karpathy.ai/lexicap/0003-large.html#00:11:54.160
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
dominance and exploitation and maximization of power and glory and resources and influence.
https://karpathy.ai/lexicap/0003-large.html#00:12:00.320
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
There's no reason to think that sheer problem solving capability will set that as one of its
https://karpathy.ai/lexicap/0003-large.html#00:12:08.720
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
goals. Its goals will be whatever we set its goals as, and as long as someone isn't building a
https://karpathy.ai/lexicap/0003-large.html#00:12:13.120
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
megalomaniacal artificial intelligence, then there's no reason to think that it would naturally
https://karpathy.ai/lexicap/0003-large.html#00:12:18.720
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
evolve in that direction. Now, you might say, well, what if we gave it the goal of maximizing
https://karpathy.ai/lexicap/0003-large.html#00:12:24.320
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
its own power source? That's a pretty stupid goal to give an autonomous system. You don't give it
https://karpathy.ai/lexicap/0003-large.html#00:12:28.960
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
that goal. I mean, that's just self evidently idiotic. So if you look at the history of the
https://karpathy.ai/lexicap/0003-large.html#00:12:34.880
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
world, there's been a lot of opportunities where engineers could instill in a system
https://karpathy.ai/lexicap/0003-large.html#00:12:40.720
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
destructive power and they choose not to because that's the natural process of engineering.
https://karpathy.ai/lexicap/0003-large.html#00:12:45.120
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
Well, except for weapons. I mean, if you're building a weapon, its goal is to destroy people,
https://karpathy.ai/lexicap/0003-large.html#00:12:49.440
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
and so I think there are good reasons to not build certain kinds of weapons. I think building
https://karpathy.ai/lexicap/0003-large.html#00:12:53.680
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
nuclear weapons was a massive mistake. You do. So maybe pause on that because that is one of
https://karpathy.ai/lexicap/0003-large.html#00:12:58.560
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
the serious threats. Do you think that it was a mistake in a sense that it should have been
https://karpathy.ai/lexicap/0003-large.html#00:13:06.480
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
stopped early on? Or do you think it's just an unfortunate event of invention that this was
https://karpathy.ai/lexicap/0003-large.html#00:13:12.800
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
invented? Do you think it's possible to stop? I guess is the question. It's hard to rewind the
https://karpathy.ai/lexicap/0003-large.html#00:13:19.200
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
clock because of course it was invented in the context of World War II and the fear that the
https://karpathy.ai/lexicap/0003-large.html#00:13:23.280
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
Nazis might develop one first. Then once it was initiated for that reason, it was hard to turn
https://karpathy.ai/lexicap/0003-large.html#00:13:28.320
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
off, especially since winning the war against the Japanese and the Nazis was such an overwhelming
https://karpathy.ai/lexicap/0003-large.html#00:13:35.120
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
goal of every responsible person that there's just nothing that people wouldn't have done then
https://karpathy.ai/lexicap/0003-large.html#00:13:42.080
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
to ensure victory. It's quite possible if World War II hadn't happened that nuclear weapons
https://karpathy.ai/lexicap/0003-large.html#00:13:47.120
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
wouldn't have been invented. We can't know, but I don't think it was by any means a necessity,
https://karpathy.ai/lexicap/0003-large.html#00:13:52.560
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
any more than some of the other weapon systems that were envisioned but never implemented,
https://karpathy.ai/lexicap/0003-large.html#00:13:57.440
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
like planes that would disperse poison gas over cities like crop dusters or systems to try to
https://karpathy.ai/lexicap/0003-large.html#00:14:02.720
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
create earthquakes and tsunamis in enemy countries, to weaponize the weather,
https://karpathy.ai/lexicap/0003-large.html#00:14:10.560
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
weaponize solar flares, all kinds of crazy schemes that we thought the better of.
https://karpathy.ai/lexicap/0003-large.html#00:14:16.000
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
I think analogies between nuclear weapons and artificial intelligence are fundamentally
https://karpathy.ai/lexicap/0003-large.html#00:14:21.120
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
misguided because the whole point of nuclear weapons is to destroy things. The point of
https://karpathy.ai/lexicap/0003-large.html#00:14:25.840
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
artificial intelligence is not to destroy things. So the analogy is misleading.
https://karpathy.ai/lexicap/0003-large.html#00:14:30.560
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
So there's two artificial intelligence you mentioned. The first one I guess is highly
https://karpathy.ai/lexicap/0003-large.html#00:14:36.080
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
intelligent or power hungry.
