00:00:00
Great to see everyone and I'm thrilled
00:00:02
to be able to talk about the issue of
00:00:04
the day and that is artificial
00:00:05
intelligence and AI in our world. Um
00:00:09
David, Michael, I'd love you to talk
00:00:12
about what where we are right now in
00:00:15
terms of the pursuit to be the number
00:00:17
one uh lead AI country. How are we
00:00:20
doing, David?
00:00:21
>> I think we're doing great. Um Maria,
00:00:24
last year uh President Trump gave a
00:00:26
major AI policy speech. is in July and
00:00:28
he declared that the United States had
00:00:30
to win the AI race. Uh he he had first
00:00:33
of all declared that we were in one. Uh
00:00:36
and I think his speech was reminiscent
00:00:38
of when President Kenny declared that we
00:00:40
were in a space race and had to win that
00:00:41
race. I think since then what you've
00:00:44
seen is that American companies have
00:00:46
only innovated more. You're seeing all
00:00:48
sorts of really incredible products
00:00:50
being released all the time. I think
00:00:52
that um American uh uh AI models, chips,
00:00:57
um data centers only just keep um
00:01:00
getting better and better. And so I feel
00:01:01
very good about the American position in
00:01:04
this AI race. Certainly we have some
00:01:05
very uh you know competent uh and
00:01:09
formidable competitors. Um China
00:01:11
obviously has a lot of very smart people
00:01:12
working in this area. But I do think
00:01:15
that uh just what you see from uh
00:01:18
American companies in Silicon Valley
00:01:19
right now is really incredible. And yet
00:01:21
there are still so many questions about
00:01:23
all of the spending underway uh to build
00:01:25
this out with regard to data centers.
00:01:27
And of course the question keeps coming
00:01:29
up are we spending too much. Will we get
00:01:31
the return on investment? How do you see
00:01:33
that?
00:01:34
>> I I think that we will. Um I think that
00:01:37
the reason why you're seeing this huge
00:01:39
infrastructure buildout is because the
00:01:41
demand is ultimately there. I I know a
00:01:43
lot of people worry and about whether
00:01:46
this could be like a dot situation. And
00:01:48
remember where we had the whole fiber
00:01:50
build out in the late '9s and then we
00:01:51
had a doc crash. The difference here is
00:01:54
that uh in the late '9s and early 2000s
00:01:57
we had a a problem known as dark fiber
00:01:59
where you had this fiber buildout and
00:02:01
then it didn't get used. There's no such
00:02:03
thing as a dark GPU right now. Every GPU
00:02:05
that's being put in a data center is
00:02:07
getting used. Uh and it's being used to
00:02:09
generate tokens and that's to power the
00:02:11
this new generation of AI chat bots or
00:02:14
coding assistants. uh and there's just
00:02:16
been some releases in the last couple
00:02:17
months on the coding front that you know
00:02:20
it's if you're following what develop
00:02:22
software developers are saying they're
00:02:24
saying it's mind-blowing it's completely
00:02:25
revolutionizing their industry so demand
00:02:28
for tokens just increases and that
00:02:30
increases the demand for this data
00:02:32
center buildout that we're seeing so I
00:02:34
don't think it's going to stop anytime
00:02:35
soon and just last year this
00:02:37
infrastructure buildout added about uh
00:02:39
2% to the GDP growth rate and I and
00:02:42
that's what helped propel us to this you
00:02:44
know four to 5% % growth rate and I
00:02:46
think you're going to see something
00:02:47
similar this year.
00:02:48
>> Well, it is certainly leading growth,
00:02:50
Michael. Um, and I'm so happy to be able
00:02:52
to get this conversation going with both
00:02:55
of you who are really leading this.
00:02:57
David, thank you. And Michael, thank
00:02:58
you. Same questions for you, Michael.
00:03:00
Assess where we are right now on AI.
00:03:02
>> I I think just a reminder for the group
00:03:03
for those who haven't been tracking as
00:03:05
closely as we do every day. The the plan
00:03:07
really had essentially three pillars and
00:03:09
it talked about how one, how can the US
00:03:12
continue to out innovate our
00:03:13
competitors? two, how can we drive the
00:03:16
infrastructure build that we need to
00:03:17
support um this this AI revolution? And
00:03:20
three, how do we actually share with the
00:03:22
world or export our great American
00:03:24
technology? And for each of those three
00:03:26
pillars, there was quite a lot of
00:03:27
actions that the federal government has
00:03:28
taken to drive that forward. Um, and I
00:03:30
think I think we're pretty proud to say
00:03:32
that we've made, I think, pretty good
00:03:33
progress on on all three. Um just
00:03:35
focusing a little bit on the innovation
00:03:36
one you were talking about earlier. I
00:03:38
think the the the the the core um
00:03:41
insight that we've always had about how
00:03:42
you drive this innovation is you have to
00:03:44
have a regulatory environment that
00:03:46
allows this technology to be developed
00:03:49
and ultimately commercialized in the
00:03:51
United States. And the US has done a
00:03:54
great job compared to the rest of the
00:03:55
world on sort of setting that up and
00:03:57
creating a framework that works. But we
00:03:59
can always do better and improve it. And
00:04:01
the president in his his speech in July
00:04:03
talked a lot about um this issue of a
00:04:06
patchwork of state regulations and how
00:04:08
can we ensure that there aren't 50
00:04:10
different rules around AI and and what's
00:04:13
most what's most important about this
00:04:14
debate which I I think a lot of people
00:04:16
sometimes don't sometimes miss is the
00:04:19
patchwork is actually most detrimental
00:04:21
to early stage young companies and
00:04:24
entrepreneurs. If you want to develop a
00:04:26
new AI technology, if you want to build
00:04:27
something on top of one of our great
00:04:28
frontier models, having to figure out
00:04:31
how to navigate 50 different rules
00:04:32
across 50 different states creates a lot
00:04:34
of friction and ultimately the big guys
00:04:36
are the ones that can succeed in in that
00:04:38
environment the best. Um, so we're
00:04:39
spending a lot of time trying to think
00:04:40
about how can you create a legislative
00:04:42
proposal that can actually um deliver on
00:04:45
um a sensible national framework to
00:04:47
solve to solve this regulatory issue.
