1
00:00:00,000 --> 00:00:02,877
What's it's going to mean to to be
[music] human like 50 years from now?

2
00:00:02,960 --> 00:00:05,957
>> I think the world is finally taking
seriously, and we've been waiting for

3
00:00:06,040 --> 00:00:09,717
this for years, that social contract is
going to have to evolve and maybe sooner

4
00:00:09,800 --> 00:00:12,797
than people thought. >> What are
your predictions for when this

5
00:00:12,880 --> 00:00:16,397
breaks [music] 85%? >> Yeah,
I think it really depends how much

6
00:00:16,480 --> 00:00:20,557
effort the frontier labs are putting
into targeting Arc AGI 3 explicitly.

7
00:00:20,640 --> 00:00:25,117
With the new set of models that I think
will come out, the impact that [music]

8
00:00:25,200 --> 00:00:28,397
the technology will have on accelerating
science and accelerating the economy. We

9
00:00:28,480 --> 00:00:31,477
may be wrong, we may be over optimistic,
but it feels like it's about to be

10
00:00:31,560 --> 00:00:34,077
absolutely massive. There are a lot
of interesting projects, [music] and yet

11
00:00:34,160 --> 00:00:37,597
there's this one new thing
that's really working.

12
00:00:37,680 --> 00:00:41,517
Hi guys, I'm D, a partner at Menlo
Ventures. I'm one of the early investors

13
00:00:41,600 --> 00:00:45,797
in India. And we are
investors in Anthropic.

14
00:00:45,880 --> 00:00:49,197
So not to say that will

15
00:00:49,280 --> 00:00:52,277
color the discussion at all.

16
00:00:52,360 --> 00:00:57,917
Um but yeah, let's get started. Um
The first thing I wanted to ask you guys

17
00:00:58,000 --> 00:01:02,557
is, you know both of you are parents,

18
00:01:02,640 --> 00:01:06,397
not just founders and CEOs of companies.

19
00:01:06,480 --> 00:01:09,677
And today when I go
and talk to a lot of parents

20
00:01:09,760 --> 00:01:13,437
they're quite concerned about
their kids and the future.

21
00:01:13,520 --> 00:01:17,357
As as fathers what do you
think about the most when

22
00:01:17,440 --> 00:01:21,437
raising your kids today? Is it any
different from before? What really

23
00:01:21,520 --> 00:01:25,277
matters to you when it
comes to raising your kids?

24
00:01:25,360 --> 00:01:27,477
Maybe we'll start with Francois.

25
00:01:27,560 --> 00:01:30,397
Yeah, and it's in my mind a lot. It's in
my mind a lot. It's like clearly they're

26
00:01:30,480 --> 00:01:33,717
going to grow up in a world that's very
different from the world in which I grew

27
00:01:33,800 --> 00:01:37,397
up in. It's expensive child. And yeah,
I think about it like the obviously their

28
00:01:37,480 --> 00:01:40,557
their relationship to technology
is going to be completely different. You

29
00:01:40,640 --> 00:01:43,637
know, they're they're they're going
to grow up in this Star Trek world where

30
00:01:43,720 --> 00:01:46,357
they can just, you know, talk to a
computer and and have the computer do

31
00:01:46,440 --> 00:01:50,797
things. Uh while back in my day,
you know, we had we had to write code like

32
00:01:50,880 --> 00:01:55,157
character by character. Um yeah,
and so actually I'm a I'm a bit

33
00:01:55,240 --> 00:01:58,637
concerned as well because, you know, I'm
I'm wondering what's going to happen to

34
00:01:58,720 --> 00:02:03,077
human human agency. And and and our
ability to you know think independently.

35
00:02:03,160 --> 00:02:07,877
Like are we are we just going to outsource
our minds? Like, you know, our

36
00:02:07,960 --> 00:02:11,877
our thoughts, our cognition, and all our
abilities to the machine? What what are

37
00:02:11,960 --> 00:02:14,357
we going to retain? You know, what
what's what's it going to mean to to be

38
00:02:14,440 --> 00:02:18,997
human in like not in 10 years from now,
but like 50 years from now? Like my son,

39
00:02:19,080 --> 00:02:22,717
my daughter, they're they're they're
going to live to see the the 22nd

40
00:02:22,800 --> 00:02:27,037
century, right? That's that's
that's something I think about a lot.

41
00:02:27,120 --> 00:02:30,597
What about you, Sam? You know, actually

42
00:02:30,680 --> 00:02:34,717
I think it'll be great. I I understand
that throughout history parents have

43
00:02:34,800 --> 00:02:38,077
always freaked out about the rate of
change, wondered how life is going to be

44
00:02:38,160 --> 00:02:40,757
for their kids. I think it
is awesome that our kids are

45
00:02:40,840 --> 00:02:43,557
going to grow up in this Star Trek
world where you just ask a computer for

46
00:02:43,640 --> 00:02:48,237
anything. Uh and it does it,
and I think they will

47
00:02:48,320 --> 00:02:52,397
you know, look at us like I kind
of looked at my parents with this

48
00:02:52,480 --> 00:02:55,517
like, "Oh man, you really had
to do things before computers?"

49
00:02:55,600 --> 00:02:57,997
Uh and my parents worried that I wasn't
going to learn anything either cuz I was

50
00:02:58,080 --> 00:03:01,477
just going to look everything up
on Google. It turns out I was able to have

51
00:03:01,560 --> 00:03:04,397
a fulfilling career of a kind they hadn't

52
00:03:04,480 --> 00:03:07,957
they didn't have a model for. And

53
00:03:08,040 --> 00:03:11,437
I assume that our kids will
have the same experience

54
00:03:11,520 --> 00:03:14,557
and they will do things um

55
00:03:14,640 --> 00:03:16,557
that absolutely astonish us.

