The OpenAI Internet Browser Has Arrived: ChatGPT Atlas w/ Dave Blundin & Alexander Wissner-Gross

OpenAI has launched a full-blown

browser. The competitive positioning

versus Google is basically all out war.

>> Today, we're going to launch Chat GPT

Atlas. This is an AI powered web browser

built around Chat GPT. We think that AI

represents like a rare once a decade

opportunity to rethink what a browser

can be about.

>> Okay, great. But Google's going to come

in and do at least this and, you know,

take back any market share they lose.

>> I don't think we should think of it as a

product. I think we should think of it

as a distribution channel for open AI's

super intelligence.

Having a local agent mode I I think is

potentially transformative.

>> If Sam wins the data aggregation race,

if he falls behind for a month or a year

in the AI race, he still has your data.

>> We're going to have an AI that is our

personal portal into everything. And I'm

not going to care what browser I use.

I'm just going to be able to have a

conversation with my AI and it will pull

up the data from wherever it is, whether

it's using super intelligence from, you

know, OpenAI or Google.

>> Now, that's a Moonshot, ladies and

gentlemen.

>> Everybody, welcome to Moonshots, another

episode of WTF Just Happen in Tech. I'm

here with my moonshot mates Dave Blondon

and AWG Alex Wemer Gross. Good morning,

gentlemen. Hey, good morning. Hey, and a

huge th shout out to the team. You know,

we were going to shoot this podcast last

night and Alex had so much material that

happened in the last three days that we

just needed to get in here. I mean,

things are changing so quickly. So,

basically, the team pulled an allnighter

last night to pull together these

stories and it's epic. So, thank you

team behind the scenes.

>> Yeah. Right now, our fourth Moonshot

mate, uh, Seem is on an airplane. Uh, I

just spent the last four days with him

here in Calamigos in Malibu for X-Prize

Visionering 2025,

which is a story I want to open up with.

Dave, I wish you were here. Alex, I wish

you were here. Uh, it was awesome. So

for those you don't know X-P prize every

year gets together uh our brain trust

and our benefactors and we debate and we

discuss what are the problems that

aren't being solved that need to be

solved or what are the challenges that

are too far out and we need to

accelerate them and bring them forward

and that's visionering. It was an

amazing uh 2 and 1/2, well really four

days in total, but two and a half days

in which we raised uh Dave, you're on my

board here at X-P Prize. Uh we raised

$3.5 million in capital last night.

>> So uh so that

>> do every night. That's a billion dollars

a year.

>> Yeah,

that would be awesome. And someday we

will.

>> Just so the audience knows the

sacrifices Peter makes to bring you all

of this information, you know. So he's

on stage all day yesterday. uh tomorrow

boards a flight to Riad. So we'll be in

Saudi. He'll be on stage the day after

that with Eric Schmidt kicking off that

event. That's 10 time zones away. So

watch the footage of him from Riad and

see see what that looks like.

>> How how wired will I be on caffeine?

>> Oh my god. It's great. But you know we

announced yesterday the uh our impact

report for X-P prize and uh the numbers

are staggering. We have massive detailed

report uh and it's we every dollar

invested in a prize we get a 60x return.

So you know milliondoll prize is driving

$60 million of R&D invested by all the

teams. They're all optimist. They all

think they can win and they're all sort

of like a Darwinian evolution to go and

solve these problems. So super pumped

about that. But I want to report you

know this is the first group to hear

about it on who won X-P prize

visionering. So we enter uh the 2 and

1/2 day program with about 20 concepts.

We have five different domains, five

different grand challenge areas and

we've got uh four concepts per. We

narrow it down to two and then down to

one which leaves us with five that enter

the uh battle royale as we call it. and

we go from five to three and then last

night uh we got down well let me just

show you the the numbers here. So X-P

prize visionering winners for 2025

uh we were expecting to just have one of

these prizes get funded to go into

development. It turned out all three of

these got funded to go into development.

Let me mention what they are because I'm

very proud of them. Uh the first prize

is called abundance and which you know

got to love the name. Uh and it was

actually two of our abundance 360

members who proposed this and raised the

capital to get this going. So what is

the abundance x-priseze? It is deliver

to a community food, water, housing,

electricity, and bandwidth for $250 a

month.

That's the goal. So everything that you

basically need and you know the

conversation last night we can talk

about this is

>> uh there's potential for a lot of civil

unrest right as people start losing jobs

as you know uh subgroups start becoming

wealthier and we've talked about this

I'm absolutely clear in the next decade

we're going to have you know

extraordinary abundance uplifting

everybody but it's this turbulent period

of the next 2 3 4 5 years that's

concerning and the idea here is if all

of a sudden moms and dads have all of

their bases covered. Um, you know, the

the basics of life for 250 bucks a

month, then they can start to think

about, okay, how do I use AI? How do I

use this technology to be an

entrepreneur to create a better life?

Any thoughts on that, Dave? Well,

especially that last fundamental of

food water shelter bandwidth

you know, if you're going to contribute

in this global revolution, I love the

fact that they added that as a

fundamental necessity.

>> Uh, you know, inside the 250 buck limit.

That's just just such a great great

idea. But that unlocks your ability to

contribute to make a living to get

educated. You know, every all education

will move to AI so you can have a you

know the healthare

>> healthcare all of that that ties to

bandwidth. So, it really is a

fundamental necessity. I love it.

>> Yeah. Alex, any thoughts?

>> Yeah, this sounds a lot like a universal

basic services concept. UBS is sort of

the symmetric duel to UBI, universal

basic income. I'm very bullish on

universal basic services in general. I I

think I I would expect it's an artifact

of a mature economy that the the cost of

living can be driven down to to near

zero as part of a sort of lifestyle

subscription and Amazon super prime if

you will.

>> Yeah. I know super excited about

>> there's a lot of studies that say that

universal basic income backfires in

terms of it causes depression, causes

alcoholism, causes drug use, but

services uh you know where you actually

get the things you need to survive still

encourages you to work and contribute on

top of the service. It's a much better

idea. But we learned in our pregame here

that that Alex doesn't even use

caffeine. So I don't know how that's

possible but

>> caffeine is a universal basic service

for sure. Uh so so this uh won the most

uh capital last night and it's going

into prize development. I'll report on

it. We'll have this team at the

abundance summit. Both of them are

abundance members. Uh and uh we'll talk

about it. The second prize uh

surprisingly that got top honors and

received enough enough capital to go

into development is a fusion XP prize.

And so here I am thinking okay there are

37 fu uh ventureback fusion companies.

There's about $10 billion invested into

fusion. What do they need a fusion prize

for? And uh amazingly and I met with uh

there were four fusion companies, you

know, four, you know, solidly funded

ongoing fusion companies as well as uh

some of the top faculty, one professor

at MIT and saying, "No, no, we need an

X-P prize to move this forward. We need

the public to understand how this

important is and how the government

needs to come in and support it." So,

uh, this one is not fully defined as a

prize, but, uh, $500,000 was committed

to develop the prize and move it

forward, uh, into potentially a prize.