00:04:49
So, so what would you say then, Michael,
00:04:51
are the basic frameworks that are uh
00:04:54
sort of must-have in in that kind of
00:04:56
federal oversight? Because some states
00:05:00
did push back in the US and say, "No,
00:05:02
no, we want to be able to control our
00:05:04
destiny when it comes to AI." What's
00:05:06
most important when you look at that
00:05:08
framework in terms of um a federal
00:05:11
oversight? Yeah, I think in the
00:05:13
executive order the president signed in
00:05:15
in December directing us to kind of work
00:05:17
through this proposal, he listed a few
00:05:19
things that um the state should continue
00:05:21
to be able to pursue individually on
00:05:23
their own. Um legislation or rules
00:05:25
around child safety was on that list. Um
00:05:27
the rules around permitting of data
00:05:29
centers and buildouts are continuing to
00:05:31
be something that states should should
00:05:32
look at. So there are a few things that
00:05:34
were enumerated, but that's the kind of
00:05:35
stuff that I guess Dave and I are going
00:05:36
to be working through. I don't know if
00:05:37
you have any thoughts on on that. Yeah,
00:05:39
I mean I I think the the basic problem
00:05:42
that we have is that I mean frankly the
00:05:44
states are going hog wild right now with
00:05:46
regulation. There's over,200 bills going
00:05:49
through state legislatores right now. I
00:05:51
think it's very much a knee-jerk
00:05:53
reaction. I know there's a lot of fears
00:05:55
and concerns about AI, but it seems like
00:05:57
for every hypothetical concern, there's
00:06:01
um multiple state bills now to try and
00:06:03
regulate that thing before we really
00:06:05
know how it's going to play out. And um
00:06:07
I think it would be better to I think
00:06:09
since this technology is so new and the
00:06:11
environment is so dynamic, I think it'd
00:06:13
be better to spend a little bit more
00:06:15
time studying how AI is actually being
00:06:17
used and what risks are actually
00:06:19
materializing before you overregulate
00:06:21
the thing. But in any event, that that's
00:06:23
the what we're seeing right now at the
00:06:25
state level. And um and and I think that
00:06:28
the president's been very consistent
00:06:30
that it would be better to have a single
00:06:32
have one rule book, a single rule book
00:06:34
at the federal level, lightweight
00:06:35
federal standard. Uh I think this
00:06:37
problem is only going to get more acute
00:06:40
over time because again you you as you
00:06:42
have 50 different states running in 50
00:06:44
different directions, the patchwork
00:06:46
problem only gets um more significant.
00:06:48
So, in any event, this is something that
00:06:49
we're going to work, I think, closely
00:06:51
together on this year, which is to see
00:06:53
if we can get enough consensus on a
00:06:55
federal framework to enact a law. Only
00:06:58
Congress can ultimately cramp the
00:07:00
states. We understand that. Um, and you
00:07:02
know, as you know, it's very difficult
00:07:03
to get a bill through Congress. You need
00:07:05
60 votes in the Senate. So, has to be
00:07:07
bipartisan to to a certain degree. So,
00:07:10
but we're going to try and see if we can
00:07:12
um work to get that consensus.
00:07:14
>> Yeah. And do you have any clarity on the
00:07:15
timing on that in terms of um support in
00:07:19
Congress for a federal oversight or do
00:07:21
you see push back there as well
00:07:23
depending on the state you're talking
00:07:24
about?
00:07:25
>> Well, there's push back in Congress to
00:07:27
the idea of preeemption without a
00:07:30
federal standard. So, in other words,
00:07:31
you can't replace something with
00:07:33
nothing. This is sort of the the thing
00:07:34
that we heard uh repeatedly. But I think
00:07:37
there is uh quite a bit of interest in
00:07:39
both the House and the Senate towards
00:07:41
having again some sort of lightweight
00:07:43
federal standard. But we're still in the
00:07:44
early stages of those conversations and
00:07:46
we're going to see what we can try and
00:07:47
get done this year.
00:07:48
>> Meanwhile, you've got some people
00:07:50
pushing back after wanting to see the
00:07:52
innovation and growth of data centers.
00:07:54
Now they're saying not in my backyard.
00:07:56
What about that? Is that an issue?
00:07:58
>> Yeah, I mean we got a letter recently
00:07:59
from Bernie Sanders saying stop all data
00:08:01
centers, all data center development.
00:08:03
And you know if we do that we will lose
00:08:06
the AI race. I mean you do need this
00:08:08
infrastructure. Uh other countries are
00:08:10
building out this infrastructure.
00:08:11
China's building out I think they're um
00:08:13
spinning up a a new uh nuclear power
00:08:16
plant or um or coal plant new energy
00:08:19
every single week and a lot of that is
00:08:21
going to power their data centers. So it
00:08:24
would fundamentally I think uh
00:08:26
the United States in the AI race if we
00:08:28
just stopped building data centers
00:08:29
altogether. At the same time, there are
00:08:31
concerns about affordability, about um
00:08:34
whether consumers would have to pay a
00:08:36
higher electrical rate because of data
00:08:38
centers. Uh President Trump's been
00:08:40
really clear that consumers should not
00:08:42
have to pay higher rates for electricity
00:08:44
because of data centers. You saw just
00:08:46
last week Microsoft stepped up and made
00:08:47
a pledge that it will that its data
00:08:50
centers will not cause residential rates
00:08:53
to increase. I think you'll likely see
00:08:55
other tech companies stepping up and
00:08:57
making similar commitments. And in fact,
00:08:59
when I've talked to the hyperscalers and
00:09:01
when I've talked to the AI companies, it
00:09:04
was never their plan to draw off the
00:09:05
grid. They all are uh saw standing up
00:09:07
their own power generation as part of
00:09:10
their buildout. Um and what Secretary
00:09:13
Wright, the our Secretary of Energy has
00:09:15
been doing is trying to um is is reform
00:09:18
the regulations that actually make it
00:09:20
more difficult for these AI data centers
00:09:22
to stand up their own power behind the
00:09:24
meter. So that basically is is our
00:09:26
vision is let and and I should say this
00:09:29
is President Trump's vision really since
00:09:31
the beginning of the administration is
00:09:32
he said let the AI companies become
00:09:35
power companies let them stand up their
00:09:37
own power generation as they built you
00:09:39
know side by side with these new data
00:09:40
centers and the um the result of that is
00:09:44
you know a we get this infrastructure b
00:09:46
residential rates don't go up. Yeah,
00:09:48
because Michael, this this race has fast
00:09:51
become it's moves from an AI race to a
00:09:54
power race.
00:09:56
>> And and I think what we're seeing is
00:09:57
that um we we need to share a good story
00:10:02
about how ultimately this buildout is
00:10:04
going to be net positive for American
00:10:06
rateayers. And I think sometimes if you
00:10:08
know and if you're in a small community
00:10:09
and someone shows up to build the data
00:10:11
center, I mean, you have to make it
00:10:12
clear that ultimately this something is
00:10:13
going to actually lower your your rates
00:10:15
long term. Um and and the president put
00:10:17
out a tweet a truth last Monday where he
00:10:19
was as as David said very clear that you
00:10:21
know if you're going to build a data
00:10:22
center you have to pay your own way for
00:10:24
it and um Microsoft has stepped up and
00:10:26
our our hope is that many others will do
00:10:27
the same.
00:10:28
>> But but some companies um because they
00:10:31
don't have the cash right now are
00:10:33
borrowing money right to to build out
00:10:35
the data centers and there's also a
00:10:37
worry that the banks will be left
00:10:39
holding the bag for some of this because
00:10:41
again the spending is is too much. your
00:10:44
thoughts on that?
00:10:46
>> Well, I think there there is obviously
00:10:48
that concern. I mean, you you know, I I
00:10:50
think it's it's um it's less I would say
00:10:53
the banks are more you see Oracle making
00:10:55
a huge investment. You see, uh
00:10:57
Blackstone making huge investments, real
00:10:58
estate companies. Um ultimately, I think
00:11:01
these are very savvy market players,
00:11:04
very deep companies, and they're doing
00:11:06
this because they see an ROI there at
00:11:08
the end of the the rainbow. Um can I
00:11:11
make one other point about just the the
00:11:12
data center? So, um, just on
00:11:14
electricity, uh, I actually think that
00:11:17
if we allow the data centers to stand up
00:11:19
their own power generation, it will
00:11:21
actually bring down rates. Not only will
00:11:23
it not increase residential rates, it'll
00:11:25
bring it down. And it'll do that in two
00:11:27
ways. One is that the data centers can
00:11:30
can give or sell power back to the meter
00:11:34
when they have excess. So, that will
00:11:35
help bring down rates. Second, there's a
00:11:38
lot of fixed costs involved in power
00:11:40
generation. It's not all variable. So
00:11:42
when you're able to amortize those fixed
00:11:44
costs over a greater supply, you bring
00:11:47
down the meter rate for everybody. And
00:11:48
so there's huge economies of scale. So
00:11:51
the more scale you get in electricity,
00:11:53
like most other things, the price comes
00:11:55
down. That's what So it's actually a
00:11:57
good thing that the uh that we have this
00:12:00
buildout going on because it will
00:12:02
ultimately reduce prices for consumers.
