WEBVTT

00:00:00.360 --> 00:00:03.277
You actually think AI's
under hyped. Why is that?

00:00:03.360 --> 00:00:06.797
>> And I'll tell I'll tell you why. Most
of you think of AI as I'll just use the

00:00:06.880 --> 00:00:10.677
general term as chat GPT. For most of you,
chat GPT was the moment where you

00:00:10.760 --> 00:00:14.237
said "Oh my god,
this thing writes." And it

00:00:14.320 --> 00:00:18.597
makes mistakes, but it's so brilliantly
verbal. Right? That was certainly my

00:00:18.680 --> 00:00:20.437
reaction. Most people
that I knew did that.

00:00:20.520 --> 00:00:24.077
>> It's visceral, yeah.
>> This was 2 years ago.

00:00:24.160 --> 00:00:28.037
Since then, the gains in what is is
called reinforcement learning, which is

00:00:28.120 --> 00:00:32.667
what AlphaGo helped invent and so
forth, allow us to do planning.

00:00:32.750 --> 00:00:32.997
>> [music]

00:00:33.080 --> 00:00:37.557
>> And a good example is look at OpenAI 03

00:00:37.640 --> 00:00:40.597
or DeepSeek R1, and you can see how it

00:00:40.680 --> 00:00:43.277
goes forward and back, forward and back,

00:00:43.360 --> 00:00:46.237
forward and back. It's extraordinary. In

00:00:46.320 --> 00:00:51.357
my case, I bought a rocket company because
it was like interesting, and I know

00:00:51.440 --> 00:00:54.957
>> As one does. >> As one does.
And it's it's an area that

00:00:55.040 --> 00:00:59.277
I'm not an expert in, and I want to be
an expert, so I'm using deep research.

00:00:59.360 --> 00:01:03.717
And these systems are spending 15
minutes writing these deep papers. This

00:01:03.800 --> 00:01:08.477
is true for most of them. Do you have
any idea how much computation 15 minutes

00:01:08.560 --> 00:01:12.757
of these supercomputers is?
It's extraordinary. So, you're seeing the

00:01:12.840 --> 00:01:16.717
arrival, the shift from language
to language, then you have language to

00:01:16.800 --> 00:01:18.917
sequence, which is how biology is done.

00:01:19.000 --> 00:01:23.077
Now you're [music] doing
essentially planning and strategy.

00:01:23.160 --> 00:01:27.597
The eventual state of this >> [music]
>> is the computers running all business

00:01:27.680 --> 00:01:31.157
processes, right? So, you have an agent
to do this, an agent to do this, an

00:01:31.240 --> 00:01:34.837
agent to do this, an agent to do this,
and you concatenate them together, and

00:01:34.920 --> 00:01:38.277
they speak language >> [music] >>
among each other. They typically speak

00:01:38.360 --> 00:01:43.157
English language. >> I mean,
speaking of of just the sheer

00:01:43.240 --> 00:01:47.277
compute requirements of these systems,
[music] let's talk about scale briefly.

00:01:47.360 --> 00:01:50.357
You know, I kind of think of these AI
systems as Hungry Hungry Hippos. They

00:01:50.440 --> 00:01:54.237
seemingly soak up all the data
and compute that we throw at them. They've

00:01:54.320 --> 00:01:58.517
already digested all the tokens on the
public internet, and it seems we can't

00:01:58.600 --> 00:02:01.237
build data centers fast enough.

00:02:01.320 --> 00:02:05.797
What do you think the real limits are,
and how do we get ahead of them before

00:02:05.880 --> 00:02:08.117
they start throttling AI progress?

00:02:08.200 --> 00:02:09.917
>> So, there's a real
limit in energy. Give

00:02:10.000 --> 00:02:13.877
an example, there's one calculation,
and I testified on this this week in

00:02:13.960 --> 00:02:15.197
Congress,

00:02:15.280 --> 00:02:18.237
that we need another 90 gigawatts of

00:02:18.320 --> 00:02:22.277
power in America. My answer, by the way,

00:02:22.360 --> 00:02:27.277
is think Canada, right? Nice people,
full of hydroelectric power, but that's

00:02:27.360 --> 00:02:30.157
apparently not
the political mood right now.

00:02:30.240 --> 00:02:31.077
Sorry.