Um, you know, Alex, I think you have

some feelings about this one. Yeah, I I

think fusion is is already well

capitalized, but I I would say

ultimately to the extent that the limits

of economic growth are bound by our

ability to solve fusion. I I think on

the margin it would be more helpful to

to allocate more capital toward fusion

energy sources and perhaps this helps

with that. Alex, you know, you know what

happens after this visionering phase is

the world's greatest experts on the

topic all get together, you know, in the

Peter versse and then they contribute

all their ideas and not all of them get

from there to actually being a prize,

but you learn so much about the state of

what's happening along the way. So, I

love it when a topic like Fusion gets

through this part of the funnel

regardless of how it ends up because the

amount of information we'll bring back

into the podcast on this will be just

immense.

>> You know, it's interesting. Uh the CEO

of Commonwealth Fusion, Bob Mumgardner,

um is

uh going to be with us in Riad and he's

going to be on stage with me at the

Abundance 360 Summit in March. And I was

on the phone with him getting ready for

what we're going to be doing in in Riad

next week. And he said, "Listen, I heard

that you're talking about a Fusion X-P

prize. I am so excited about that." And

so here we have the best funded, most

advanced fusion company.

Actually excited about a Fusion X-P

prize. So I'm excited to dig in further.

All right. The third prize is actually

something I love. Uh it's called Wall-E.

We'll have to be uh uh you know in

debate and discussion with with Disney

about this. But here's the prize. Dump a

machine into a garbage dump. And the

machine sorts the trash and and

generates uh you know piles of metals

and foods and paper and basically can we

take our current uh you know what do you

want to call them landfills and actually

reutilize them. So I have the way I

would actually win it but uh I don't

know I think this is a convergence of

technologies. It's going to be AI it's

going to be robotics. It's going to be

material sciences. Any thoughts, Dave?

>> Alex brought us Alex keeps bringing us

deal after deal after deal and every one

of them so far has been a winner. Um, so

it's really exciting. But, uh, Alex, you

brought us that rare earth company. You

want to talk about that? And I learned a

lot about this just studying that

company.

>> Maybe just a a broader comment on the

space. I I I think there is such a long

tale of physical world service jobs that

are ripe for automation, not just

limited to repurposing junkyards, as it

were. But I I I think if you look around

the world today, I I I often sort of

look out in the street and you can

ponder where are all the robots where

we're supposed to be living in the

future. Why haven't we seen anything

that that looks facially transformative

when you look out in the street? I I

think in the next 5 to 10 years, we will

look out onto the street and we will see

an abundance of robots and physical

automation that enables communities to

be visually transformed. Uh aesthetics

that would otherwise be out of reach for

an economy our size, as the economy

starts to grow radically, we'll start to

deploy robots everywhere for for even

the most minor tasks that would be

otherwise economically inaccessible

today. So I think this is actually just

maybe a special case of a much broader

opportunity over the next 5 to 10 years

of just deploying automation everywhere.

Every week my team and I study the top

10 technology metat trends that will

transform industries over the decade

ahead. I cover trends ranging from

humanoid robotics, AGI and quantum

computing to transport energy longevity

and more. There's no fluff, only the

most important stuff that matters that

impacts our lives, our companies and our

careers. If you want me to share these

meta trends with you, I write a

newsletter twice a week, sending it out

as a short two-minute read via email.

And if you want to discover the most

important metat trends 10 years before

anyone else, this report's for you.

Readers include founders and CEOs from

the world's most disruptive companies

and entrepreneurs building the world's

most disruptive tech. It's not for you.

If you don't want to be informed about

what's coming, why it matters, and how

you can benefit from it. To subscribe

for free, go to dmmandis.com/metatrends

to gain access to the trends 10 years

before anyone else. All right, now back

to this episode. I just wanted a robot,

you know, just walking up and down the I

10 freeway, the 405, picking up the

trash on the side of the road. But the

idea that we can actually take our

landfills, which are have so many

different problems and everything from

methane production to just, you know,

disease and we're sending so much of our

trash overseas to Southeast Asia uh with

heavy metals. I mean, the idea that we

can actually use it as a feed stock um

is amazing. So, I don't want to belabor

the point. Congratulations to the teams

that won X-Prize Visionering.

Congratulations to Nusan and Shari and

the entire leadership team of X-Prise.

It was an awesome two and a half days.

Uh and we recorded a podcast which we

dropped a couple of days ago. So uh you

know we had Immodust uh and Eric Pulier

See and myself being a podcast.

Hopefully everyone listening has heard

that one. We're going to be expanding on

some of the ideas because I want to make

sure to bring in the, you know, the

brilliance and vision of AWG and Dave.

All right, let's move on. Uh, the, uh,

main course today, AI chips and data

centers as it is every day.

All right. Do you want to introduce this

video uh, Dave or or AWG?

>> Yeah. Uh, so so this is one of the

reasons we needed to get together

quickly. This is this just came out. Uh

so OpenAI has launched a full-blown

browser. Uh the functionality it won't

blow you away yet, but the positioning

the competitive positioning versus

Google is basically all out war. So I

went back and researched you know Google

launched Chrome. Chrome was not in the

world. You know it's people don't

remember this and it they leveraged

their user base to install it and now

they have two-thirds market share of

browsers and so this is people's point

of contact with information goes through

Google. They get to see everything you

do. Then later on they turned on Chrome

sync so they watch everywhere you

navigate. All that information goes back

into Google's great AI machine and

serves you ads. Brilliant and kind of

kind of scary. Uh so then OpenAI, you

know, Sam being the strategic genius

that he is says, "Okay, this is one of

those fundamental Bill Gates style

points of control we absolutely have to

play in the browser game." So, we're

going to launch the Atlas browser, and

what's going to make it better than

Chrome is it's going to learn what you

like and don't like far far better and

use our our AI advantage to serve up

better ideas. And the integration of GPT

and what you're browsing will be

completely seamless. It'll be advising

you. It'll be taking you to the next

website. It'll be curating your news all

through that integrated browser.

>> It'll be taking your data.

>> Yeah. Yeah. That too. I mean, that's

that's the key, right? You know,

>> let's watch this a short video of Sam

and his team announcing Atlas and then

we'll talk about it.

>> We're going to launch Chat GPT Atlas,

our new web browser. We think that AI

represents like a rare once a decade

opportunity to rethink what a browser

can be about and how to use one and how

to sort of most productively and

pleasantly use the web. And then there's

three special core features of Atlas

that Ryan's going to walk you through in

a bit. The first is chat comes with you

anywhere as you go on the web. The

second big feature is browser memory.

The third which we're really excited

about and uh Justin's going to show this

later is agent which is in Atlas Chatbt

now can take actions for you. It can do

things.

>> All right. So my first reaction is okay

great but Google's going to come in and

do at least this and you know take back

any market share they lose. I don't

know. Do you agree with that? Alex, what

are your thoughts? I I I think there's a

misconception that Atlas is a product. I

don't think we should think of it as a

product. I think we should think of it

as a distribution channel for OpenAI's

super intelligence. I I think all of

these products, these discrete products

are just going to dissolve over the next

few years into a uniform medium of

distribution for super intelligence. So

whether it's one browser on the desktop

versus another browser competing, I

almost think it's the wrong question.

And I think the the right question is

what form of back-end super intelligence

is being surfaced via which channels.