00:12:05
But we do have to make sure that these
00:12:08
new data centers aren't just plugging
00:12:09
into the grid and using, they have to be
00:12:11
contributing back.
00:12:12
>> And I think what a great policy change
00:12:14
has made under this administration, the
00:12:16
the B administration had as a matter of
00:12:18
policy had made it such that you
00:12:19
couldn't do this behind the meter energy
00:12:21
generation. Um, and if you wanted to
00:12:23
bring your own power, you couldn't. You
00:12:24
had to be part of the larger grid. So I
00:12:26
think um that rule has has changed uh uh
00:12:29
by by Secretary Wright and by FK to kind
00:12:31
of allow this to happen. And ultimately
00:12:33
I agree with David. I think once you
00:12:34
have sort of greater scale in in the
00:12:36
power generation, you'll be contributing
00:12:38
back into the grid in a way that that
00:12:39
benefits rateayers.
00:12:41
>> Let's go back to the uses and how AI is
00:12:44
changing our lives. You you mentioned
00:12:46
earlier um all of the uses and and and
00:12:49
the impact the AI is having. What do you
00:12:51
see as the most important use and where
00:12:54
AI is being deployed and implemented
00:12:57
best right now?
00:12:58
>> Well, it's interesting. I think it
00:12:59
there's been an evolution. So I think we
00:13:01
started with you know AI chat bots like
00:13:04
chatgpt and in a sense that was kind of
00:13:07
like better web search. Um it was really
00:13:09
great for research asking it questions
00:13:11
and give you answers to anything. Then
00:13:13
we saw um we saw models add chain of
00:13:16
thought and they could start to do you
00:13:18
know deeper reasoning. Then we saw
00:13:20
coding assistance. This is really and
00:13:22
and I think over the past few months
00:13:23
there's been a real breakthrough. If you
00:13:26
talk to people, software developers, it
00:13:28
really seems like there's been, you
00:13:29
know, a major shift in in just
00:13:31
improvement in the quality of the coding
00:13:33
assistants. And I think where that's
00:13:35
going next is um tools for knowledge
00:13:38
workers. So the same types of assistants
00:13:41
that have been outputting code can now
00:13:44
output any type of format. So whether
00:13:46
it's like Excel models, PowerPoints,
00:13:49
websites, you name it, knowledge workers
00:13:52
are now going to be able to generate all
00:13:54
these different types of things the same
00:13:56
way that coders have been gen that
00:13:58
software developers have been using AI
00:14:00
generate code. I think that's one of the
00:14:01
big things you're going to see in 2026
00:14:04
is again just this um this productivity
00:14:07
boom for knowledge workers. So I think
00:14:09
that's like one of the things you're
00:14:10
seeing on the ground. And then
00:14:12
separately there's there's a bunch of
00:14:13
things happening in industry verticals.
00:14:16
So different industries being impacted
00:14:18
by AI. So in healthcare I think there's
00:14:21
a tremendous opportunity uh to improve
00:14:25
um or to to reduce sort of
00:14:27
administrative bureaucracy to uh to
00:14:30
improve this um processing of paperwork
00:14:32
that happens. also to use uh AI and
00:14:35
medical and scientific research to help
00:14:37
find new uh cures. You're already seeing
00:14:39
users tell all sorts of stories about uh
00:14:42
diagnosis. They've been able to put in
00:14:44
their medical records into chat GBT or
00:14:47
you know other chat engine or chat bots
00:14:49
and get like remarkable results. They've
00:14:52
been able to you know finally figure out
00:14:54
what was you know what what was wrong
00:14:56
with them and they've been able to take
00:14:57
that to a doctor. You have doctors using
00:14:59
it too. So medical I think is a really
00:15:01
interesting area but there's a whole
00:15:03
bunch of these um examples of different
00:15:05
different industries are now being
00:15:06
impacted.
00:15:08
>> The the one area I think a lot about is
00:15:09
is AI for science and and back to to to
00:15:13
David's initial point about the progress
00:15:14
we've seen these frontier models. I
00:15:16
think the very early ones sort of
00:15:18
started with just general knowledge and
00:15:20
you have to go back and understand like
00:15:21
why and the question was what was the
00:15:23
data available for those model builders
00:15:25
to start training their models and for
00:15:27
the early ones you could just scrape the
00:15:29
internet and just kind of cram
00:15:30
everything into a model and train it and
00:15:32
and that's where you kind of had this
00:15:33
this first phase of of large language
00:15:35
models and the second one was coding and
00:15:38
if you think about how do you get a
00:15:39
really good coding model you again you
00:15:40
have to trade it you have to train it on
00:15:42
existing code and that's again something
00:15:44
that is you know relative atively easier
00:15:47
to to acquire than other types of data
00:15:48
and you saw great progress and jumps in
00:15:50
in the coding models. I think the the
00:15:52
third big sort of shift that hasn't
00:15:54
really been touched on yet which the
00:15:56
government itself is trying to do a good
00:15:58
uh push on is the AI for science
00:16:00
question and why it's so challenging for
00:16:02
scientific discovery to like tie in with
00:16:05
the way that LMS are are traditionally
00:16:07
trained is that the science data is
00:16:09
extraordinarily fragmented and it's not
00:16:12
done in a way or formatted in a way that
00:16:14
um can easily be applied to a large
00:16:16
language model sort of like training run
00:16:18
and if you think about scientific
00:16:19
discovery it's spread out across so many
00:16:21
different disciplines. You have
00:16:22
chemistry data, you have math data, you
00:16:24
have material science data and all of
00:16:26
that is is all types of different
00:16:28
formats. And our effort in
00:16:30
administration um we launched something
00:16:32
called the Genesis mission which are is
00:16:34
our attempt to sort of make these big
00:16:36
bold leaps in AI for scientific
00:16:39
discovery. and our national labs at the
00:16:41
Department of Energy are have been doing
00:16:43
incredible research over the last you
00:16:45
know 50 60 years and all of that has is
00:16:48
sitting and is ready to be used to be
00:16:49
trained for for for these models. So my
00:16:52
hope is that over the next year we're
00:16:53
going to see a lot more work in this in
00:16:55
scientific discovery to be able to
00:16:56
actually accelerate how quickly we can
00:16:58
choose which experiments to run, run
00:17:00
those experiments, go back and figure
00:17:02
out what we did wrong and run them
00:17:03
again. And and this ties in with lots of
00:17:05
interesting ideas that people have
00:17:06
around some of these AI labs where you
00:17:09
essentially have you can put in the the
00:17:10
the thesis or the hypothesis and
00:17:13
ultimately these labs can do lab
00:17:14
experiment itself and move forward. So
00:17:16
that's kind of the dream that I have
00:17:17
that that ultimately we as a country can
00:17:19
can almost double our our R&D output
00:17:22
over the next 10 years because of AI.