56
00:03:16,640 --> 00:03:19,757
That they will look back on us with look

57
00:03:19,840 --> 00:03:24,997
back on our lives with some like pity
is maybe too strong of a word, but

58
00:03:25,080 --> 00:03:26,757
man, you all had it hard and bad and

59
00:03:26,840 --> 00:03:29,117
miserable. And this the quality of life

60
00:03:29,200 --> 00:03:34,037
will be so incredible.
Um I I know actually I think all those

61
00:03:34,120 --> 00:03:39,317
things. I think it's
amazing that my kids will

62
00:03:39,400 --> 00:03:44,757
never be smarter than a computer, and they
will never expect that a computer can't

63
00:03:44,840 --> 00:03:48,317
do these incredible things, and they
will they will achieve and they will

64
00:03:48,400 --> 00:03:52,917
grow up expecting to achieve these amazing

65
00:03:53,000 --> 00:03:57,037
feats, and I think they'll have more
agency and more will than any of us do,

66
00:03:57,120 --> 00:04:00,957
and they'll like they'll grow up
expecting that. It won't be weird. Um

67
00:04:01,040 --> 00:04:03,317
and they'll develop with that way
in the same way that we developed with the

68
00:04:03,400 --> 00:04:07,557
computers and thought of all the things we
could do that our parents didn't think of.

69
00:04:07,640 --> 00:04:12,717
Emotionally I'm like oof.

70
00:04:12,800 --> 00:04:16,077
You know what's this
really going to be like?

71
00:04:16,160 --> 00:04:19,317
Do you think just a quick follow-up
on that? Do you think in that world, you

72
00:04:19,400 --> 00:04:22,277
know, when we grew up, intelligence
probably mattered a lot. You know, our

73
00:04:22,360 --> 00:04:24,317
parents probably were like, "Be smart."

74
00:04:24,400 --> 00:04:27,157
You know, you guys went to some pretty
good schools, you learned a lot. Do you

75
00:04:27,240 --> 00:04:31,317
think that still continues to matter
in this world or not so much or maybe a

76
00:04:31,400 --> 00:04:34,837
different kind of intelligence?
What are your thoughts?

77
00:04:34,920 --> 00:04:39,557
So intelligence is really the ability
to adapt, and the the more rapidly the

78
00:04:39,640 --> 00:04:42,597
world changes, the more important
it is to be able to adapt to it. So

79
00:04:42,680 --> 00:04:47,157
personally, I trust my kids that, you
know they will see a level of change uh

80
00:04:47,240 --> 00:04:50,997
technological change and social change
that I haven't seen in my life, but I I

81
00:04:51,080 --> 00:04:53,477
trust them to adapt like kids kids are

82
00:04:53,560 --> 00:04:56,157
the most adaptable ones in the world.

83
00:04:56,240 --> 00:04:59,437
I'd love to like get a peek into your

84
00:04:59,520 --> 00:05:03,517
into your lives. Like what's something
that's happened in the last week or two?

85
00:05:03,600 --> 00:05:05,717
Like a small anecdote you could share

86
00:05:05,800 --> 00:05:11,560
that just tell us about tell us
about what goes on behind the scenes.

87
00:05:13,000 --> 00:05:18,157
Sure. I mean I can yeah. [laughter]
So what one thing I've I've been doing

88
00:05:18,240 --> 00:05:22,357
for the last few weeks with my son.
So my son is a huge Minecraft fan. Like

89
00:05:22,440 --> 00:05:26,077
he's he's he's turning five soon, and so
that's it's playing so much Minecraft

90
00:05:26,160 --> 00:05:29,837
and watching Minecraft videos on YouTube
and so on. And what I've been doing is

91
00:05:29,920 --> 00:05:33,637
that I've been live coding effectively

92
00:05:33,720 --> 00:05:37,597
sort of like simple pretty
future complete Minecraft clone.

93
00:05:37,680 --> 00:05:41,717
And in fact using Codex,
so thank you Sam for that. Um

94
00:05:41,800 --> 00:05:44,997
and it's been it's been so fun.
And basically so my son is telling me, you

95
00:05:45,080 --> 00:05:47,517
know, what what what he wants to see
in the game, what features he wants, and

96
00:05:47,600 --> 00:05:51,357
he's he's pretty creative about it.
And then I tell Codex to do it. And I

97
00:05:51,440 --> 00:05:54,317
haven't I haven't written a single line
of code. So it's it's it's it's been a

98
00:05:54,400 --> 00:05:57,877
very fun experience. What about you, Sam?

99
00:05:57,960 --> 00:06:00,440
Um

100
00:06:01,160 --> 00:06:04,357
I I think one interesting

101
00:06:04,440 --> 00:06:08,317
like I'm very grateful to get to work
on so many interesting things and

102
00:06:08,400 --> 00:06:11,237
built my kids can't build the Codex yet,
but we can, you know, put blocks

103
00:06:11,320 --> 00:06:13,477
together and stuff. Um that's the most

104
00:06:13,560 --> 00:06:18,317
fun, but the most kind
of intellectually stimulating thing is

105
00:06:18,400 --> 00:06:21,477
I think the world
is finally taking seriously

106
00:06:21,560 --> 00:06:24,637
and we've been waiting for this for years,
that the social contract is going

107
00:06:24,720 --> 00:06:26,437
to have to evolve and maybe sooner than

108
00:06:26,520 --> 00:06:29,197
people thought. And it finally feels to

109
00:06:29,280 --> 00:06:32,117
me like there is a moment where in many

110
00:06:32,200 --> 00:06:36,557
different countries, many different
political parties, many people who

111
00:06:36,640 --> 00:06:41,077
normally would think
very differently about how

112
00:06:41,160 --> 00:06:45,117
the economy should work and how
power structures should be developed

113
00:06:45,200 --> 00:06:49,197
there is a real finally,
I think it's real, desire to

114
00:06:49,280 --> 00:06:52,877
sit down and say, "Let's start talking
about ideas. Let's not just keep saying,

115
00:06:52,960 --> 00:06:55,077
'Man, stuff is really going to change
and this is really going to be a big

116
00:06:55,160 --> 00:06:57,117
effect on the economy.'" But

117
00:06:57,200 --> 00:06:59,877
the Overton window the

118
00:06:59,960 --> 00:07:04,157
interest sort of willingness
to get to work on

119
00:07:04,240 --> 00:07:08,117
this with the people that can actually
make it happen. This last week, few

120
00:07:08,200 --> 00:07:10,757
weeks, whatever, first time I believe
like, "All right, maybe the world is

121
00:07:10,840 --> 00:07:14,917
really going to do this." Um and that's
been an exciting development. That's

122
00:07:15,000 --> 00:07:18,717
awesome. That's awesome. Let's let's
move into some AI questions. This is a

123
00:07:18,800 --> 00:07:22,117
question I just have to ask. I know
Sam probably just saw this benchmark.