Browser is one, intelligent code editor

environments are another. I think robots

and and various wearable devices are

going to be another over the next few

years. And I think it's really the super

intelligence at the end of the day

that's the differentiation less the the

particular Chrome if you will that is

that that's just an embodiment of it to

deliver it to the user. And I I think

along those lines, the most interesting

for me part of the Atlas launch was the

agent mode. Uh less so the the other

features having a local agent mode I I

think is potentially transformative for

for a number of use cases and feels a

little bit more sophisticated than prior

agent launches that we've seen from

OpenAI. If you remember Operator or if

you remember the cloud-based chat GPT

agent, this one is at least partially

local.

>> So you've got you've got the the big big

guys with infinite budgets. So you've

got Google, you've got Zuck, and you've

got Elon. But then you've got the two

little startup, you know, super hyper

creative startup guys. So that's Daario

and Sam. And and so, you know, Sam,

OpenAI,

uh, is playing a very different game

from Daario. Daario is relying on

exactly what you just said. I'm going to

build a more intelligent fundamental

machine, and because it's more

intelligent, people will navigate to it

and we'll go out through corporate

channels. Then Sam is playing the old

Bill Gates game where like I'm not going

to take for granted that my AI is better

than Google's. But right now I have

twice as big an installed base as Google

does. So what can I add to protect my

position that makes me the default

choice in the case where the two AIs are

on rough par. So he gets Johnny IV to

build a device. Uh he's building his own

data centers with Broadcom and now he's

adding a browser. And so he'll add

everything that Bill Gates would have

added that's a user point of control or

an entry point into the use of AI in

order to defend that turf and encourage

more of the innovation to come through

him rather than work around him through

Google. But it's like all that warfare

all of a sudden between Google and

OpenAI and it's just really fun to

watch.

>> Dave and Alex, you know, my favorite

model for this is still Jarvis from Iron

Man, right? We're going to have an AI

that is our personal uh compatriate and

and our portal into everything. And I'm

not going to care what browser I use.

I'm just going to be able to have a

conversation with my AI and it will pull

up the data from wherever it is, whether

it's using super intelligence from, you

know, OpenAI or Google.

>> Well, the only thing I'd add to that is

that that very very soon that data will

be your personal health data, your

personal preferences, your everything

about yourself. So, and when you have

your your virtual girlfriend or

boyfriend, everything you like and don't

like in life will be in there. So, if

Sam wins the data aggregation race, if

he falls behind for a month or a year in

the AI race, he still has your data. And

that personalization might create a much

more compelling experience, allow him to

catch up again. Uh, so, you know, the

the personal data warfare is like

kicking off in a huge way right now. You

mentioned it a second ago, Peter.

>> What's what's the downside of what we

see here with Atlas? I mean, we have the

ability of OpenAI to not only look at

the data you have on your browser, but

probably every tab that you have open

and everything you have going on in your

computer. And they're not promising to

keep it confidential. Um, thoughts on

that Alex?

>> I I I I think we'll see forcing

functions for for greater forms of

confidentiality and privacy, but I'm

just reminded, do you remember the

browser wars?

>> Yeah, of course. Right. And Google

Google one with 70% market share today.

>> Yeah.

>> Right. So, so there's sort of a long

history of sleepy periods of relatively

low innovation separated by Cambrian

explosions of functionality. I I I

remember all of the browser wars and I I

think a browser war today over competing

among other factors on whose browser is

most private while also being AI

agentic. I I think that's a valid front

for competition and I welcome the

competition.

>> Amazing. Uh Alex, would you introduce

this next uh this next slide here? You

built ch a chess game. But before I play

it, explain what you built here.

>> Yeah. So, uh with with computer use

assistance, uh CUAS of which arguably

this this new chat GPT atlas agent mode

is is one example. I I have my own

emails. Um, one of my my favorite evals

for for testing these CUAs is to see

whether they can win at simple uh and or

complicated single player web games. So,

uh favorite easy example is is to see

whether I turn Atlas loose on a single

player, not not double player, single

player game of web chess and see whether

it can win. Uh, I've I've used this eval

against uh historically operator from

OpenAI. What we're seeing here is is a

time lapse uh of it just being asked I I

turned it loose on on a web chess single

player asked it to win. Uh, and

interestingly, this is the best

performance I've seen to date from a a

web-based CUA turned loose on just

sometimes I'll turn it loose on a game

of web civilization if folks are

familiar with the Civilization

franchise. But in in this case,

intriguingly, it asked for hints, which

I've never seen before. Uh, so it used

it sort of used the helpline built into

the the web game to ask for hints and

was winning at the end of the day. I I

think this is a preview. In short,

>> did it ask you for hints or did it ask

some other?

>> It asked the website for hints once it

discovered, which it did pretty quickly

that it could ask for hints. It asked

for hints and use that to win the game.

And and I I think this is a preview of

CUAS for everything, not just winning

easy games of chess.

>> Amazing. Amazing. All right, I'm going

to jump into anthropic. And this is a

conversation between Jonah Cool, who's

the head of life science partnership and

development, and Eric uh CH Cower

Abrams, who's the head of biology and

life sciences research. You know, in

January at the World Economic Forum, uh,

we heard Dario Amade, the CEO of

Anthropic, talk about one of his

passions, which is the ability of AI to

accelerate biology and longevity. And

very famously, he said, you know, if

we're able to hit the targets we have

for AI, we could see the doubling of the

human lifespan the next 5 to 10 years,

which perked everybody's ears up,

including mine. uh you know are we going

to see longevity escape velocity within

this decade? Uh increasingly the answer

is yes. Let's take a listen to uh to

Jonah and Eric have this conversation.

>> I'll start with why are we focused on

the life sciences when we talk about the

beneficial use cases of AI and all the

amazing things that we can do in the

world with the frontier AI that we're

developing. Actually the number one

place that we at anthropic are excited

about applying it is within biology in

the life sciences. Right? If you read

our foundational material that's the

primary area where we're really focused

on on delivering um the the beneficial

impact. We need claude to be conversant

with all of the tools that scientists

are using every day. Right? And so

there's a whole ecosystem of important

tools and partners out there that we are

integrating with. Right? So we talk

about benchling on the you know

experiment administration lab notebook

side of things. TEDx Genomics with Cell

Ranger, right? Incredibly important

platform for um analyzing single cell

experiments and then PubMed, for

example, for being able to query the

literature, right? And so these are just

three of a three incredibly important

partners in a much larger ecosystem. And

so that that base level is we need to um

make sure that cloud can can talk to all

the major sources that scientists are

using throughout, you know, their their

daily. We want to bring Claude to

performing at the level of a superhuman

research assistant that can assist you

as as a scientist throughout all stages

of your project. Alex,

>> I I I speak from time to time on this

pod about super intelligence solving

math, science, engineering, medicine. I

I think this is likely how biology gets

solved. I I think I was talking a moment

ago about computer use assistance, CUAS.

I I think we're entering the era of CUAS

for biology where we have baby super

intelligences that are completely fluent

and and well-versed in the tools of

computational biology and are able to

read PubMed fluently and then go and

perform experiments even. I I think this

is what solving biology with AI looks

like.

>> Yeah. You know, there's a company I just

recently invested in that I'm very

excited about. It's called LIA, L I L A.