00:17:24
>> So so what kind of breakthroughs um
00:17:26
would you expect or would you like to
00:17:29
see? Yeah, I I think there um the ones
00:17:32
that I think can make a big impact are
00:17:34
uh first the the the the experimentation
00:17:37
and training runs around fusion um are
00:17:41
extraordinarily computationheavy
00:17:43
and they themselves if we can if we can
00:17:46
have a a a faster feedback loop on how
00:17:49
we do these these simulations for fusion
00:17:51
we can move the timelines in for fusion.
00:17:53
So that could be a big a big step.
00:17:55
Material science is also a very a very
00:17:57
big area where you want to be able to
00:17:59
test all types of of different molecules
00:18:01
and how interact with each other. This
00:18:02
is important for all the big things
00:18:04
we're trying to do in space. Whether
00:18:05
it's our lunar base or getting to Mars
00:18:07
or bringing nuclear energy to space,
00:18:09
having advanced material science is
00:18:11
important. And the third is one that
00:18:12
everyone always cares about is is
00:18:14
healthcare and and therapeutics. How can
00:18:16
you more quickly be able to identify the
00:18:18
the best molecules to solve a particular
00:18:20
particular health challenge? and how do
00:18:22
you more quickly iterate to a point
00:18:24
where you can move to a to a clinical
00:18:26
trial
00:18:26
>> and on a everyday level I mean you also
00:18:29
have the auto sector I think as a big
00:18:31
beneficiary here I think that's one area
00:18:33
that seems to be spending a lot on this
00:18:36
as well you agree with that
00:18:37
>> well I mean with like self-driving or
00:18:41
>> I mean self-driving for sure is going to
00:18:43
be huge it feels like we've hit some
00:18:45
sort of new inflection point there where
00:18:47
the quality's gotten to the point where
00:18:49
you're starting to see robo taxis now
00:18:51
Whimo and Tesla. Um
00:18:54
>> what what about an AI assistant? I mean,
00:18:56
is that going to be something that is
00:18:59
sort of common place? I someone said to
00:19:01
me the other day that oh, in China we're
00:19:02
doing things so much differently because
00:19:05
you're using AI for research as as you
00:19:08
said, but we're using it as I have my AI
00:19:11
assistant and I'm um you know, they're
00:19:14
paying my bills and cleaning my house
00:19:15
and buying my wife a birthday present
00:19:17
and and doing everything for me. I I
00:19:20
think so. I think that'll happen
00:19:21
probably this year. So the the product
00:19:24
that just came out recently that
00:19:26
everyone's kind of going crazy over is
00:19:28
the latest iteration of flawed code uh
00:19:31
which is uh powered by uh anthropics uh
00:19:34
opus 4.5 model which seems to be a real
00:19:37
breakthrough in in in coding and so
00:19:40
again this is you know the software
00:19:41
developers are really imp impressed with
00:19:43
it but in inside of cloud code they had
00:19:46
they introduced a new tab called co-work
00:19:49
again you can as a non coder
00:19:53
uh or as someone who is looking for um
00:19:56
to create output other than code, you
00:19:58
can now use it to uh to basically create
00:20:01
all sorts of other kinds of outputs.
00:20:02
Like I mentioned, you can do uh
00:20:04
spreadsheets or powerpoints, things like
00:20:07
that. And you can have it, you can point
00:20:09
it to your file drive and it can look at
00:20:12
the work you've already done. So if
00:20:14
there's a particular type of format for
00:20:16
a PowerPoint you like, you just point it
00:20:18
to the work you've already done and say
00:20:20
I want to do you know a new um you know
00:20:23
presentation but using this style but on
00:20:26
this topic and it'll actually emulate
00:20:29
you know your style and and the work
00:20:31
your format the work you've already
00:20:33
done. And um people are very impressed
00:20:36
with this and you can also point it at
00:20:37
your email and have it analyze your
00:20:39
email pull things out of it. So it right
00:20:42
now it's very taskbased. You you the
00:20:44
user have to prompt it for each task.
00:20:47
But you can see there the beginning of a
00:20:49
personal digital assistant where you
00:20:52
connect it to your file drive to your
00:20:54
email to all of your data sources and it
00:20:57
can start to do tasks for you and again
00:21:00
it understands the format and the style
00:21:02
that you like to produce work in. So, it
00:21:05
feels to me like we just need one more
00:21:09
layer of abstraction on top of a tool
00:21:11
like that and you'll have your own
00:21:13
personal digital assistant
00:21:15
>> and um you know there'll be like a voice
00:21:17
interface. You ever seen the movie Her
00:21:20
you know with uh walking Phoenix and um
00:21:23
I think Scarlett Johansson is just the
00:21:25
voice but uh you know he's telling her
00:21:27
what what to do through an earpiece. I
00:21:29
mean we're very close to something like
00:21:30
that. I mean, I'm not saying that, you
00:21:32
know, the AI is going to become sensient
00:21:34
or whatever, but um but no, we're like I
00:21:36
think in 2026, you could see that that
00:21:40
these types of of tools again started as
00:21:42
coding assistants, but now they become
00:21:44
personal digital assistants. That could
00:21:46
definitely happen this year.
00:21:48
>> Michael, what what don't people
00:21:49
understand about AI? What what do you
00:21:51
think is most important for us to
00:21:52
understand about the innovation underway
00:21:54
right now with science and and AI? I I
00:21:58
think some people I I think it's easy to
00:22:00
underestimate the the long-term impact
00:22:03
this is going to have across so many
00:22:04
industries and and domains. Um I think
00:22:07
very much, you know, it's easy to to to
00:22:10
quickly think about AI as a as just a
00:22:12
sophisticated chatbot because that's
00:22:14
what most people interact with every day
00:22:16
and and that's what they they they touch
00:22:18
and feel. Um, but I think that to me I
00:22:20
think the long-term impacts and not to
00:22:21
keep harping on the science, I think
00:22:23
there is a there's a a real fundamental
00:22:25
shift happening in the velocity and pace
00:22:28
that we can test and uh and evaluate and
00:22:31
execute scientific discovery and
00:22:33
endeavors. And I think I think that's
00:22:34
going to have huge repercussions for the
00:22:36
way that we as a country innovate
00:22:37
broadly speaking in the years ahead
00:22:39
>> which is why we're watching what China
00:22:41
is doing. Let's talk a bit about China
00:22:42
and where it is relative to the United
00:22:45
States. Are we winning? Is it about
00:22:47
chips? What's the race specifically
00:22:49
really about?
00:22:50
>> Well, I I I think that in general we're
00:22:53
ahead of China. There's different layers
00:22:55
of the stack. So, you've got the the the
00:22:58
models, then you've got the chips, and
00:23:01
you know, then you've got the chipm
00:23:02
equipment, you know. So, you go down the
00:23:04
stack. I would say that the deeper in in
00:23:06
the stack that you go, the greater the
00:23:07
American advantage. Um I think on models
00:23:10
most people would say that we're our
00:23:13
models are maybe 6 months ahead or so
00:23:15
plus or minus of the Chinese models. You
00:23:18
look at chips maybe 2 years ahead. You
00:23:20
go to the semiconductor manufacturing
00:23:23
equipment it could be like 5 years. So
00:23:25
the US does have sign significant
00:23:27
advantages there. There's only maybe a
00:23:30
couple of areas where I think China has
00:23:33
has an advantage. Um one is on energy
00:23:37
production. And if you look at the their
00:23:38
grid, their grid has roughly doubled in
00:23:41
the last 10 years, whereas ours has only
00:23:43
grown by about 2 to 3%. Energy
00:23:46
production in the US has been a
00:23:49
relatively sleepy industry before AI
00:23:52
came along. And a lot of that had to do
00:23:54
with regulations and the antipathy of
00:23:57
the previous administration towards
00:23:59
energy production. Obviously, President
00:24:01
Trump had a very different view on this.