124
00:07:22,200 --> 00:07:25,197
Francois, you probably have a good
idea of this or at least a prediction.

125
00:07:25,280 --> 00:07:26,797
You know, Arc AGI, Greg covered this,

126
00:07:26,880 --> 00:07:30,277
right? Remember 2019, 5 years later, O3

127
00:07:30,360 --> 00:07:33,997
preview crushes it, 88%. Arc AGI 2

128
00:07:34,080 --> 00:07:37,437
launches about a year ago.
Um right now I think the top score is

129
00:07:37,520 --> 00:07:39,997
Gemini Deep Thinking 84.6.

130
00:07:40,080 --> 00:07:43,717
Today Arc AGI 3 is 0.37.

131
00:07:43,800 --> 00:07:45,917
What are your predictions for when this

132
00:07:46,000 --> 00:07:50,917
breaks 85%? Yeah, I think
it really depends how much

133
00:07:51,000 --> 00:07:56,037
effort the frontier labs are putting
into targeting Arc AGI 3 explicitly. Um

134
00:07:56,120 --> 00:07:59,357
like for Arc 2 for instance, a lot
of the very fast progress we saw in the in

135
00:07:59,440 --> 00:08:02,717
the late 2025 was really because, you
know, there was just so much effort, so

136
00:08:02,800 --> 00:08:06,837
much compute that was
thrown at the benchmarks.

137
00:08:06,920 --> 00:08:11,477
Um and I think if if the the frontier
labs decide that they're just not going to

138
00:08:11,560 --> 00:08:14,357
look at Arc 3 whatsoever, they're not
going to they're not going to target it,

139
00:08:14,440 --> 00:08:16,277
they're completely they're going
to completely ignore it, they're not going

140
00:08:16,360 --> 00:08:18,917
to let the scores increase organically

141
00:08:19,000 --> 00:08:21,517
just as a function of the G of the model

142
00:08:21,600 --> 00:08:23,797
increasing, I think in this in this

143
00:08:23,880 --> 00:08:28,677
scenario it would take several years
to saturate Arc 3. Uh but what's actually

144
00:08:28,760 --> 00:08:31,877
going to happen in practice is they
they're they're going to throw a ton of

145
00:08:31,960 --> 00:08:35,277
compute at it. Um and uh I would say at

146
00:08:35,360 --> 00:08:38,317
least a year until saturation because

147
00:08:38,400 --> 00:08:42,717
uh it's it's significantly harder than
than Arc 2. Uh we were deliberately

148
00:08:42,800 --> 00:08:47,237
trying to make it, you know, like out
of distribution. Like the concepts, the

149
00:08:47,320 --> 00:08:51,357
dynamics you will see in the in the test
environments are very different than the

150
00:08:51,440 --> 00:08:54,317
ones that you see in the in the public
in the public games. Uh it's going to be

151
00:08:54,400 --> 00:08:56,597
much much harder to put forth, for sure.

152
00:08:56,680 --> 00:08:59,197
So at least a year. Thankfully, we have

153
00:08:59,280 --> 00:09:02,757
a big lab CEO right here. So I guess

154
00:09:02,840 --> 00:09:07,157
this is a kind of freebie question,
but does Arc AGI 3 matter to you, Sam? And

155
00:09:07,240 --> 00:09:07,797
uh

156
00:09:07,880 --> 00:09:11,757
how do you think about Arc AGI 3?

157
00:09:11,840 --> 00:09:15,037
Um yeah, of course. I I think any

158
00:09:15,120 --> 00:09:18,077
benchmarks that are interesting and real

159
00:09:18,160 --> 00:09:22,717
and not currently
saturated matter to us. Um

160
00:09:22,800 --> 00:09:25,357
there aren't many left.
I I mean there are metrics about

161
00:09:25,440 --> 00:09:28,037
discovering new science.
There are benchmarks about discovering new

162
00:09:28,120 --> 00:09:31,077
science. There are benchmarks about
economic acceleration. Those I think

163
00:09:31,160 --> 00:09:35,557
really matter. But there are not
a lot of benchmarks in this sense.

164
00:09:35,640 --> 00:09:39,517
Uh so of course it matters. Uh I thought
that was a very good and very nuanced

165
00:09:39,600 --> 00:09:43,637
answer, which is how long it takes will
depend how hard how much people target

166
00:09:43,720 --> 00:09:46,997
it. Um I hope people don't target
it that much, but I'm sure, you know,

167
00:09:47,080 --> 00:09:49,117
people like scoreboards and they will.

168
00:09:49,200 --> 00:09:53,917
Um is

169
00:09:54,000 --> 00:09:56,237
AGI wasn't supposed to be numbered.

170
00:09:56,320 --> 00:10:00,157
You know, in the way that people first
talked about AGI, it was you have AGI or

171
00:10:00,240 --> 00:10:01,477
not. You're not supposed to have an AGI

172
00:10:01,560 --> 00:10:05,077
one, AGI two, AGI three. And

173
00:10:05,160 --> 00:10:07,317
and yet

174
00:10:07,400 --> 00:10:09,997
we learn how hard it is to precisely

175
00:10:10,080 --> 00:10:14,357
define what AGI is and how we're going
to test for it. And I bet there'll be a

176
00:10:14,440 --> 00:10:16,837
four and a five and then a ASI one and

177
00:10:16,920 --> 00:10:23,600
ASI two and then it'll keep going.
We're working on it. There we go. Um

178
00:10:23,720 --> 00:10:27,397
but I think the main
thing that points to is not

179
00:10:27,480 --> 00:10:31,877
that this team doesn't do incredible
work, which obviously they do, but

180
00:10:31,960 --> 00:10:37,357
the goal posts move. AGI is hard to define
and great benchmarks are hard to make.

181
00:10:37,440 --> 00:10:38,757
Another benchmark that's talked about a

182
00:10:38,840 --> 00:10:41,677
lot that sort of has no ceiling is this

183
00:10:41,760 --> 00:10:44,637
uh METER benchmark, which is how long of

184
00:10:44,720 --> 00:10:47,917
a task can an AI agent autonomously do

185
00:10:48,000 --> 00:10:50,677
with some success rate. So, right now

186
00:10:50,760 --> 00:10:55,637
um there's 12 hours with a above 50%
success rate. There's a model called

187
00:10:55,720 --> 00:10:58,517
Opus that does that. What do you What
do you think of that benchmark? Like, is

188
00:10:58,600 --> 00:11:03,717
that not valid? How's that different
from ARC-AGI? Francois, any thoughts? Uh

189
00:11:03,800 --> 00:11:06,197
I like METER.
I think it's good signal in it.