People can look it up. It's out of MIT

and Harvard. Uh George Church is the

chief scientist. Uh uh Jeffrey Vmolson

is the CEO. And what they're doing in a

similar fashion, but I think more

advanced is they've set up these uh

science data factories, right? So they

have a a super intelligence model

they're building and these these science

data factories are basically 24/7 lights

out robotic uh you know robotic farms

looking for information out of nature.

So if you imagine the super intelligence

will come up with a scientific theorem

or you know a proposed research. They'll

program the robots to go do the research

at night, gather the data, bring it

back, check their their theory, iterate,

put the next experiment forward and

running on this 24/7 cycle to sort of

mine data out of science uh itself and

focusing on biology first and foremost,

but chemistry and material sciences. And

I love this as we're searching for new

data out there in the world to help us

understand what's going on in our 40

billion cells. You know, C, you know,

it's 5 to 10 chemical 10 5 to 10 billion

chemical reactions per second per cell.

Uh, we need to we need to be able to

reach in and get the data out to build

our models even better.

>> I I think that it as as as I I think

Peter, you might know, Jeff was a

labmate of mine when we were undergrads

at MIT. And I'm a huge evangelist for

dark labs. I I would like to see dark

labs for everything.

>> Yeah.

>> Well, and Jeffrey von Maltson, I know

it's a harder name to find on the

internet than Jonas Cool or Jonah Cool,

but uh but definitely look him up. The

guy is going to be huge. Uh you you

know, you can see it coming out and Alex

will reaffirm this, but he will be one

of the key figures cracking life

sciences. And and I'll tell you what

else. um you know we'll see later in the

pod

there there are some people saying look

we got to slow down AI we got to stop

it's not going to actually happen we're

going to move full throttle and there

are two reasons one is China the other

one is this people are not going to sit

and let people die unnecessarily from

illnesses if AI can discover solutions

to that that's not going to happen

>> so that's why the AI labs are talking

about this use case so much because it's

it's life it's preserving lives

>> yeah And uh by the way uh Jeff uh

Jeffrey Van Molton and uh and Laya will

be at the Abundance Summit. Uh super

excited for him to present our theme in

March of 26 at the summit is super

intelligence and the rise of uh humanoid

robots. So he said okay that's

definitely a subject I want to cover.

All right let's move on. Uh Wikipedia

says uh human traffic has been dropping

down 8% yearon year. less humans are

coming to Wikipedia. Uh, we can dive

into this. I'm still waiting for

Guacipedia to come online.

>> Alex, what are your thoughts here?

>> Yeah, I get asked the question a lot.

How do we incentivize humans to create

new knowledge in an era of generative

AI? And I I I suspect the question

itself is is probably faulty. I I think

knowledge gathering is likely itself to

transition to AI. I think we'll see

investigative reporting that's AI based.

So I I I I'm not losing sleep over human

traffic dropping in an era when

knowledge synthesis is abundant, but

knowledge generation by AI is not yet

abundant. I I think AI generated

knowledge is right around the corner.

>> You know, I have a Okay, Dave, I'm going

to go ahead and then I have a rant on

this.

>> All right. Well, this is this is right

in my wheelhouse, so I need to wax

poetic for a minute on this topic. So,

you know, I've been the founder of 20

direct to consumer AI companies. First

and foremost, every time someone

complains about their traffic going

down, it's going somewhere else. It's

not going away. Traffic overall traffic

is going up very very quickly. And so,

you know, I'm involved in a company I

can't name right now that's gone from

from nothing to 600 million of revenue

purely from online arrivals, 100 million

of profit on the bottom line. And so,

when when Wikipedia says, "Hey, traffic

is going down, it's going to some other

place." And the formula for getting the

traffic is is well known now. You know,

first and foremost, you need to create

huge amounts of AI generated content,

but it has to be good content, but you

also have to pay the man. You got to pay

Google. You got to pay Facebook. And if

you do that concurrently with putting

your your content out there, then

they'll give you the traffic. Also, you

need to reformat your content so it's

it's easily readable and interpretable

by the AI, hence GEO at the bottom of

the slide, generative engine

optimization. Because in the future, you

know, people do not go to Wikipedia for

their content. They just ask the AI. The

AI's got all the information, but it

still needs to be factually accurate and

correct. And so that that role, and I'm

a big Wikipedia fan, but you know, I was

at the Washington Post when it was

getting obliterated by the internet, and

it felt like, hey, we're we're important

for the country. We're we're factual. It

doesn't matter. You're going away. And

so that's what's happening. My my my

rant on this, you know, I've been trying

to update my Wikipedia page for

literally two years. I hired consultants

to update my Wikipedia page and every

time it's updated, they bring it back to

what it was. It's like so stuck 20 years

ago. And, you know, I don't know. I I

used to use Wikipedia. I don't anymore.

and the ability for an AI to actually

search the web and get consistent and

relevant and accurate information about

me. So I think maybe Graedia will be a

solution here or in fact any AI that

just says you know spin up a page on

Dave London that can send somebody u

that's going to be awesome. I

>> I'll give you one other you know pro

tip. Get a get a similar web account

similar.com get a similar web account

and you can see exactly where that user

went. the the guy that would have gone

to Wikipedia yesterday, where did he go

instead today. And so then if you track

where it's all moving, replicate that

behavior and you'll succeed.

>> Amazing. All right. Uh, next article

here is GPT5 rediscovers longforgotten

math connections. Uh, this has Alex

Wizzer Gross written all over it. Dr.

Gross, please tell us. Uh Peter, I I I

talk frequently about how super

intelligence is and will be solving

math science engineering medicine

other fields. There was a lot of hand

ringing o over the past week plus about

a specific set of math problems and

whether AI in general and GPT5

specifically was actually uncovering new

math. And I I think this this story sort

of beautifully encapsulates the fog of

war we're in right now. the the level

the water level of intelligence is

rising day by day and some of the

earliest math problems open math

problems to to be solved are I think

will will be math problems that where

the solutions were known to a subset of

humanity but not to all of humanity like

and and and we're going to ring our

hands collectively as a civilization

quite a bit over well was this open

problem in math really open or was it

solved or was it half open where some

people knew how to solve it and other

people didn't know that it had even been

solved. That that's the fog of war phase

that we're in. So I there was a lot of

discussion over the past week like was

this a real accomplishment, a real

discovery in math by AI, one of the the

Erdish problems um number 143 for

example. But there was I I think

ultimately a lot of really revealing

discussion and and commentary on on this

particular problem and also other Erdish

problems that actually this is just a

phase right now like early days we're

we're still cleaning up house as it were

in terms of understanding even which

problems are open closed or somewhere in

between and after this phase I I predict

we'll get to a phase where a lot of the

uncertainty is reduced regarding whether

a given problem is actually open or not.

>> Yeah. I think

>> open means solved, right? You mean

solved.

>> Open means unsolved. Closed means

solved.

>> I think this is also a great little case

study and how the academia world is

like, well, this proves that it didn't

really solve it. It looked up an ancient

like when you're trying to do something,

you don't care a wit how it solved it.

It came back with the right answer. This

is a lot like uh you know uh ThinkStruct

in our lab. you know it's a company that

does academic research and now patent

research using AI. So Nikki Abate and

Julius Hidekutter and it is actually

turning out to be a really good hybrid

of writing your patent application while

doing all the background research for

all prior applications and all prior

knowledge. And so those two things are

are integrated. And this is where you're

seeing AI being superhuman because

normally you'd say, "Oh, well research

of old documents is this guy, but

thinking of new things is this other

guy.