00:02:31.160 --> 00:02:35.157
So, 90 gigawatts is 90 nuclear power

00:02:35.240 --> 00:02:39.277
[music] plants in America.
Not happening. We're building zero,

00:02:39.360 --> 00:02:42.957
right? How are we going to get all that
power? This is a major, major national

00:02:43.040 --> 00:02:45.997
issue. Um you can use
the Arab world, which is

00:02:46.080 --> 00:02:50.997
busy building 5 to 10 gigawatts of data
centers. India is considering a 10

00:02:51.080 --> 00:02:52.637
gigawatt data center.

00:02:52.720 --> 00:02:55.237
To understand how big gigawatts are, is

00:02:55.320 --> 00:02:58.437
think cities per data center. That's how

00:02:58.520 --> 00:03:03.517
much power these things need. And the
people look at it, and they say, "Well,

00:03:03.600 --> 00:03:06.557
there's lots of algorithmic
improvements, and you will need less

00:03:06.640 --> 00:03:10.397
power." There's an old rule,
I'm old enough to remember,

00:03:10.480 --> 00:03:14.457
right? Grove giveth, Gates taketh away.

00:03:14.540 --> 00:03:17.317
>> [laughter] >> Okay?
The hardware just gets faster and

00:03:17.400 --> 00:03:20.157
faster. The physicists are amazing.

00:03:20.240 --> 00:03:23.957
Just incredible what they've been able
to do. And us software people, we just

00:03:24.040 --> 00:03:26.237
use it, and use it, and use it.

00:03:26.320 --> 00:03:30.757
And when you look at planning, to at
least in today's algorithms, it's back

00:03:30.840 --> 00:03:35.037
and forth, and try this, and try And just
watch it yourself. Their estimates,

00:03:35.120 --> 00:03:39.597
and you know this from from Industry
Horowitz reports, uh it's been well

00:03:39.680 --> 00:03:42.557
studied, [music] that there's
an increase in at least a

00:03:42.640 --> 00:03:46.837
factor of a hundred, maybe a factor of
thousand in computation required just to

00:03:46.920 --> 00:03:48.557
do the kind [music] of planning.

00:03:48.640 --> 00:03:51.997
There the technology goes
from essentially deep learning to

00:03:52.080 --> 00:03:55.357
reinforcement learning to something
called test time compute, where not only

00:03:55.440 --> 00:03:58.917
are you doing planning, but you're also
learning while you're doing planning.

00:03:59.000 --> 00:04:01.157
That is the, if you will, the zenith or

00:04:01.240 --> 00:04:03.917
what have you, of computation needs.

00:04:04.000 --> 00:04:09.317
That's problem number one, electricity
and hardware. Problem number two is we

00:04:09.400 --> 00:04:12.837
ran out of data, so we
have to start generating it,

00:04:12.920 --> 00:04:15.957
but we can easily do that because
that's one of the functions.

00:04:16.040 --> 00:04:19.757
And then the third question that I
don't understand is what's the limit of

00:04:19.840 --> 00:04:22.357
knowledge? Um I'll I'll
give you an example. Let's

00:04:22.440 --> 00:04:26.437
imagine we are collectively all of the
computers in the world, and we're all

00:04:26.520 --> 00:04:29.877
thinking, and we're all thinking
based on knowledge that exists that was

00:04:29.960 --> 00:04:32.397
previously invented. How do we invent

00:04:32.480 --> 00:04:35.517
something completely [music] new?

00:04:35.600 --> 00:04:39.317
So, Einstein. So, when you
study the way scientific

00:04:39.400 --> 00:04:42.637
discovery works, biology,
math, so forth and so on,

00:04:42.720 --> 00:04:48.157
what typically happens is a truly
brilliant human being looks at one area

00:04:48.240 --> 00:04:52.237
and says, "I see a pattern
that's in a completely

00:04:52.320 --> 00:04:55.957
different area. Has nothing to do with
the first one. It's the same pattern."

00:04:56.040 --> 00:04:58.197
And they take the tools from one, and

00:04:58.280 --> 00:05:00.437
they apply it to another. Today, our

00:05:00.520 --> 00:05:03.117
systems cannot do that. If we can get

00:05:03.200 --> 00:05:05.277
through that, um I'm working on this.

00:05:05.360 --> 00:05:07.677
It's a general technical term for this

00:05:07.760 --> 00:05:12.317
is non-stationarity of objectives.
The rules are keep changing.

00:05:12.400 --> 00:05:15.597
Um we will see if we can solve that
problem. If we can solve that, we're

00:05:15.680 --> 00:05:17.957
going to need even more data centers,

00:05:18.040 --> 00:05:22.717
and we'll also be able to invent
completely new schools of scientific and

00:05:22.800 --> 00:05:25.920
intellectual [music] thought.