190
00:11:06,280 --> 00:11:10,757
Um it's also a bit of a noisy signal.
Like, I wouldn't take the claim that uh

191
00:11:10,840 --> 00:11:12,797
current AI can do like 12 hours of

192
00:11:12,880 --> 00:11:16,157
autonomous work. Um and I mean, as a as

193
00:11:16,240 --> 00:11:20,437
a counter claim, I mean, look look at uh
any Oxbury environment. It doesn't take

194
00:11:20,520 --> 00:11:24,237
12 hours to to solve one of them.
It takes uh about 5 minutes like 5 to 10

195
00:11:24,320 --> 00:11:28,157
minutes for a human. And uh current
functional models cannot do that. So,

196
00:11:28,240 --> 00:11:31,717
we're not I don't think we're at 12
hours over arbitrary tasks. So, if you

197
00:11:31,800 --> 00:11:35,277
look at METER, I think uh some of the
issues is if you're looking at longer

198
00:11:35,360 --> 00:11:38,637
time horizons, and it's definitely
the regime that we're in now,

199
00:11:38,720 --> 00:11:42,317
uh it's less reliable because they don't
have a lot of tasks for for for long

200
00:11:42,400 --> 00:11:46,757
time horizons. So, it's small sample size.
Uh they don't have very good human

201
00:11:46,840 --> 00:11:50,637
baselines as well. Um so,
yeah, let's let's not let's not

202
00:11:50,720 --> 00:11:54,397
just take it at face value. A big
difference between uh their approach to

203
00:11:54,480 --> 00:11:58,157
benchmarking AI and our approach
is that we are really targeting general

204
00:11:58,240 --> 00:12:01,997
intelligence. And I feel like METER
is more of a domain-specific or even

205
00:12:02,080 --> 00:12:05,277
task-specific benchmark. Like,
for instance, uh they're benchmarking the

206
00:12:05,360 --> 00:12:08,637
hardness. I mean, if you start learning
hardness, it means that you're expecting

207
00:12:08,720 --> 00:12:12,317
a certain kind of task. Uh your system
is already task-specific like it's

208
00:12:12,400 --> 00:12:16,677
domain-specific. And they're heavily
indexing on uh software engineering. And

209
00:12:16,760 --> 00:12:18,757
meanwhile, we're really trying to uh

210
00:12:18,840 --> 00:12:20,677
benchmark the the the G like the the

211
00:12:20,760 --> 00:12:24,317
team in in AGI. Um so, in principle, you

212
00:12:24,400 --> 00:12:28,437
know, if you have AGI, uh it should
be able to look at any benchmark and it

213
00:12:28,520 --> 00:12:32,437
should not need any harness. Should not
need, you know, repurposing uh with a

214
00:12:32,520 --> 00:12:33,997
human in the loop for the new task.

215
00:12:34,080 --> 00:12:37,637
Should just, you know,
go and do it the way humans do.

216
00:12:37,720 --> 00:12:42,637
I I think it's a interesting thing
to measure for sure, but I think it's

217
00:12:42,720 --> 00:12:44,557
inadequate in a few different ways that

218
00:12:44,640 --> 00:12:48,357
are very important to understand. One is

219
00:12:48,440 --> 00:12:51,157
that as these tasks get to five-year or

220
00:12:51,240 --> 00:12:55,517
50-year tasks, it's like very hard
to evaluate and we don't want to wait five

221
00:12:55,600 --> 00:12:57,677
years and

222
00:12:57,760 --> 00:12:59,317
you know, that's going to be

223
00:12:59,400 --> 00:13:03,357
weird. And then the other thing is

224
00:13:03,440 --> 00:13:08,397
there are things that I could work
on for years and not solve and that a

225
00:13:08,480 --> 00:13:12,637
genius mathematician could work
on for a few minutes and solve easily.

226
00:13:12,720 --> 00:13:17,517
There are things that no one on Earth can
do or haven't been able to do yet without

227
00:13:17,600 --> 00:13:20,197
the help of tools, no matter how much
time they have, that we'd really like to

228
00:13:20,280 --> 00:13:23,837
get done. There's something that

229
00:13:23,920 --> 00:13:28,477
matters about um like,
you know, can we discover new

230
00:13:28,560 --> 00:13:31,477
science that humans have not been able
to discover and whether that takes a

231
00:13:31,560 --> 00:13:34,477
huge amount of compute and a small
amount of time or a small amount of

232
00:13:34,560 --> 00:13:36,077
compute and a lot of time,

233
00:13:36,160 --> 00:13:38,757
like we want the new knowledge. And very

234
00:13:38,840 --> 00:13:41,957
soon, I think we'll be talking about

235
00:13:42,040 --> 00:13:46,717
things like that. Can we Can we do
things of real value to people? New

236
00:13:46,800 --> 00:13:51,677
new scientific knowledge,
new economic acceleration. And

237
00:13:51,760 --> 00:13:57,037
you know, time on task is an interesting
but very perfect proxy for that.

238
00:13:57,120 --> 00:14:00,677
Thanks. Um guys, it would be remiss if
I didn't ask you about, you know, both of

239
00:14:00,760 --> 00:14:03,077
you are a part of companies that have

240
00:14:03,160 --> 00:14:08,797
sort of different approaches
to achieving this goal. So, namely in up

241
00:14:08,880 --> 00:14:11,597
the OpenAI world, at least what
is publicly known is, you know, you

242
00:14:11,680 --> 00:14:15,997
pre-train a big model, you do some
post-train based on RL data, and you get

243
00:14:16,080 --> 00:14:20,237
the model to be really good. And it's
we had reasoning happen along the way. It

244
00:14:20,320 --> 00:14:24,117
sounds like we got really, really good
just doing that. And that paradigm has

245
00:14:24,200 --> 00:14:27,997
has worked so far. Obviously, and uh
you have a slightly different approach,

246
00:14:28,080 --> 00:14:33,037
Francois. Um you want to do deep
learning based on program synthesis. So,

247
00:14:33,120 --> 00:14:35,797
maybe the way I'd want to frame this
question is, Francois, can you talk a

248
00:14:35,880 --> 00:14:39,837
little bit more about your approach?
I'm sure more of the audience knows less

249
00:14:39,920 --> 00:14:43,877
about the approach and uh I'd like
to compare and contrast that to sort of

250
00:14:43,960 --> 00:14:46,677
Sam, what you're seeing at OpenAI.