>> The AI doesn't care. It just does both."

>> I'm so excit so excited about the use of

AI in in writing up uh and submitting

patents and talk about something that is

extraordinary. But one of my most one of

my favorite applications of AIS and

patents were the following. Uh this was

a conversation with an abundance member

who was like, you know, I want to figure

out how to use these technologies on my

business. I said, well, why don't you

just ask? And so what I what I showed

her said, okay, here's, you know, here

are three patents you're interested in.

Uh put them in the browser and say, this

is my business. How would I combine

these three patents together to make a

new product or service in my business?

And oh my god, it's extraordinary,

right? This is literally a creative

engine.

>> All right.

>> Well, anyone who's a real fan of this

podcast by now has to have read

Accelerando because Alex Wisner Gross

says it's the best piece of writing in

the history of humanity. If you if you

heard that and then didn't read it,

something's wrong with you. But the very

first chapter, the opening scene is

exactly what we're talking about right

now. the the lead character makes a

living with with AI generated patent

filing.

>> Yes. Consistently and gives it away.

Anyway, let's not go there. All right.

Our next article here is Uber tests

microwork for drivers to train AI. So

Uber is paying between 50 cents to a

dollar per task that can take two to

three minutes uh and get processed

within 24 hours. So is this sort of a

digital task rabbit? What is this Dave?

Uh this is uh this is really really cool

because you know Meror is is almost you

know closing in on a billion of revenue

going all over the world g grabbing

expertise and getting it into a format

where the AI can assimilate it and then

the AI can be an expert in that topic

too. Well you got all these Uber drivers

driving around. They're sitting sitting

around a lot of the time. Do they have

knowledge that may be a contributor back

into the great AI machine? you know,

because a lot of what's missing is

physical motion, common sense, you know,

just all this information. So, you know,

why not use that same platform you've

already got to be another another

Merkore type AI data gathering machine?

>> Alex, your thoughts on this?

>> Yeah, I think this points directionally

to the future of the gig economy. The

gig economy historically was focused on

the physical world, physical tasks,

inclusive of driving other people uh to

their locations or or driving food to a

person's location. I I think this points

toward a near future where training

robots to perform service economy tasks

is the new deacto gig economy.

>> Yeah. So fascinating that Uber turned

this way. it, you know, it's all about

the relationships it has, right? It has

a relationship with a large number of

people that it knows wants to earn money

on the margin. Uh, and we'll probably

see other companies follow suit as well.

>> Well, you know, during co, you know, got

annihilated and Uber did fine because

they had launched Uber Eats. Uh so

they're you know they're very very

thoughtful about this you know in fact

when when I don't know if you remember

Travis Colick when Uber was going public

but he got on stage and he said Uber is

not a ride sharing hailing cab company

we're a internet fabric it was some some

like really ethereal but now they're

actually doing it it makes sense in in

hindsight so they don't view their

platform as being about cars and rides

they view it about like

>> we're going to spend time with Dra the

CEO of Uber he's going to

on stage with us at the abundance

relationship with DAR. We'll talk about

what he's doing in the data side, but

also you know they're now partnered with

Whimo. Uh you can in certain places hire

a Whimo through Uber and they're you

know they're hooking up I think with

Joby on the you know flying cars let's

call it that for the moment. So Uber's

been an incredible platform for

experimentation and sort of integration

of various exponential technologies.

>> So that'll be fun.

>> All right,

>> Alex, I'm going to turn to you on this

one. Deepseek is packing text into

images. Uh talk about this, pal. What's

this significant uh transformation,

isn't it?

>> Yeah, this is a major advance from from

Deepseek. So a new model that Deepseek

announced, Deepseek OCR. Uh so maybe a

bit of background first. Foundation

models, frontier models like GPT aren't

thought to perceive text in the way that

humans perceive text. Humans look at

text on a page and we see text visually.

The frontier models, the foundation

models, most of them are are believed to

to still consume text in the form of

chunks of letters called tokens. and

they don't perceive have any based on

publicly available information any

visual perception of letters on a page.

So they don't visually see the shape of

a character or formatting or desktop

publishing type layout on a page. They

perceive none of that. They perceive at

best maybe like HTML formatting

instructions. So I I think DeepSk OCR,

which is I again if you squint at at the

the model architecture, it's it's sort

of an autoenccoder that that does

optical character recognition after a

fashion, but in a really in a really

interesting way. It it consumes raw

images of entire pages and encodes those

as image tokens, not as text tokens, and

then tries to decode those image tokens

into text tokens. So a few things fall

out of this. One, optical character

recognition at at high accuracy rates,

which is pretty incredible. But

secondly, this is able to perceive

formatting at the way humans do. And I I

I think the the practical upshot of this

would be better grounding. Like wouldn't

it be wonderful if we could have desktop

publishing type formatting of outputs

from from Frontier models with beautiful

layouts? I I would expect that to fall

out for free or better understanding of

mathematical equations that are

dependent on the way the equations are

written and how they appear visually. I

think better understanding of fonts, all

of these I expect eventually to fall out

of this line of research.

>> Interesting. And we're going to be

seeing an article later about Amazon

getting into the AR uh glass

marketplace. And we're going to see from

Meta and Google and probably Open AI and

all of them were transforming, you know,

from a from a phone as the medium of

interface to glasses at this medium

interface. So, uh I'm assuming that this

kind of technology is going to help your

your glass effectively translate

everything you're seeing into something

it can be uh understood, read, and uh

responded to. I I think that that's

table stakes. Uh so yes to that, but

also having AI that understands at a

visual level all the text, I I think

that is is going to be quite

transformative.

>> Dave, you want to comment on this?

>> Well, I'm still blown away that when I'm

if I'm writing code in cursor or winerf

and I take a screenshot and say, "Hey,

there's a bug in here somewhere. It's an

image. It's not text." And I just slap

that right back into cursor. it it has

no problem with it at all. Now I know

under the covers it's not doing this.

It's actually converting it to text and

then moving forward from there. So this

will put the AI engine much more in tune

with human with human thinking because

you're using the same exact pixel by

pixel interface that we use with our

eyeballs for everything whether it's

text or images or whatever. So it'll

it'll be a big advantage in multimodal.

But what works already is just

mind-blowing to me. Do you expect, Alex,

that we're going to see this type of OCR

uh come into all the models next?

>> Yeah, I I think we're moving towards a

near future with universal tokens, that

tokens that span modalities. And I I I

I've long thought, wouldn't it be

wonderful aspirationally if if we had

just a single modality that everything

else flowed through? So rather than

having a text modality and images and

audio and video, if we just had maybe

like a single universal maybe video

style modality that everything else

flowed through, it it might have certain

benefits. You know what's really

interesting about that, Alex, is that

that that's happening and it puts the

models much more in touch with humans

and at the same time it's going the

other direction in in very specific

domains like you know magnetic bottles

and and you know quantum computing where

the knowledge is so far out of the human

domain that you want completely

different data representations at the

front end of the funnel. And so these

these first models are going to use the

second models as tools. It's It's really

cool to watch the two kind of spread

apart and and think about how they're

going to end up interacting.

