1
00:00:00,000 --> 00:00:01,757
foreshadowed a couple times already that

2
00:00:01,840 --> 00:00:04,597
you have some good models on the way.

3
00:00:04,680 --> 00:00:08,317
What is Spud? Uh the information said uh

4
00:00:08,400 --> 00:00:11,677
that you finished uh pre-training Spud,

5
00:00:11,760 --> 00:00:16,757
and Sam Altman, the CEO of OpenAI,
has told the staff that uh they should

6
00:00:16,840 --> 00:00:20,637
expect to have a very strong model
in a few weeks. This was a few weeks ago.

7
00:00:20,720 --> 00:00:24,117
And the team believes it can really
accelerate the economy, and things are

8
00:00:24,200 --> 00:00:26,757
moving faster than many of us expect

9
00:00:26,840 --> 00:00:29,677
expected. So, what's Spud?

10
00:00:29,760 --> 00:00:33,957
It's a good model. Um but I think
that it's really not about any one model.

11
00:00:34,040 --> 00:00:37,957
Okay, right? The the the way that our
development process works is you have

12
00:00:38,040 --> 00:00:42,077
pre-training, so you produce a new base
model that then is the foundation that

13
00:00:42,160 --> 00:00:45,757
we build further improvements
on top of. And that

14
00:00:45,840 --> 00:00:49,477
is always a huge effort across many
people in the company, and that's where

15
00:00:49,560 --> 00:00:52,797
I've actually been spending most of my
efforts over the past 18 months has been

16
00:00:52,880 --> 00:00:56,837
really focused on our GPU
infrastructure, on supporting the

17
00:00:56,920 --> 00:00:59,997
teams that do all of the training
frameworks to scale up at these big

18
00:01:00,080 --> 00:01:04,077
runs. But then there's a reinforcement
learning process. So, you take this AI

19
00:01:04,160 --> 00:01:08,077
that has learned lots of things
about the world, and it applies that

20
00:01:08,160 --> 00:01:10,397
knowledge. And then we do
a post-training process where you really

21
00:01:10,480 --> 00:01:13,717
say, "Okay, now you know how to solve
problems. You practice it in in all

22
00:01:13,800 --> 00:01:16,637
these different contexts." And then
here's kind of the last mile of of

23
00:01:16,720 --> 00:01:18,557
behavior and usability. So, I think of

24
00:01:18,640 --> 00:01:22,037
Spud as a new base, as a new pre-train,

25
00:01:22,120 --> 00:01:25,917
and that we have had this uh

26
00:01:26,000 --> 00:01:30,117
I'd say it's like we have maybe 2 years
worth of research that is coming to

27
00:01:30,200 --> 00:01:33,717
fruition in this model.
It's going to be very exciting.

28
00:01:33,800 --> 00:01:37,357
And I think that the way that the world
will experience it is just improved

29
00:01:37,440 --> 00:01:41,397
capabilities, and that for me
it's never about any one release,

30
00:01:41,480 --> 00:01:45,797
because as soon as we have this one
release, it'll be an early version of

31
00:01:45,880 --> 00:01:49,117
what we have coming. We'll do much
more of each of these steps of the

32
00:01:49,200 --> 00:01:53,997
improvement process. And so, I think
that where we're going is is almost just

33
00:01:54,080 --> 00:01:58,237
we have this engine of progress that
just moves faster and faster, and that

34
00:01:58,320 --> 00:02:01,237
Spud is just one step along the way.
So, what do you think it'll be able to do

35
00:02:01,320 --> 00:02:04,357
that today's models can't? I So, I think

36
00:02:04,440 --> 00:02:06,237
it's going to be able to solve both much

37
00:02:06,320 --> 00:02:11,437
harder problems. I think it will be
much more nuanced. It'll understand

38
00:02:11,520 --> 00:02:15,077
instructions better. It'll understand
the context much better. That there's

39
00:02:15,160 --> 00:02:18,317
this thing called big model smell that
people talk about, where it's just like

40
00:02:18,400 --> 00:02:19,757
there's something about like when these

41
00:02:19,840 --> 00:02:22,157
models are just actually just much

42
00:02:22,240 --> 00:02:25,837
smarter, much more capable, that they

43
00:02:25,920 --> 00:02:28,357
bend to you much more, and you feel it.

44
00:02:28,440 --> 00:02:31,637
Right? When you ask a question and the AI
doesn't quite get it, it's always so

45
00:02:31,720 --> 00:02:34,437
disappointing. Right? When you have
to like explain it, you're like, "You

46
00:02:34,520 --> 00:02:37,517
really should be able to figure this out."
And so, I would just think of it

47
00:02:37,600 --> 00:02:42,517
as in some ways just
qualitatively there will be

48
00:02:42,600 --> 00:02:45,757
like quantitatively lots of shifts,
right? And qualitatively there will just

49
00:02:45,840 --> 00:02:48,757
be new things where you would be
frustrated before you never use an AI

50
00:02:48,840 --> 00:02:51,917
for it, and now you just use it without
without thinking very much. And I think

51
00:02:52,000 --> 00:02:55,557
that that that is what we're going
to see across the board. I'm super excited

52
00:02:55,640 --> 00:02:57,397
to see how it raises the ceiling, right?

53
00:02:57,480 --> 00:03:00,077
We've already seen these physics
applications, things like that. And I

54
00:03:00,160 --> 00:03:04,357
think we will be able to just solve like
way more open-ended problems, way longer

55
00:03:04,440 --> 00:03:07,997
time horizons. And then also
very excited to see how it raises

56
00:03:08,080 --> 00:03:11,557
the floor, where just for anything you
want to do, it's just so much useful

57
00:03:11,640 --> 00:03:14,880
more useful for you.