251
00:14:46,760 --> 00:14:50,237
Sure. So, you know, when I mentioned
program synthesis, lots of people ask

252
00:14:50,320 --> 00:14:53,477
me, "Is are you doing like coding agents?
Are you doing code gen?" And

253
00:14:53,560 --> 00:14:57,397
that's not at all what we're doing. Uh
we're not like adding a new layer on top

254
00:14:57,480 --> 00:15:00,557
of the the other end stack.
We're actually trying to rethink the

255
00:15:00,640 --> 00:15:04,597
foundations. So, we call
our approach symbolic learning.

256
00:15:04,680 --> 00:15:09,357
And it's basically uh a new branch
of machine learning. Like, if you go back

257
00:15:09,440 --> 00:15:12,917
to uh the the machine
learning problem in its

258
00:15:13,000 --> 00:15:16,597
simplest form, you have some input data,
you have some target data, you're

259
00:15:16,680 --> 00:15:20,197
looking for a function that will map the
inputs to targets and hopefully it will

260
00:15:20,280 --> 00:15:22,997
generalize to uh new inputs as well.

261
00:15:23,080 --> 00:15:26,277
Um and if you're doing parametric
learning, uh which includes deep

262
00:15:26,360 --> 00:15:30,637
learning, then your function is a
parametric curve and you're trying to

263
00:15:30,720 --> 00:15:33,957
find the right parameters
via uh gradient descent.

264
00:15:34,040 --> 00:15:38,117
And uh all of this all this picture
is basically what we're doing. What we're

265
00:15:38,200 --> 00:15:40,957
replacing the curve with a symbolic

266
00:15:41,040 --> 00:15:45,277
program that is minimal. It's like
the shortest program that can accurately

267
00:15:45,360 --> 00:15:50,037
explain the data. And when you're
doing this, you're much closer to uh

268
00:15:50,120 --> 00:15:54,477
producing a form of optimal machine
learning like optimal learning. Uh

269
00:15:54,560 --> 00:15:58,477
and uh there are a few consequences when
you when you get closer to optimality.

270
00:15:58,560 --> 00:16:03,277
Uh if your model is genuinely uh
the simplest possible to to to explain the

271
00:16:03,360 --> 00:16:06,117
data, then it's also going to be
the model that will be capable of

272
00:16:06,200 --> 00:16:10,277
generalizing the most strongly. Uh it's
going to be the most compute-efficient

273
00:16:10,360 --> 00:16:15,117
at this time and it's going to need uh
the least uh data points uh to find that

274
00:16:15,200 --> 00:16:17,757
model because, you know, if if the model
is too biased, you don't need you don't

275
00:16:17,840 --> 00:16:20,957
need uh trillions of tokens to find it.

276
00:16:21,040 --> 00:16:24,637
Um so, that's basically the story. We're
trying to rebuild machine learning from

277
00:16:24,720 --> 00:16:29,277
the ground up on top of uh a symbolic
substrate that would bring us much, much

278
00:16:29,360 --> 00:16:32,837
closer uh to optimal AI.
And eventually, you

279
00:16:32,920 --> 00:16:35,517
know, we're going to rebuild the entire
stack. Like, I expect in the future

280
00:16:35,600 --> 00:16:39,437
we'll have uh very
large-scale foundation models

281
00:16:39,520 --> 00:16:41,517
that will be symbolic in nature. We want

282
00:16:41,600 --> 00:16:46,477
our approach to be as scalable
as compute-efficient as deep learning.

283
00:16:46,560 --> 00:16:48,717
Sam, what do you think about
that approach? And, you know, OpenAI

284
00:16:48,800 --> 00:16:53,277
obviously takes a very different approach.
Is that going to continue to be the case?

285
00:16:53,360 --> 00:16:54,917
How do you think of different ways to

286
00:16:55,000 --> 00:16:57,677
get to this amorphous somewhat amorphous

287
00:16:57,760 --> 00:17:00,717
concept of of AGI? What do you think?

288
00:17:00,800 --> 00:17:04,477
I I think that is a great thing to try.

289
00:17:04,560 --> 00:17:08,077
I think that people
should try many approaches.

290
00:17:08,160 --> 00:17:11,397
We we try to bet with conviction on ours,

291
00:17:11,480 --> 00:17:16,677
but we could very easily hit some sort
of unexpected roadblock and I think

292
00:17:16,760 --> 00:17:19,997
you know, when we started with ours,
it was very unpopular and turned out to

293
00:17:20,080 --> 00:17:24,797
have some value and uh I'm sure other
people will start out with things that

294
00:17:24,880 --> 00:17:29,397
Well, this one isn't even very unpopular.
Uh this one is a great thing to try. Uh

295
00:17:29,480 --> 00:17:31,637
our approach is to push on the paradigm

296
00:17:31,720 --> 00:17:34,917
we understand and try to build towards

297
00:17:35,000 --> 00:17:39,197
uh an aligned AI researcher that can
discover whatever remaining stuff is to

298
00:17:39,280 --> 00:17:43,117
discover. Um and

299
00:17:43,200 --> 00:17:46,677
you know, we like seeing promising
signs, but still a ways to go.

300
00:17:46,760 --> 00:17:50,757
There's a specific question I got
from Twitter for about your timeline

301
00:17:50,840 --> 00:17:52,957
specifically for AI research, Sam. You

302
00:17:53,040 --> 00:17:55,597
at some point said automated AI research

303
00:17:55,680 --> 00:17:58,597
interns would be done by September 2026

304
00:17:58,680 --> 00:18:02,717
and fully automated AI researchers
by March 2028. Do you have an update on

305
00:18:02,800 --> 00:18:05,557
those timelines or is that still >>
say would be done. I said it was our

306
00:18:05,640 --> 00:18:08,877
goal. Uh but I feel like we are on

307
00:18:08,960 --> 00:18:10,837
trajectory for those goals.