00:24:03
I think he was preient on this issue.
00:24:05
you go back 10 years and he was talking
00:24:07
about we got a drill baby drill and um
00:24:10
and I think he understood that energy
00:24:13
growth was the precondition for economic
00:24:16
growth and it's definitely the pre
00:24:18
precondition for this uh AI
00:24:20
infrastructure growth. So this is an
00:24:23
area where again we have to basically
00:24:24
expand our energy production um and I
00:24:27
and and and so I think that is an area
00:24:30
where we need to catch up. The other
00:24:32
area where I would say, you know, I I
00:24:35
don't know if I would call this an
00:24:36
advantage exactly, but if you but you
00:24:40
could argue that China
00:24:44
has the edge in what is what's being um
00:24:48
called AI optimism. So there was a a
00:24:50
polling done by Stanford across
00:24:52
countries and they asked the citizens of
00:24:54
all these different countries uh do you
00:24:57
feel that the benefits of AI will be
00:24:59
more beneficial or more harmful? And if
00:25:03
if you thought that it that overall be
00:25:05
more beneficial than harmful they call
00:25:07
that AI optimism. Well in China AI
00:25:10
optimism was 83%. So 83% of the
00:25:13
population feels that it's being more
00:25:15
beneficial than harmful. That number in
00:25:16
the United States is only 39%.
00:25:19
So for some reason people in China are
00:25:22
more optimistic about AI than in the
00:25:25
United States and you generally you
00:25:26
generally see this that uh Asian
00:25:29
countries are very high on AI optimism
00:25:32
in the western countries are lower and I
00:25:35
think it's a interesting or open
00:25:37
question about why this is. I think
00:25:39
there's a few possible explanations for
00:25:41
it. I I think that um first of all, the
00:25:43
the media tends to focus on the doom and
00:25:45
gloom stories with with AI,
00:25:48
>> the fear,
00:25:48
>> the fears. Um and we can talk about some
00:25:50
of those fears and uh and and how, you
00:25:53
know, whether we think they're they're
00:25:54
real. Um but I think the media has a lot
00:25:57
to do with it. I think that the the way
00:25:59
that Hollywood has portrayed AI over the
00:26:03
decades, you know, with whether it's the
00:26:04
Terminator or 2001, uh has you has
00:26:08
portrayed this dystopian view of the
00:26:10
future. And I think that plays into
00:26:11
fuel's thinking. And then frankly, I
00:26:13
would say that part of the the fault
00:26:16
lies with our tech leaders who haven't
00:26:19
necessarily done a great job describing
00:26:21
the benefits of AI. In fact, when
00:26:23
they're talking about, you know, AI
00:26:25
eliminating 50% of knowledge workers,
00:26:28
that doesn't sound like a, you know,
00:26:29
very utopian scenario. That sounds
00:26:31
dystopian to most people. And so I do
00:26:34
think that unintentionally some of our
00:26:36
tech leaders have played into this um AI
00:26:39
pessimism. And the reason why I think
00:26:42
this could be a disadvantage for the
00:26:43
United States is because again it's
00:26:45
feeding into this regulatory frenzy
00:26:48
we're seeing again 1,200 bills at the
00:26:50
state level. And right now I think you
00:26:53
know we are winning this AI race. We're
00:26:55
ahead in all the key dimensions chips
00:26:57
models and so on. But we could shoot
00:26:59
ourselves in the foot, you know, if we
00:27:01
end up overregulating this thing to to
00:27:03
death, we could actually cost ourselves
00:27:05
this AI race. So, I do worry about this
00:27:09
question of AI optimism,
00:27:11
>> right? It's a great point. And how what
00:27:12
would happen if the US is not number one
00:27:14
in this, Michael?
00:27:15
>> Yeah, I I I think we we need to be and
00:27:17
that's why we put put the plan out. I
00:27:19
think you know when I think about the
00:27:20
the China question and about the the
00:27:22
sort of larger question of how do we win
00:27:24
the AI race what always what I always
00:27:26
like to think about is this question of
00:27:28
adoption and I think sometimes there's
00:27:30
this overemphasis on the leaderboard
00:27:32
it's like which frontier model is number
00:27:34
one on some sort sort of metric and the
00:27:37
reality is we're neck and neck and as
00:27:39
David said we're probably had you know
00:27:40
six to 12 months on our frontier models
00:27:42
but I think what we have seen over over
00:27:44
time and over history is that um you
00:27:46
don't necessarily need to have the very
00:27:48
best model or very best piece of
00:27:50
technology in the world for it to
00:27:52
perforate globally. And a lot of us who
00:27:54
were part of the first Trump
00:27:55
administration saw this very firsthand
00:27:57
with the telecom wars of that era of
00:28:00
what Huawei was able to do globally. And
00:28:02
at the time when when Huawei first
00:28:04
started their their sort of global
00:28:05
export push um they certainly were not
00:28:08
the very best technology in the world.
00:28:10
They were current they were certainly
00:28:12
you know you know subpar compared to to
00:28:14
Ericen and Nokia. that they were good
00:28:16
enough and they were subsidized enough
00:28:18
such that they became sort of the
00:28:20
default telecom um system for a lot of
00:28:22
the world and we've learned a lot of
00:28:23
lessons from that and we take that very
00:28:25
seriously. When it comes to AI, we know
00:28:27
there's ambition for the Chinese to
00:28:29
export their models and have them be the
00:28:32
models that are powering all these
00:28:33
different use cases across across the
00:28:36
global south and across the rest of the
00:28:37
world. Um, that's why the president
00:28:39
launched something called the American
00:28:40
AI export program and our mission and I
00:28:42
think we're in a very lucky position
00:28:43
here compared to what we're dealing with
00:28:45
with Huawei is as David said, we are
00:28:48
dominant in almost every part of the
00:28:49
stack. We have the very best models. We
00:28:51
have the various applications. We have
00:28:52
the very best chips. So, we are in a
00:28:54
position of power now and is up to us as
00:28:56
a country to share that technology with
00:28:58
the world with all of our partners and
00:28:59
allies. make sure that any developer
00:29:01
anywhere in the world that wants to
00:29:03
build a new application using AI is
00:29:05
using is fine-tuning an American model
00:29:07
on top of an American chip. And that
00:29:09
isn't that isn't a a hard reality to
00:29:11
see. That is something that I think we
00:29:12
can very easily do just because we have
00:29:14
the very best tech. That's a program
00:29:15
that um we launched last late last year
00:29:17
and we're doing a big push this year to
00:29:19
get that get that out the door.
00:29:20
>> It's an important point that you make in
00:29:22
terms of exporting AI to the rest of the
00:29:24
world. Is it true that China is telling
00:29:27
its companies don't use American chips,
00:29:30
don't use American AI right now?