>> This episode is brought to you by

Blitzy, autonomous software development

with infinite code context. Blitzy uses

thousands of specialized AI agents that

think for hours to understand enterprise

scale code bases with millions of lines

of code. Engineers start every

development sprint with the Blitzy

platform, bringing in their development

requirements. The Blitzy platform

provides a plan, then generates and

pre-ompiles code for each task. Blitzy

delivers 80% or more of the development

work autonomously while providing a

guide for the final 20% of human

development work required to complete

the sprint. Enterprises are achieving a

5x engineering velocity increase when

incorporating Blitzy as their preIDE

development tool, pairing it with their

coding co-pilot of choice to bring an AI

native SDLC into their org. Ready to 5x

your engineering velocity? Visit

blitzy.com to schedule a demo and start

building with Blitzy today.

>> All right, let's jump into our next

article here. This is OpenAI hires

bankers to automate junior work. So

OpenAI has hired over 100 bankers paying

them 150 bucks an hour to train AIS on

M&A, LBOs, IPOs. Effectively, uh these

bankers are in one sort of the way

traders helping to eliminate jobs of

fellow bankers. So I don't know my first

take on this is open AI is basically

eliminating what I would have imagined

an entrepreneurial startup would do. I

imagine lots of startups are looking to

do this and this is sort of a shot over

the bow. Well, you know is open a going

to do this for every field. Um you know

get rid of white collar work across the

board. Um and well the answer to that is

absolutely yes. They're going to do this

for every field and they're going to do

it quickly. And I just had this

conversation with two of our companies.

Make sure that you're the guy that Sam

calls and they're they're like, "Well,

Sam's not going to Sam's not going to

call me." Like, why would Sam call me?

They said, "Okay, eliminate everybody

else. He's not going to call State

Street Bank. He's not going to call like

he doesn't want to talk to big legacy

bloated entities." All right? And he's

also not likely to call two 22 year olds

out of Y Cominator who haven't even

gotten to market yet. So if you're

somewhere between those two things,

you're you're just need to position

yourself like Brendan Foody did at

Merkore, get into the building and be

the person that solves that problem for

open AI. But yes is the answer. He'll do

this in absolutely every category of

human endeavor.

>> So I would imagine this kind of a

vertical would be something an

entrepreneur would say, okay, we're

going to go and do this ourselves for

whatever field. But uh again, we talked

about this when we were up with Kevin

Kevin Will that is open AI going to be

moving into and eliminating all the

entrepreneurial vertical channels? Um I

find this fascinating.

>> Yeah, I go ahead.

>> I I I think that the sort of the

superficial story is that this is what

the end of so-called white collar work

looks like. vertical by vertical, labor

category by labor category. Each

existing form of the service economy,

each manifestation of it gets digested

and and turned into AI automation. But I

I think that creates enormous

entrepreneurial opportunities for for

everyone. There are thousands if not

tens of thousands of labor categories

with domain specific knowledge that will

require automation. Same with industry

subverticals. And every platform company

always instills maybe a modicum of fear

in other companies. Oh well, the

platform will just absorb what I'm I'm

doing and we'll we'll lose our footing.

But I I don't think that's an accurate

representation of of the real economy

where there is just tens of trillions of

dollars of service economy labor that

can be automated. And I I do not expect

a singleton scenario where any one

company or any one platform or any one

model just consumes the entire economy.

We'll have I think a completely

heterogeneous economy indefinitely into

the future.

>> So bottom line here is that this effort

by OpenAI could eliminate between a

quarter and half of the junior headcount

across Wall Street within two years.

>> All right. Uh moving on. Uh I love this

article. So Google is prepping Genie 3

for public experiments. So uh Genie 3 is

going to let users create interactive

worlds with text prompts. We talked

about this um extraordinarily powerful.

This is this is persistent and uh

consistent worlds that are generated

from a text prompt that are

photorealistic

that uh you can get into and you can use

for a variety of different uh of

different areas. Um, Alex, do you want

to jump in? Well,

>> first of all, you have to admire that

the user interface, which we're now

seeing previews of, looks identical

almost to the grid of the Holodc in Star

Trek.

>> Yes, I love that.

>> Have to admire that that we're catching

up with the future. It's very exciting.

A level deeper though, I I I do think

world models, so-called right now, are

going to merge with the foundation

models. I I think this is very likely to

be an instrumental element of general

purpose generalist foundation models and

frontier models that you'll not just be

able to have textbased conversations

with them or audio-based conversations.

They'll create entire worlds that you'll

be able to walk around on the one hand.

That's the consumer use case. And the

enterprise use case is these world

models that are fully interactive will

enable us to create new inventions,

create new products. it this is the mode

through which AI understands and will

understand the physical world and be

able to create economically

transformative inventions. It's the

democratization of interactive content

creation, right? Uh at a level of

reality and resolution that is shocking.

Just to hit on on some of the ideas,

right, for the individual, if you're

thinking about this, how would I use it?

Uh you can build personalized gaming,

it's creative storytelling, it's

customized education. Uh for companies,

I think a lot of companies could be

using this for game development, uh for

education tech. I I personally think the

most extraordinary way to educate and

learn about something is to dive into

that world. I've used this example so

many times. If you want to learn Greek

history, you can read a very dry

textbook. You can even watch a movie.

But imagine being able to drop into

ancient Greece. You see a guy in a toga

on a chunk of marble and you walk over

and he says, "Hey, I'm Socrates. Let's

go for a walk. Let me show you around.

Meet my friends."

That kind of immersive experience is the

future of education uh without for me

any question at all. Dave,

>> well, just not to disappoint everybody,

but this is going to be another one of

those things that everybody instantly

loves. uh just like deep research. If

you've tried using Gemini deep research

lately, you'll sit there for like 10 15

minutes unnecessarily and then it'll

give you something great back, but it's

just enough to frustrate the hell out of

you. It's entirely

>> more GPUs.

>> More GPUs, man. Keep tiling cuz people

are going to love it. It's incredible.

And unless you buy your own Nvidia box,

you're not going to be able to get the

speed you want. All right, our next

article here is Meta borrows $27 billion

to build an AI data center. So Meta SPV

is borrowing this money at 6.8% to fund

a multi-gawatt le Louisiana data center.

Dave, you had some thoughts on this one.

>> Yeah, absolutely. So, so Mark Zuckerberg

has taken every penny of cash flow uh

from one of the biggest tech companies

in the planet, you know, from Meta,

Facebook, and pumped it all into this AI

initiative uh and now is borrowing, you

know, going to the next level. And the

market, stock market loves it. So, what

does that tell you as a CEO? Like, if if

you're a true AI company on a and on a

true AI mission, you can invest like

crazy from your public capital or from

your from your venture capitalists. uh

and they love it because they see the

the future is is here. But it's I don't

think it's probably unprecedented in

history for a company that used to be an

absolute bottomline cash cow producing

huge amounts of EBIT to take every penny

of it and then more then borrow even

more to pump it into an initiative.

>> I mean Mark has said over and over again

he will do whatever he needs to get to

digital super intelligence first.