308
00:18:10,920 --> 00:18:13,357
Makes sense. So, good. I also want to

309
00:18:13,440 --> 00:18:16,077
ask a question that's come up to me a

310
00:18:16,160 --> 00:18:18,317
lot, which is

311
00:18:18,400 --> 00:18:21,197
so many people in the research world

312
00:18:21,280 --> 00:18:23,517
specifically feel like they've just been

313
00:18:23,600 --> 00:18:25,997
it's it's been dominated by AI. And

314
00:18:26,080 --> 00:18:30,797
they're really looking to people like
yourself to say, "What are the kind of

315
00:18:30,880 --> 00:18:34,597
research directions that are less
explored that you'd you wish people

316
00:18:34,680 --> 00:18:37,237
would do more of?"
Almost like call for research. Do you

317
00:18:37,320 --> 00:18:40,517
have anything top of mind that you think,
"Yeah, I wish more people tried

318
00:18:40,600 --> 00:18:44,077
these approaches or tried more
experiments in this direction?"

319
00:18:44,160 --> 00:18:45,237
Francois, do you have any thoughts?

320
00:18:45,320 --> 00:18:48,677
>> Like, outside of AI? No,
uh inside of AI different than the

321
00:18:48,760 --> 00:18:53,277
current models or the current approaches
that we're taking in in some of the labs.

322
00:18:53,360 --> 00:18:56,117
Or it could be outside AI if you have
something interesting. I think there are

323
00:18:56,200 --> 00:18:59,637
so many directions uh that
people have not tried or not

324
00:18:59,720 --> 00:19:04,037
tried seriously yet. And in in a sense,
you know, I believe compute is a is a

325
00:19:04,120 --> 00:19:07,397
great equalizer. Like, if you throw
a ton of compute at one approach, you will

326
00:19:07,480 --> 00:19:09,077
see some really cool stuff happen.

327
00:19:09,160 --> 00:19:12,677
Famous rule of uh evolutionary
algorithms in Pascal.

328
00:19:12,760 --> 00:19:14,317
Uh I think it's been very underexplored.

329
00:19:14,400 --> 00:19:17,277
We have not tried to scale them, but we
could. I think we would see some really

330
00:19:17,360 --> 00:19:21,517
cool stuff. Uh that's true of uh program
synthesis and that's precisely what we

331
00:19:21,600 --> 00:19:23,237
are trying to scale right now.

332
00:19:23,320 --> 00:19:26,797
Um so, yeah, these these two I think are
pretty good candidates, but in general,

333
00:19:26,880 --> 00:19:30,557
they are just they're just
this infinite supply of uh

334
00:19:30,640 --> 00:19:35,597
novel and cool ideas that could work
maybe. And in fact, probably would if

335
00:19:35,680 --> 00:19:38,317
not work, at least you would see some
really interesting results if you were

336
00:19:38,400 --> 00:19:41,957
to to pursue them seriously.
Sam, what do you think?

337
00:19:42,040 --> 00:19:47,077
There was a time after right after we
had put out GPT-4, where I was having

338
00:19:47,160 --> 00:19:50,117
dinner with one of the great OpenAI
researchers and he said, you You the

339
00:19:50,200 --> 00:19:53,637
thing that is most sad to me right now,
there're all these brilliant people in

340
00:19:53,720 --> 00:19:56,917
academia, there're all
of these great ideas to try.

341
00:19:57,000 --> 00:20:01,757
Everybody just wants to reproduce GPT-4,
which at the time was, you know, still a

342
00:20:01,840 --> 00:20:05,157
difficult thing to do. And he's like,
there there's this incredibly wide

343
00:20:05,240 --> 00:20:08,437
space. How do we convince every academic

344
00:20:08,520 --> 00:20:12,397
department around the world to stop
saying like, you know, we just need to

345
00:20:12,480 --> 00:20:17,317
we just need to train a big model
in the in the same style. And

346
00:20:17,400 --> 00:20:21,237
we never found an answer to that, but I
think finally people are saying again,

347
00:20:21,320 --> 00:20:25,317
let's try some of these
very different ideas.

348
00:20:25,400 --> 00:20:28,997
Uh, I'm not sure what's changed there,
but it does feel exciting that that's

349
00:20:29,080 --> 00:20:34,317
happening. However, I will give
an answer outside of AI, which is

350
00:20:34,400 --> 00:20:37,037
for the first time in the last few
months, I think you can use the latest

351
00:20:37,120 --> 00:20:40,157
frontier models hugely

352
00:20:40,240 --> 00:20:42,637
accelerate some other kinds of research.

353
00:20:42,720 --> 00:20:46,637
And, you know, like AI research has no
lack of smart people working on it and

354
00:20:46,720 --> 00:20:51,277
no lack of funding and I think
it's great. But the stack that

355
00:20:51,360 --> 00:20:54,757
other directions of research, you know,
physics, mathematics, bio, whatever

356
00:20:54,840 --> 00:20:59,277
you'd like, that have been difficult
to make fast progress and all of a sudden

357
00:20:59,360 --> 00:21:02,997
feel like they might get burst
open and all these factors,

358
00:21:03,080 --> 00:21:06,677
um, that feels very underexplored
relative to the potential.

359
00:21:06,760 --> 00:21:09,557
Makes sense. I have one last question
for you, Sam, and then we can do the

360
00:21:09,640 --> 00:21:13,717
rapid fire. I know you have a hard
stop at 7:30. Um, and let me ask you the

361
00:21:13,800 --> 00:21:17,557
OpenAI question, Sam. Um, OpenAI
is doing a lot of different things, right?

362
00:21:17,640 --> 00:21:22,637
So, you have different modalities,
you have Sora, you have Operator, Atlas.

363
00:21:22,720 --> 00:21:25,597
There's a bunch of different experiments
you have run and I think everyone's

364
00:21:25,680 --> 00:21:29,477
grateful that you're you're running all
these experiments. Um, but very recently

365
00:21:29,560 --> 00:21:34,277
you said you'd shut down Sora. Um, there
was a leak about this new pre-train that

366
00:21:34,360 --> 00:21:38,037
you have and you want to focus all
of your effort into sort of coding and and

367
00:21:38,120 --> 00:21:43,117
the enterprise. Um, can you sort of add
some throw some color on that and uh

368
00:21:43,200 --> 00:21:47,557
talk a little bit about what
are OpenAI's top priorities?