00:29:33
>> It it it seems so. Um I mean China is
00:29:36
developing its own models. Obviously
00:29:37
about a year ago you had the Deepseek
00:29:39
moment where you you had a powerful
00:29:42
model released by Deep Seek and I think
00:29:43
that kind of put Chinese uh AI on the
00:29:47
map in a way. I think people in the west
00:29:49
didn't realize you in a way how good
00:29:51
China was at producing models and there
00:29:53
was a little bit of complacency
00:29:56
uh towards our relative position. People
00:29:59
weren't really talking about the global
00:30:00
competition two years ago. It wasn't
00:30:03
really discussed at all. Uh I remember
00:30:05
when you know the Biden administration
00:30:06
created this you know 100page Biden
00:30:09
executive order regulating AI. No one
00:30:12
was talking about whether this might
00:30:13
slow whether all this regulation would
00:30:15
slow us down. Visa v China wasn't even
00:30:17
part of the conversation. Then Deepseek
00:30:18
launched and I think we did realize
00:30:20
we're in a global competition and we
00:30:21
have to win and that's why we have to
00:30:23
actually be quite careful about how we
00:30:25
regulate this and not make sure we're
00:30:26
not overregulating it. But I think you
00:30:29
know China definitely wants to compete.
00:30:31
Um there have been some stories recently
00:30:35
I think uh Bloomberg and Reuters
00:30:36
reported that they actually are not
00:30:38
allowing Nvidia chips into their country
00:30:41
and the reason for that we think is that
00:30:44
they want to indigenize chip production.
00:30:46
They want to stand up Huawei as their
00:30:48
national champion and effectively
00:30:50
they're creating a market subsidy for
00:30:52
Huawei by keeping out the competition.
00:30:55
So, they're protecting their market to
00:30:56
stand up Huawei. And I think their plan
00:30:59
would be to have Huawei dominate chips
00:31:02
in China first and then use that to
00:31:04
scale up and then try to take over the
00:31:07
rest of the world. Chip production is a
00:31:09
scale up business. So, you know, if they
00:31:11
can dominate the Chinese market first,
00:31:13
that gives them a powerful platform to
00:31:15
then proliferate to the rest of the
00:31:16
world.
00:31:17
>> So, so where are we in that, Michael? I
00:31:19
mean, first you all came up with the AI
00:31:21
action plan, then came up with another
00:31:23
plan in terms of exporting AI to the
00:31:26
rest of the world. What can you tell us
00:31:27
in terms of where we are in that?
00:31:28
>> Yeah, so the the progress is is moving
00:31:30
on that. We um we closed a request for
00:31:33
information from the commerce department
00:31:34
late last year, which went out to
00:31:36
industry and said, "Hey, if we want to
00:31:37
export the American AI stack, what
00:31:38
should we be thinking about? How should
00:31:40
we be designing these packages that we
00:31:41
share with the world?" Commerce is now
00:31:43
ingesting that that information.
00:31:45
There'll be a request for proposals that
00:31:47
comes out very shortly. And that's where
00:31:48
we actually want companies to come
00:31:50
together to form consortia and say like
00:31:52
look this is what a package looks like.
00:31:54
And I think what um you know what what
00:31:56
people need to sort what I always try to
00:31:58
remind people is that the the the the
00:32:01
buyers of AI around the world um vary
00:32:04
quite dramatically in their level of
00:32:07
sophistication. So in the US, if you're
00:32:09
a very sort of, you know, if you're a
00:32:11
Fortune50 company and you want to deploy
00:32:13
AI, you have a pretty sophisticated sort
00:32:15
of CIO or CTO shop, you are thinking
00:32:17
very carefully about like which cloud
00:32:19
you want to buy, which potential model
00:32:20
you want to use, do you want to
00:32:21
fine-tune it on your own data? Do you
00:32:23
want to build your own application? You
00:32:24
know, what application you go and see?
00:32:25
You can like test various things. You
00:32:27
like go to all these third parties and
00:32:29
evaluate which is best. And it's a very
00:32:31
sort of complicated mix of how you end
00:32:32
up creating something that's optimum for
00:32:34
your particular company. for a lot of
00:32:36
countries around the world that are
00:32:37
aspiring to to use AI for their people
00:32:40
or to support the services whether it be
00:32:42
health care or um you know tax
00:32:44
collection or whatever it may be um you
00:32:46
know they don't have a a you know
00:32:49
billion dollar IT budget you know
00:32:51
they're just trying to figure out what
00:32:52
is a tool that I can use in my country
00:32:54
to deliver the benefits of AI to my
00:32:56
people so we think very carefully around
00:32:58
how can we craft solutions which you
00:33:01
know turn keys could be one way to put
00:33:03
it or how do you how do you provide a
00:33:04
solution that can easily be deployed in
00:33:06
a country and what's often you know what
00:33:09
often sort of gets caught up in this
00:33:10
debate is this question of you know how
00:33:12
many chips is the US going to be sending
00:33:14
around the world and and what I always
00:33:15
try to remind people is that you know
00:33:17
outside of the US China and maybe a few
00:33:20
other countries most countries around
00:33:22
the world do not have the capital or the
00:33:24
aspiration to do largecale training runs
00:33:27
or development of their own frontier
00:33:28
models there are very few countries
00:33:30
around the world that are going to build
00:33:32
sort of colossus style training centers
00:33:34
most countries around the world need
00:33:36
smaller data centers that just have
00:33:37
inference related chips that can drive
00:33:39
and and and do the you know do the
00:33:42
inference on on the particular um runs
00:33:45
that the government wants to have. So I
00:33:47
think what we're working very hard to do
00:33:49
is is is create sort of these these
00:33:51
these turnkey manageablysized AI
00:33:54
solutions that then we can partner with
00:33:57
a lot of our export finance
00:33:59
organizations like Development Finance
00:34:00
Corporation or the Export Import Bank to
00:34:03
make the export of that particular stack
00:34:05
much more appealing and commercially
00:34:08
viable in countries that are not
00:34:09
extraordinarily deeped. Um so we're
00:34:12
going to be in India next month for the
00:34:13
India um AI impact summit. Um, this is
00:34:16
sort of the largest global gathering for
00:34:18
for AI folks. Um, and we're going to be
00:34:20
sharing a lot more on on the progress of
00:34:22
this uh of this program there.
00:34:24
>> You want to weigh in?
00:34:25
>> Well, I just just to build on that. I
00:34:27
think people sometimes ask, you know, h
00:34:30
how do you how will you know if if
00:34:32
you've won the AI race, you know, with
00:34:34
with with China with with other
00:34:35
countries and I think there's a very
00:34:37
simple answer to that which is market
00:34:39
share. You know, if 5 years we look
00:34:41
around the world and we see that it's
00:34:44
American chips and models are being used
00:34:46
everywhere, well that means we won. But
00:34:47
if in 5 years we look around the world
00:34:49
and it's Huawei chips and Deepseek
00:34:51
models, then that would be very bad,
00:34:53
right? That would be a bad sign. That
00:34:55
means that we lost. So I do think that
00:34:56
the proliferation or diffusion of
00:34:59
American technology is really critical
00:35:01
to winning this AI race. We know from
00:35:03
Silicon Valley that the companies that
00:35:06
end up becoming huge are the ones that
00:35:08
create ecosystems. It's the, you know,
00:35:10
you you you as a technology company, you
00:35:13
want to have the most apps in your app
00:35:14
store. You want to have the most
00:35:15
developers writing on top of your API.