>> It's like his is his war cry. Alex,

>> I I I think it's also worth adding the

credit markets are just as interested in

financing this project, call it tiling

the earth with compute, as the equity

markets. Uh, and the the fixed

income/credit/

debt markets are enormous. And I I I

think we're we're starting to see this

financial model where the lower half of

the AI infra stack is being funded by

credit as we're seeing here. And then

the the upper half where the the models

and the applications live is being

funded by equity. So so we're we're

seeing sort of a whole of economy

financing model emerge for this full AI

infrastruct.

>> It's you know the implications of this

though is capital from public equities

from sovereigns from debt all flowing

into AI at the exclusion of so many

other technology areas. Well, one of our

best partners, uh, Kush Pavaria,

phenomenal guy, uh, just co-founded a

company called ORN.

Uh, yeah, check it out. But my point in

in this is that people that are kind of

technical and engineering wouldn't

normally get into the finance side of

things. But, but, you know, kind of like

Chase Lock Miller, they're getting drawn

into this in a big way. And it's a very

very good strategic move. If you if you

have any interest in finance whatsoever

and you understand GPUs, chips, data

centers or just math, uh it's a great

direction to go. It's just a huge amount

of capital, you know, redirecting into

this direction and and and you know,

making it move intelligently, the right

investments, the right locations, that's

not a trivial problem at all. So, if you

have an engineering mindset and you're

interested in this area, you can really

do well.

>> Yeah. Amazing. Okay. On the data center

world, here's the uh news from Oracle.

Oracle is planning a 16 Zetaflop AI

supercomputer. We don't talk about

Zetaflops all that often. So, it's

announced a nextgen cloud computer

designed scaling to 800,000 GPUs. You

know, is Zetaflop here? Is Zetaflop

there? All right, I'm going to feed Alex

on this one. I can't wait to hear what

he has to say. But I remember on that

podcast a couple months ago, we were

talking about uh the 10 E26 models. So

they, you know, E26, that's that's the

regulatory definition of a AGI super

intelligent register it with the

government type thing. So that's one

E26.

>> Uh, a Zaflop is what 1E21.

So that's per second though. That's, you

know, that's that many flops per second.

>> So we're talking about an exponent of

>> Yes. Go ahead.

>> Yeah. So So to get from, you know, 10 16

to 10 21, you need you need five more.

So that's 100,000. So every 100,000

>> those are orders of magnitude just to

translate

>> five more. Five more. So 100,000x. So so

every 100,000 seconds a one zetaflop

computer can create a foundation

frontier level AI model every 100,000

seconds. So 100,000 seconds is 1.1 days

as it turns out. So every day you get a

new foundation and that's at one zoflap.

This is 16 zeta flops. So 16 times a day

you build a foundation frontier level

model. Does that sound right, Alex? Did

I get any of that wrong? I'm doing

>> I need to double check, but it sounds

approximately right.

>> I I I I would add. So uh Oracle is uh

this is all in the public reporting.

Oracle is both financing and operating

Stargate Abolene. And Stargate Abene is

I I think uh together with this 16

Zetaflop super cluster. It is emblematic

of a new form factor for computing. The

the personal computer was a major new

form factor. The smartphone was arguably

a major new form factor. These

superclusters with approximately a

million GPUs and tens of zeta flops.

This is a fundamentally new form factor

for computing with high-speed

interconnect which we're not talking

about but which is arguably just as

important as the raw compute power being

a key architectural innovation and it's

not going to stop with Stargate

Appaline. Th this form factor again in

in the spirit of tiling the earth with

compute we are unless something radical

changes we are going to tile the earth

and and maybe near-earth solar system

with this type of new form factor of

computer.

>> Incredible. All right. Uh continuing on

this conversation anthropic uh to expand

to 1 million TPUs on Google cloud. So

their goal is to bring this compute

online by 2026.

Uh there's a I think there's a a very

loving relationship between Google and

Anthropic. Anthropic is sort of the

little brother there and they're growing

closer and closer. Alex, what do you

make of this?

>> Well, I I I I I want to say something a

little bit glib per perhaps, which is

that when you have when you have super

intelligence that's incredibly thirsty

for compute, it makes for some

interesting combinations in in the

market. I I think the the the thirst for

compute is is creating enormous pressure

on the frontier labs to diversify their

infrastruct. So we're seeing Nvidia GPUs

up against Google TPUs up against Amazon

traniums up against AS6 including

Frontier Lab specific AS6. I think these

are all in the mix. So for for those who

are worried about some sort of

architectural monopoly or singleton

where only one GPU or accelerated

compute architecture completely

dominates the market. I I think this is

a healthful dose of both diversity and

reality that now actually we're seeing

heterogeneous architectural combinations

at multiple levels of the stack. The the

future light cone of compute

architectures is not going to be

dominated by any single company. Alex,

for those who don't know the difference

between TPUs and GPUs, would you uh give

us a 101 here?

>> GPUs, this is branding that was

popularized by Nvidia. So, graphics

processing unit. This was originally

conceived and went to market for

accelerating video games where Nvidia

was the arguably the the chief actor for

accelerating compute specifically for

video game purposes and professional

graphics as well. Then eventually it

found its way to Bitcoin and other

crypto minings. And then fortunately the

need and the thirst for for accelerated

compute for AI arrived just in time to

uh to sort of recover from a bit of a

mini crypto winter and step in.

Meanwhile, TPUs, tensor processing

units, this is a a term from Google, but

the the underlying architecture is

pretty similar to uh to the way GPUs

from Nvidia and other firms handle AI

operations. the the t the tensor refers

is is a reference to this idea that the

central operation that they need to

perform in support of AI and machine

learning is taking large matrices which

if generalized become tensors sort of

highdimensional matri matrices of of

numbers and multiplying and adding them.

Uh so that that's sort of the to to

oversimplify that that is the core

operation of accelerated compute for

machine learning just taking matrices of

numbers and multiplying them.

>> That's a great point.

>> Appreciate that. Uh we talked about this

on the podcast we recorded at

Visionering uh a few days ago. Hopefully

you enjoyed that episode. But I wanted

to bring Alex and Dave into the

conversation here. Uh this is StarCloud

bringing data centers to space. I'm

going to play a short video from Philip

Johnston. Actually, Philip, who's the

co-founder and CEO of StarCloud Cloud,

was here with me for the last few days.

So, it was fun to see uh his points of

view. Let's play the video and we'll

talk about it afterwards.

>> The reason we're building data centers

in space is mainly for the energy that

we can draw from solar energy in space.

So, there's almost unlimited access to

abundant solar energy in space. The

problem on Earth is we're very quickly

running out of space and actually energy

on Earth to build large data centers. In

space, we can have these enormous solar

panels um which can power these data

centers. And then another advantage is

we can then run large radiators to

dissipate that heat and infrared out

into the the vacuum of space. So, it's

interesting. Uh, Philip Johnson was on

stage pitching, uh, a prize called, uh,

you know, the spa, uh, let's see, the

space cool X-P prize. It was something

like that. And basically, one of the

challenges they still have is radiative

cooling. Uh, space is very cold, but

there's no there's very little uh, you

know uh

atoms to carry the heat away. So, you're

focused on infrared radiative cooling,

which is a challenge. So, I'm so

curious, Alex, what do you make of this?

Is this the future, or is this something

that isn't going to happen?