369
00:21:47,640 --> 00:21:50,837
Um, with the new set of models

370
00:21:50,920 --> 00:21:55,197
that I think will come
out, the impact that

371
00:21:55,280 --> 00:21:58,997
the technology will have on accelerating
science and accelerating the economy.

372
00:21:59,080 --> 00:22:02,197
Uh, I mean, we may be wrong, we may be

373
00:22:02,280 --> 00:22:03,877
over-optimistic, but it feels like it's

374
00:22:03,960 --> 00:22:09,357
about to be absolutely massive. And we

375
00:22:09,440 --> 00:22:13,197
think that will be the most important
area for us to dedicate our compute, our

376
00:22:13,280 --> 00:22:15,717
resources. Um,

377
00:22:15,800 --> 00:22:19,077
I think things are going to get pretty

378
00:22:19,160 --> 00:22:22,837
heated. Uh, and we

379
00:22:22,920 --> 00:22:26,557
I mean, I think there's many
great things but to do, but

380
00:22:26,640 --> 00:22:30,157
several times in OpenAI's
history we've said, you know,

381
00:22:30,240 --> 00:22:33,197
there are a lot of interesting projects,
and yet there's this one new thing

382
00:22:33,280 --> 00:22:38,397
that's really working. Uh,
as happened in the original GPT days, I

383
00:22:38,480 --> 00:22:41,957
want to shut down a bunch of other
projects to concentrate our compute. And

384
00:22:42,040 --> 00:22:45,557
it feels like it's really time to
concentrate our compute. I mean, to give

385
00:22:45,640 --> 00:22:49,437
like an obvious example,
if we will be able to put

386
00:22:49,520 --> 00:22:53,477
a million GPUs towards curing all
cancers or a million GPUs towards making

387
00:22:53,560 --> 00:22:55,517
videos, I love the videos, I pick the

388
00:22:55,600 --> 00:22:57,997
cancer research. Um,

389
00:22:58,080 --> 00:23:02,957
and as we look at the next couple
of years, I think it is a time, like we've

390
00:23:03,040 --> 00:23:06,477
had a couple of times in our history, to
say, "Okay, we see a really big thing on

391
00:23:06,560 --> 00:23:10,437
the horizon. Like, let's point
all of our firepower at it."

392
00:23:10,520 --> 00:23:14,797
That makes total sense. All right, let's
do the rapid fire. Let's start with AGI.

393
00:23:14,880 --> 00:23:17,237
This is our AGI launch after all.

394
00:23:17,320 --> 00:23:19,997
I want you to answer
on a scale of 1 to 10,

395
00:23:20,080 --> 00:23:23,997
how far are we to your definition of

396
00:23:24,080 --> 00:23:27,797
AGI? One being nowhere close, 10 being

397
00:23:27,880 --> 00:23:30,757
probably going to happen this year.

398
00:23:30,840 --> 00:23:32,357
Friends, why don't we start with you?

399
00:23:32,440 --> 00:23:34,077
Are we talking a linear scale?

400
00:23:34,160 --> 00:23:36,557
>> [laughter] >> And it's linear.

401
00:23:36,640 --> 00:23:40,317
Let's let's I'll go with with five.
I I I think so.

402
00:23:40,400 --> 00:23:43,917
Obviously, we have models now that have
like genuine fluid intelligence.

403
00:23:44,000 --> 00:23:47,517
Uh, they are intelligent uh in hence
they're not human level and it's

404
00:23:47,600 --> 00:23:51,797
actually still we're still in this regime
where I'd say easy to find things

405
00:23:51,880 --> 00:23:55,717
uh that are, you know, easy far far
from normal people or from average people.

406
00:23:55,800 --> 00:23:58,997
Uh, you don't have to you don't have
to look for super geniuses to beat the

407
00:23:59,080 --> 00:24:02,597
models. Uh, and uh and you
know, things that are

408
00:24:02,680 --> 00:24:05,637
basically impossible for current AI.
So, we're still in this regime where it's

409
00:24:05,720 --> 00:24:09,997
not hard to find these tasks that humans
can do, AI cannot. I think as we move

410
00:24:10,080 --> 00:24:12,837
closer to AGI, it's going
to be harder and harder.

411
00:24:12,920 --> 00:24:15,757
And at some point, you know, uh it's
it's going to have to be like, you know,

412
00:24:15,840 --> 00:24:19,917
PhD level uh tasks only and and and it's
and it's going to be very very short

413
00:24:20,000 --> 00:24:22,277
list of times it's going to be very
difficult to come up with them. And

414
00:24:22,360 --> 00:24:26,717
that's the key when we have AGI. Makes
sense. Sam, what do you think? It feels

415
00:24:26,800 --> 00:24:29,877
very close to me. I think the only thing
I'd say is really missing is some sort

416
00:24:29,960 --> 00:24:32,917
of continuous learning or long-term
memory. Uh, but other than that, it

417
00:24:33,000 --> 00:24:35,197
feels very AGI-like to me already.

418
00:24:35,280 --> 00:24:38,317
Awesome. What is one
product or experience that

419
00:24:38,400 --> 00:24:40,277
you've tried, other than your own,

420
00:24:40,360 --> 00:24:46,800
that has sort of blown your mind,
AI or otherwise, in the last month or so?

421
00:24:47,400 --> 00:24:49,797
So, it's probably um as I commented to

422
00:24:49,880 --> 00:24:52,837
me, it's a coding like Codex and Code

423
00:24:52,920 --> 00:24:56,957
Code. It's really really impressive
and really useful as well. That's awesome.

424
00:24:57,040 --> 00:25:01,317
Sam, what do you think? >> These
magnetic tiles for small children.

425
00:25:01,400 --> 00:25:03,677
Every face connects to every other face.

426
00:25:03,760 --> 00:25:07,477
Okay. I had to like cut one open to see
how it worked. Um, but kids really love

427
00:25:07,560 --> 00:25:11,397
them, they're awesome.
What do you know what they're called?

428
00:25:11,480 --> 00:25:15,117
Magna- The tiles, they're little cubes,
like a full cubes so that every side can

429
00:25:15,200 --> 00:25:17,567
connect. >> going to say
they're sold out after this.

430
00:25:17,650 --> 00:25:18,597
>> [laughter]

431
00:25:18,680 --> 00:25:20,957
>> That's awesome. Um, okay.