00:35:19
You want to be a platform company. And
00:35:21
so in all these technology races,
00:35:24
biggest ecosystem wins. And we want to
00:35:26
have the so that's basically why I think
00:35:28
this program is so important is we want
00:35:30
to create the biggest ecosystem. Now
00:35:33
this is not only about benefiting the US
00:35:35
because in order to have a successful
00:35:37
ecosystem you have to create value for
00:35:39
your partners and that's really
00:35:41
important like Michael's saying not
00:35:43
every country is going to be on the
00:35:44
cutting edge of developing its own chips
00:35:46
or developing its own frontier models
00:35:48
but they can use these tools to derive
00:35:51
value to apply them to their businesses
00:35:54
to their economies to extract value and
00:35:56
be part of this technological
00:35:58
revolution. So I think that you know we
00:36:00
have to think in this with this partner
00:36:03
mindset and I do think that this this
00:36:06
type of mindset is actually very common
00:36:08
to Silicon Valley. Like I mentioned I
00:36:09
think every great technology company
00:36:11
thinks in terms of how do we get the
00:36:13
most people on top of our tech stack but
00:36:15
it is a form of thinking that's pretty
00:36:17
alien to the bureaucracy in Washington
00:36:21
which has much more of a command and
00:36:22
control type of mindset. Y
00:36:24
>> and when President Trump came into
00:36:26
office, just give a couple examples of
00:36:28
this, the regulations that were sitting
00:36:30
on our desk that had just been handed
00:36:32
down by our predecessors, again, we had
00:36:34
this 100page Biden executive order on AI
00:36:38
that was all this new regulation and
00:36:40
there was a 200page
00:36:42
uh was called the Biden diffusion rule,
00:36:44
which was 200 pages of regulations
00:36:47
uh on the export of semiconductors. So
00:36:50
we were turning the the AI industry
00:36:53
models and chips into a highly regulated
00:36:55
industry. That was that was basically
00:36:57
the direction that Washington was going
00:36:59
in. And the first thing President Trump
00:37:02
did his first week in office was rescend
00:37:04
all of those unust regulations which I
00:37:06
think was absolutely critical. You know
00:37:08
the thing that really makes Silicon
00:37:10
Valley special is this concept of
00:37:12
permissionless innovation. you know,
00:37:14
since um Hulin and Packard started 85
00:37:17
years ago, started building Silicon
00:37:20
Valley, it's it the idea has always been
00:37:22
that just a couple of founders kind of a
00:37:24
great idea start their company, they get
00:37:26
some angel investors to write, you know,
00:37:29
a check for, you know, seed capital.
00:37:32
Those investors think they're probably
00:37:33
going to lose their money, but they
00:37:34
figure there's a shot. and you know and
00:37:37
it's so it could be the two guys in
00:37:39
their garage or it could be the college
00:37:40
dropout in the dorm room and they don't
00:37:43
need to go to Washington to get
00:37:44
permission for their idea right it's
00:37:46
permissionless innovation that's what's
00:37:48
has made Silicon Valley the crown jewel
00:37:50
of the world it's why so many of the I
00:37:52
think heads of state who are here are
00:37:54
always asking how do we create our own
00:37:56
Silicon Valley that was not the
00:37:58
direction we were on when President
00:38:00
Trump came into office the new 300 pages
00:38:03
of regulations concerning AI the Biden
00:38:06
administration left us with would have
00:38:08
changed this um environment of
00:38:10
permissionless innovation to an
00:38:12
environment of you have to go to
00:38:13
Washington to get approval for your idea
00:38:16
and I think that President Trump really
00:38:18
corrected that and since then we've been
00:38:20
implementing you know his AI action plan
00:38:23
uh which is all about you know pro
00:38:24
innovation pro infrastructure pro energy
00:38:26
and pro export so it's been I think a
00:38:29
total change and I think just in the
00:38:31
past year you've seen the results of
00:38:33
that
00:38:33
>> and I think what one thing to to add
00:38:35
there Um part of the the international
00:38:38
uh agenda that we have on AI is one
00:38:40
obviously let's let's do the export but
00:38:42
the other piece is trying to share with
00:38:44
all of our partners and allies how you
00:38:46
can actually create a regatory
00:38:48
environment that allows this technology
00:38:49
to succeed and here we are in Europe and
00:38:52
I think many of us that sort of have you
00:38:54
know tried to work with technology
00:38:56
companies in Europe have have hit sort
00:38:58
of a lot of roadblocks and a lot of
00:38:59
stumbles and no matter you know the drug
00:39:01
reporting came out and and he can say
00:39:04
that there's a lot of issues But things
00:39:05
don't ever seem to seem to really
00:39:07
change. And I think all of that that the
00:39:10
the the way that our regulatory
00:39:12
structure is is designed in the US and
00:39:15
the way that the entrepreneurial spirit
00:39:16
thrives in the US is something that we
00:39:18
try to share with countries all around
00:39:20
the world. And I think the the the
00:39:23
general um knee-jerk reaction for most
00:39:26
policy makers around the world is one
00:39:28
that moves to a corner that is obsessed
00:39:31
with the precautionary principle. this
00:39:34
concept that every time something new
00:39:35
comes out, the role of the policy maker
00:39:37
is to sort of like sit in a room and
00:39:39
whiteboard everything that could go
00:39:41
wrong and then design regulations to
00:39:44
make sure those wrong things, these
00:39:46
hypothetical wrong things don't happen.
00:39:48
When in reality, what we do in the US,
00:39:50
what we try to do is sit in a room and
00:39:51
whiteboard what rules we can create to
00:39:53
actually unlock innovation. What are the
00:39:55
ones we should remove to allow more
00:39:57
innovation to happen? And I think that
00:39:59
mindset is something that we constantly
00:40:01
try to share at all these international
00:40:02
fora. The US has you know there has been
00:40:05
an AB test on what regulatory structure
00:40:08
works and what succeeds. You know we've
00:40:10
seen how the how how Europe has
00:40:12
approached this in the last 20 years and
00:40:13
we've seen what the US has done. So I
00:40:15
think the the recipe is kind of obvious
00:40:16
but but sometimes we have to just keep
00:40:18
repeating it to to our counterparts. And
00:40:20
I love the Draghy report because it was
00:40:22
so clearly uh identifying companies that
00:40:26
are in Europe that you know like no Novo
00:40:29
Nordus is like 350 billion or $400
00:40:32
billion company and in America we've had
00:40:34
companies of trillion dollar companies
00:40:37
Nvidia hitting$5 trillion. So so what is
00:40:40
the path to innovation?