https://karpathy.ai/lexicap/0003-large.html#00:14:39.920
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
Yeah, it's a system that we design ourselves where we give it the goals. Goals are external to
https://karpathy.ai/lexicap/0003-large.html#00:14:42.080
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
the means to attain the goals. If we don't design an artificially intelligent system to
https://karpathy.ai/lexicap/0003-large.html#00:14:46.800
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
maximize dominance, then it won't maximize dominance. It's just that we're so familiar
https://karpathy.ai/lexicap/0003-large.html#00:14:55.200
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
with homo sapiens where these two traits come bundled together, particularly in men,
https://karpathy.ai/lexicap/0003-large.html#00:15:00.800
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
that we are apt to confuse high intelligence with a will to power, but that's just an error.
https://karpathy.ai/lexicap/0003-large.html#00:15:06.320
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
The other fear is that will be collateral damage that will give artificial intelligence a goal
https://karpathy.ai/lexicap/0003-large.html#00:15:15.520
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
like make paper clips and it will pursue that goal so brilliantly that before we can stop it,
https://karpathy.ai/lexicap/0003-large.html#00:15:21.440
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
it turns us into paper clips. We'll give it the goal of curing cancer and it will turn us into
https://karpathy.ai/lexicap/0003-large.html#00:15:27.440
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
guinea pigs for lethal experiments or give it the goal of world peace and its conception of world
https://karpathy.ai/lexicap/0003-large.html#00:15:32.800
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
peace is no people, therefore no fighting and so it will kill us all. Now I think these are utterly
https://karpathy.ai/lexicap/0003-large.html#00:15:38.720
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
fanciful. In fact, I think they're actually self defeating. They first of all assume that we're
https://karpathy.ai/lexicap/0003-large.html#00:15:43.680
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
going to be so brilliant that we can design an artificial intelligence that can cure cancer,
https://karpathy.ai/lexicap/0003-large.html#00:15:49.040
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
but so stupid that we don't specify what we mean by curing cancer in enough detail that it won't
https://karpathy.ai/lexicap/0003-large.html#00:15:53.600
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
kill us in the process and it assumes that the system will be so smart that it can cure cancer,
https://karpathy.ai/lexicap/0003-large.html#00:15:59.600
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
but so idiotic that it can't figure out that what we mean by curing cancer is not killing everyone.
https://karpathy.ai/lexicap/0003-large.html#00:16:06.240
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
I think that the collateral damage scenario, the value alignment problem is also based on
https://karpathy.ai/lexicap/0003-large.html#00:16:12.880
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
a misconception. So one of the challenges, of course, we don't know how to build either system
https://karpathy.ai/lexicap/0003-large.html#00:16:18.320
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
currently or are we even close to knowing? Of course, those things can change overnight,
https://karpathy.ai/lexicap/0003-large.html#00:16:23.200
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
but at this time, theorizing about it is very challenging in either direction. So that's
https://karpathy.ai/lexicap/0003-large.html#00:16:27.440
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
probably at the core of the problem is without that ability to reason about the real engineering
https://karpathy.ai/lexicap/0003-large.html#00:16:33.840
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
things here at hand is your imagination runs away with things. Exactly. But let me sort of ask,
https://karpathy.ai/lexicap/0003-large.html#00:16:39.600
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
what do you think was the motivation, the thought process of Elon Musk? I build autonomous vehicles,
https://karpathy.ai/lexicap/0003-large.html#00:16:45.120
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
I study autonomous vehicles, I study Tesla autopilot. I think it is one of the greatest
https://karpathy.ai/lexicap/0003-large.html#00:16:52.320
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
currently large scale application of artificial intelligence in the world. It has potentially a
https://karpathy.ai/lexicap/0003-large.html#00:16:57.680
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
very positive impact on society. So how does a person who's creating this very good quote unquote
https://karpathy.ai/lexicap/0003-large.html#00:17:04.400
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
narrow AI system also seem to be so concerned about this other general AI? What do you think
https://karpathy.ai/lexicap/0003-large.html#00:17:10.880
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
is the motivation there? What do you think is the thing? Well, you probably have to ask him,
https://karpathy.ai/lexicap/0003-large.html#00:17:19.280
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
but there, and he is notoriously flamboyant, impulsive to the, as we have just seen,
https://karpathy.ai/lexicap/0003-large.html#00:17:23.040
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
to the detriment of his own goals of the health of the company. So I don't know what's going on
https://karpathy.ai/lexicap/0003-large.html#00:17:31.520
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
in his mind. You probably have to ask him, but I don't think the, and I don't think the distinction
https://karpathy.ai/lexicap/0003-large.html#00:17:37.360
Steven Pinker: AI in the Age of Reason | Lex Fridman Podcast #3
between special purpose AI and so called general AI is relevant that in the same way that special
https://karpathy.ai/lexicap/0003-large.html#00:17:42.560