432
00:25:21,040 --> 00:25:24,677
What is one industry or asset class that

433
00:25:24,760 --> 00:25:29,397
you are super bullish on and super
bearish on in the next 10 years?

434
00:25:29,480 --> 00:25:33,837
I'm very bullish on hardware as especially
uh robots and drones. I think

435
00:25:33,920 --> 00:25:37,277
that's going to be more and more
important, you know, as AI capabilities

436
00:25:37,360 --> 00:25:40,557
advance, we're going to want to put more
and more of these capabilities into the

437
00:25:40,640 --> 00:25:43,477
real world and we're going to need
a support for that. We're going to need

438
00:25:43,560 --> 00:25:46,677
robots, we're going to need drones.
So, I'm very bullish on drones.

439
00:25:46,760 --> 00:25:48,757
Uh, very bullish on hardware in general.

440
00:25:48,840 --> 00:25:53,157
Uh, as uh as software gets easier
and easier with like HNC coding and and

441
00:25:53,240 --> 00:25:56,877
eventually AGI, the the importance to
compare to the importance of hardware is

442
00:25:56,960 --> 00:25:59,797
only going to rise.
What do you say, Sam? I could give like

443
00:25:59,880 --> 00:26:03,477
a 3-hour answer to this question
with and I had so much fun.

444
00:26:03,560 --> 00:26:07,157
I will just limit myself to a single
observation though. The there was this

445
00:26:07,240 --> 00:26:10,397
quote that was bouncing around Twitter
for a while, which is all of the things

446
00:26:10,480 --> 00:26:13,797
we've learned about society, about
managing scarcity, are not really going

447
00:26:13,880 --> 00:26:17,517
to apply to managing abundance. And it
seemed really smart and really profound

448
00:26:17,600 --> 00:26:20,037
and then I was like after a day I
was like, "Yeah, that's kind of like not

449
00:26:20,120 --> 00:26:23,797
that interesting." And then it kept like
rattling around my head and I was like,

450
00:26:23,880 --> 00:26:27,397
"You know what?
There is one thing about this that is

451
00:26:27,480 --> 00:26:31,877
blowing my mind, which is countries
like us have a great deal of

452
00:26:31,960 --> 00:26:35,557
debt. And the plan
on this is we are going to

453
00:26:35,640 --> 00:26:37,797
uh inflate it away as far as I can tell.

454
00:26:37,880 --> 00:26:41,197
You know, we have 130% someone will
know the number, something like that,

455
00:26:41,280 --> 00:26:45,037
debt-to-GDP ratio, but, you know, we're
like a powerful country and inflation's

456
00:26:45,120 --> 00:26:47,917
going to keep going
and it's not that scary.

457
00:26:48,000 --> 00:26:52,717
If we're in a forever deflationary world,
I don't know how you're supposed to

458
00:26:52,800 --> 00:26:55,837
think about that as an asset
class and what's going to

459
00:26:55,920 --> 00:27:00,357
happen. That's a great answer.
Um, I want to do

460
00:27:00,440 --> 00:27:05,077
favorite book or a book that's
impacted you most in life.

461
00:27:05,160 --> 00:27:07,197
Feel free to go anyone. Let's go first.

462
00:27:07,280 --> 00:27:10,277
>> to go with the Asimov.
Okay. Uh, robots,

463
00:27:10,360 --> 00:27:12,357
the foundation the series.

464
00:27:12,440 --> 00:27:17,357
Um, I've wanted to build uh human level
AI since I was like 16 and I think one

465
00:27:17,440 --> 00:27:20,437
of the the big influences
on me at the time was the

466
00:27:20,520 --> 00:27:25,237
Asimov's uh robot stories.
What about you, Sam? Um, that's a great

467
00:27:25,320 --> 00:27:29,317
choice. I would going to say The Beginning
of Infinity, but uh the Asimov

468
00:27:29,400 --> 00:27:34,157
stories maybe did have more of a impact
on me when I was younger. I in general,

469
00:27:34,240 --> 00:27:37,837
I think about this now with kids,
like books that you read as a kid where

470
00:27:37,920 --> 00:27:41,077
you're like, "I can go do anything,
I can be a high-agency person, I can dream

471
00:27:41,160 --> 00:27:45,317
about the future." Those are really great.
Was that the book for you?

472
00:27:45,400 --> 00:27:48,637
I mean, it wasn't one. Like, I was like
all of the young adult fiction that had

473
00:27:48,720 --> 00:27:52,517
some vibe of that I loved. Um, last one.

474
00:27:52,600 --> 00:27:54,717
What is a person

475
00:27:54,800 --> 00:27:58,437
alive today that you respect the most?

476
00:27:58,520 --> 00:28:02,720
Could be family, could be
a famous person, could be

477
00:28:05,640 --> 00:28:07,677
So many great names, you know.

478
00:28:07,760 --> 00:28:11,717
Um, Okay, let me go
with a really random answer.

479
00:28:11,800 --> 00:28:14,757
Uh, I want to I want
to highlight like a researcher

480
00:28:14,840 --> 00:28:18,877
whose work has influenced me and is not
well known at all. Uh, his name is uh

481
00:28:18,960 --> 00:28:22,037
Pierre Boudier. Uh,
he's a he's a French researcher so

482
00:28:22,120 --> 00:28:25,277
we are researchers. He's been doing
AI for for a very very long time.

483
00:28:25,360 --> 00:28:28,517
And he was one of the pioneers on uh AI

484
00:28:28,600 --> 00:28:32,437
curiosity, like artificial curiosity,
which is really a fascinating concept.

485
00:28:32,520 --> 00:28:35,997
So, he's done a lot of great work. Uh,
he's very successful but also not not so

486
00:28:36,080 --> 00:28:39,917
well known. So, yeah, I respect him a lot,
but you know, that's just I I could

487
00:28:40,000 --> 00:28:42,877
go with like so so many different people.
Yeah, that's awesome. I like

488
00:28:42,960 --> 00:28:46,997
that. Also, Sam, you can answer well
known or not well known, up to you. Um,

489
00:28:47,080 --> 00:28:52,157
my mom. Love that. Love that. Well,
with that, I think we should wrap.

490
00:28:52,240 --> 00:28:56,337
Thank you guys so much for being
here and thank you all for attending.

491
00:28:56,420 --> 00:28:58,440
>> [applause]