00:40:43
Well, I I I think part of it is and I I
00:40:46
I think this is the difference between
00:40:49
maybe the American mindset and the
00:40:50
European mindset towards this is that
00:40:53
ultimately the innovation in the United
00:40:55
States comes from the private sector. It
00:40:57
comes from the entrepreneurs, the
00:40:58
founders, the innovators, the geniuses
00:41:00
with an idea. And I think that the
00:41:04
government sees its role, at least when
00:41:06
it's thinking properly about this, as
00:41:08
being an enabler and as just setting the
00:41:11
rules of the road. um and maybe putting
00:41:13
in some guard rails, but basically it's
00:41:15
letting the entrepreneurs cook and
00:41:18
that's how you get innovation. And now I
00:41:21
don't want to bash our European host too
00:41:24
much, but you know the when when the uh
00:41:27
when the EU talks about AI leadership,
00:41:30
they're talking about the regulators and
00:41:32
they think their value ad is well, we're
00:41:34
going to we're going to show the whole
00:41:35
world the regulatory model for AI. So,
00:41:37
it's kind of a bad case of u main
00:41:40
character syndrome where uh you know
00:41:42
where like the regulators think they're
00:41:44
the main characters in this. No, look,
00:41:46
the regulators are the supporting
00:41:48
players. The main characters always have
00:41:50
to be the entrepreneurs. It's got to be
00:41:52
the innovators. That's how you unlock
00:41:54
innovation. When the when you start to
00:41:57
see yourself, I mean, the the regulators
00:41:59
and the policy makers as the as the main
00:42:01
characters, that's not a great recipe
00:42:03
for innovation. And I think just just a
00:42:05
minor point on the on the AI stuff in
00:42:07
Europe that you know the EUAI act which
00:42:09
has been so detrimental to to the AI
00:42:12
ecosystem here here in Europe was passed
00:42:15
before chat GPT was even invented and
00:42:18
that shows the challenge here. You're
00:42:20
you're you're believing that you can
00:42:22
solve some kind of problem or some
00:42:23
you're solving something but the end of
00:42:25
the day innovation is moving so much
00:42:26
more quickly and ultimately that that
00:42:28
rule makes no sense now in a world of of
00:42:30
frontier models large language models
00:42:32
and they have to sort of edit it. So,
00:42:33
let me push back before we go and ask
00:42:36
you to identify any risks or threats or
00:42:40
downside risks in all of this. What
00:42:43
should we be worried about if anything
00:42:45
with regard to AI usage?
00:42:48
>> Well, I I think there are Orwellian
00:42:51
scenarios uh of AI that I think we
00:42:54
should be concerned about. And again, I
00:42:57
I tend to think that those scenarios
00:42:59
were described by George Orwell, not by,
00:43:01
you know, James Cameron and the
00:43:02
Terminator. And specifically, it's
00:43:04
misuse of AI by government. I do think
00:43:07
that AI could be used as a tool to um
00:43:12
surveil, to censor, to even potentially
00:43:15
brainwash the population. This is why
00:43:17
the administration has taken such a firm
00:43:19
stance against what it's called woke AI,
00:43:21
which I almost think that that name
00:43:23
maybe trivializes the magnitude of the
00:43:26
problem we're talking about. We're
00:43:27
talking about AI having a political bias
00:43:29
built into it. Um, and it the bias can
00:43:33
be so subtle that people don't even
00:43:34
necessarily notice over time, but it has
00:43:36
a huge impact on what people are allowed
00:43:39
to learn and think and know and what you
00:43:41
know children learn. And so I think it's
00:43:44
very important that we try to make sure
00:43:46
that AI was politically unbiased. Um
00:43:49
there just in this regard, one of the
00:43:52
things that we were so concerned about
00:43:54
with that Biden executive order on AI
00:43:56
that we were sended in the first week is
00:43:58
that it had 20 pages of language on DEI
00:44:02
and it was promoting this idea that AI
00:44:05
models need to build in a DEI layer.
00:44:08
Well, you know, this is how you ended up
00:44:10
with, you know, the the the black George
00:44:12
Washington uh you know, story where that
00:44:15
the first version of of uh Gemini came
00:44:18
out and it was, you know, it was
00:44:19
basically rewriting history to serve a
00:44:22
current political agenda of DEI. And um
00:44:25
you know that that was in a way that
00:44:27
that that case of bias was so ludicrous
00:44:30
that everyone kind of laughed at it. But
00:44:32
it gives you a sense of what could
00:44:34
happen if you start to build the the
00:44:37
bias into AI and you know that same you
00:44:41
know so-called trust and safety
00:44:43
apparatus that was starting to be built
00:44:45
into social media sites as a way to
00:44:48
censor and deplatform and shadowban. You
00:44:51
could see that being built into AI
00:44:53
models as a way to control uh the the
00:44:57
public discourse in in a very serious
00:44:59
way. And I think that, you know,
00:45:02
President Trump again just put a total
00:45:04
halt to that, you know, rescended that.
00:45:07
But it was also we also um President
00:45:09
Trump signed an executive order saying
00:45:11
that the federal government would not
00:45:12
procure politically biased AI. So look,
00:45:16
on a first amendment basis, if an AI
00:45:18
company wants its AI to be biased in
00:45:21
some direction, they probably have a
00:45:23
first amendment right to do that. But we
00:45:26
have as the federal government have the
00:45:28
discretion not to buy that software and
00:45:30
we've said that we won't. So, I feel
00:45:33
very good that during President Trump's
00:45:35
uh term in office for the next three
00:45:37
years, this idea of Orwellian AI is not
00:45:40
going to be a problem. But I do worry
00:45:42
that at some point in the future, if you
00:45:44
had a different regime in Washington,
00:45:47
you know, if the federal government
00:45:49
started to pressure AI companies to
00:45:52
build in this political bias, that would
00:45:54
be a very serious threat, I think, to to
00:45:57
our freedoms. It's a it's a great point
00:45:58
to make. Before we wrap up real quick on
00:46:00
jobs, can either of you explain what
00:46:03
Elon Musk is is saying about the impact
00:46:05
of AI said we're not going to need to
00:46:07
work. You know, the the AI AI is going
00:46:10
to do it all. I I just I'm trying to
00:46:12
understand what he's saying that it's
00:46:14
we're going to go on holiday. Um jobs
00:46:16
are going away and AI is going to do
00:46:18
everything. Well, Elvon's a friend of
00:46:21
mine and um I'll I'll uh I I'll disagree
00:46:25
with him slightly on this, but um but
00:46:27
but let me just
00:46:29
the his comment about the the job loss
00:46:32
obviously is what gets all the
00:46:33
headlines, but at the same time he's
00:46:34
saying that he's also saying that in
00:46:37
this future there's going to be so much
00:46:39
abundance that everyone's going to have
00:46:40
what they want and there's not going to
00:46:42
be any money. So people people leave out
00:46:46
that part of the story and they just
00:46:47
report Elon says everyone's going to
00:46:48
lose their jobs. No, we're talking about
00:46:50
a radically different future. It could
00:46:52
be the future. It's kind of described in
00:46:53
Star Trek, you know, where like there is
00:46:56
no money because we have everything.
00:46:58
Look, I I think that, you know, Elon is
00:47:01
directionally correct about the future.
00:47:02
I think we are heading toward toward uh
00:47:05
towards a world of much greater
00:47:06
abundance, rising living standards for
00:47:09
everybody, greater productivity. I think
00:47:11
that will lead to rising wages. I don't
00:47:13
think it's going to put everyone out of
00:47:14
work. I don't think that's going to
00:47:15
happen. Uh but again, the timelines
00:47:18
matter a lot. And you know, getting to a
00:47:20
world with no money is not something
00:47:22
that's going to happen in the next 5
00:47:23
years.
00:47:24
>> And of course, Michael, this is helping
00:47:25
us um in terms of longevity and living
00:47:29
longer, right? In terms of the impact on
00:47:31
science.
00:47:32
>> Totally. I think generally the the
00:47:34
abundance story extends itself well into
00:47:36
into, you know, health care and
00:47:39
everywhere else that and and just
00:47:40
quality of life. So, good things ahead.
00:47:42
I think
00:47:43
>> we'll leave it there. Michael Katzios
00:47:45
and David Saxs, thanks so much.
00:47:46
>> Thank you.