>> Well, I I think at at the heart of this

is what I would argue is one of the most

important civilizational questions that

we face. We don't know the answer, but

the the question is, does a mature

intelligent civilization build a Dyson

swarm or not? Dyson swarm. swarm,

meaning taking apart the planets in our

solar system to to build lots of

computers that orbit the sun. I don't

know the answer. I I suspect the answer

will depend on physics discoveries that

haven't happened yet. And just jumping

out a few decades, play playing this

tape forward as it were, playing the

recording forward. I I think if humanity

ends up being permanently latency

constrained, we're probably going to do

it. that this this probably then is the

beginning of the construction of a Dyson

swarm. On the other hand, if if physics

make it ergonomic to easily travel to

other star systems, presumably with

physics that that we're not aware of

yet, then I I could imagine scenarios

where actually building a Dyson swarm,

you know, turning StarCloud and and and

other orbital computing platforms into a

fullon Dyson swarm probably doesn't make

that much sense. One could also imagine

other contingencies. Maybe the demand as

unconscionable as it is right now. That

demand for accelerated compute might

peak at some point in the future if that

ever happens. I could also imagine we

don't build the Dyson swarm. Otherwise,

I I think just straight shot uh this is

the beginning of a a long-term trend.

Mark this point in time. We're at the

beginning unless something changes of

the construction of a Dyson swarm.

>> Yeah. Just to clue folks in uh Dyson's

form of the terminology comes from Dr.

Freeman Dyson who was at Institute for

Advanced Studies at Princeton who

basically said as you become an advanced

civilization you're going to want to

capture all of the energy coming out of

your star. So you'll dismantle your

solar system and you'll basically build

a shell around the star that captures

all of it. Um this is the earliest days.

So, you know, I just want to point out,

I had this conversation with Philip. You

know, we have 8,000 times more energy

that hits the surface of the Earth today

than we consume as a species. And the

challenge is, can we build the square

meterage of solar and uh and dissipation

arrays in space? You know, there's going

to be a lot of robotics required to do

that. And when do we get there? um you

know is it 10 years from now 20 years

from now uh we're going to find out.

Along these lines uh we saw Caruso uh

you know basically announce that they

plan to support this by 2027. Uh and I'm

not exactly sure what they mean by

supporting it. They're going to put an

H100 uh up in space and H100 in space

represents 100 times more compute than

any other satellite has had. But it's

it's a single H100. It's not a uh uh

it's not a cloud, not a Cruso cloud.

Alex, did you dig into this further?

>> Yeah, I I would maybe also just comment

on time scales. So putting a single H100

in low Earth orbit or LEO may not sound

like that much now, but I if if you just

starting from from physics, like if if

we have this notion that we know or at

least have a prediction that the end

state of all of this is taking apart our

solar system, you could actually just do

a few calculations to figure out the

time scale for when that would happen.

So, one of my favorite statistics, if

you ask like if if we could completely

encircle the sun with solar collectors,

capture all of its luminosity and

channel all of that that power to say

unbinding Jupiter. Basically

disassembling Jupiter. Jupiter's created

it its own gravity well. So, uh so we

the the term of art would be unbinding

it from its own gravity. it would only

take approximately two centuries if we c

captured all the the light from the sun

to disassemble or unbind Jupiter. So I I

I view you know one H100 going into

space in in the next couple of years.

This is the the first step in a

potentially a a two century journey to

to deploy compute at scale in our solar

system. And I think

>> exponential growth double something 30

times you get a billionfold increase.

Dave, what are your thoughts on this?

>> Well, Peter, you said, you know, that

the sunlight hitting the Earth every day

is 8,000 times more energy than we

consume. But have you ever done the math

on the fraction of all the sun's energy

that hits the Earth in the first place?

>> Oh, yeah. It's it's it's a you know, far

far less. It's a fraction of 1%.

>> Yeah, I know. I don't know how many

decimal points are in there, but it's

like there's a monster amount of energy

in that that Dyson sphere, Dyson swarm

view. Uh so yeah it's it's you know 200

years sure why not. Um what's

interesting in the short term this could

be a great idea or a terrible idea for

Cruso and it depends entirely on the

timeline diffusion which we're about to

talk about.

>> Uh so so that's an interesting factor in

all this.

>> It's worth pointing out while the term

StarCloud sounds like it's got Musk

behind it uh Elon is not involved in

this. uh he did retweet uh the StarCloud

announcement, but uh you know I I love

Elon. He's incredibly brilliant, but at

the end of the day, if he were to take

this on, he would probably do it on his

own. Uh that's my experience. All right.

Uh moving forward. Uh okay, now on to

Elon here. So Elon says the A15 chip uh

by some metrics will be 40 times better

than A14. We deleted the legacy GPU.

It's basically a GPU. I poured so much

life energy into this personally. It'll

be a real winner. So, you know, we've

seen this before where Elon goes heads

down and focuses on a very specific

element, you know, all the way down to

the engineers, scientists, the

production line.

Alex, you've been tracking this. What

does the A15 mean for, you know, for

Tesla, for Optimus, for XAI? the the uh

so I've spoken in in the pod on the pod

in the past about this notion that super

intelligence is not going to stay just

bottled up in the data centers. It is I

I've argued in past it is literally

going to walk out the doors of the data

centers in in humanoid robotic form in

in driverless car form. I I think what's

most intriguing about the AI5

architecture is it's a unified

architecture that this is a this is a

single accelerator that is planned for

use both in the data center side and in

the robotic/car side single chip which

is this is something new that that the

world hasn't seen before a single

unified architecture for both cloud data

center compute and also embodied in

robots and cars and so I I I think this

is quite literally potential eventually

the embodiment of intelligence walking

out the door of the data center into

into our homes and into our lives.

>> Well, and this ties back to our last

story, too. Uh, you know, all the big

guys now have their own chips as Sam

announced in our last podcast that that

he has uh his own Broadcom custom

designs. So, Enthropic is the one

exception. And so, they're going to

adopt the Google TPUs. that was in that

other slide. But that's not a very

comfortable place to be if all the other

competitors have their own chip designs

and they're as they're modifying their

algorithms. They're tweaking the AI is

tweaking the chip design. So once you're

in bed with Samsung or TSMC or Intel and

you have your whole supply chain going

right into your own data centers, you

can innovate innovate redesign the chip

and get it back into production very

very quickly. You know, Google's already

got that cycle down cycle time way down.

So it leaves Anthropic in kind of this

uncomfortable position where well we're

buddying up with Google. Yeah, but

you're on their TPUs. They're going to

give you whatever they want to give you.

>> Fascinating. But all of this comes back

to TCMC production capability, right? In

Sam in Samsung, there are basically

choke points. Uh

>> yeah, there's no doubt that any one of

these companies would be buying TSMC,

Intel or Samsung tomorrow if the

regulators would let it happen because

because that's the choke point and they

all know it. So all these really, you

know, high level partnerships and

relationships are really really forming

and it's a very competitive playing

field

>> you know week by week we're seeing the

the shifting relationships and uh in

capital flow here. All right, this next

article comes from Amazon and their new

delivery glasses. Let's take a look at

the video here and then talk about the

implications for this. It's fascinating

what this means for uh for labor.

>> Well, check out these nerdy smart

glasses. These are smart glasses

developed by Amazon for their delivery

drivers. So, they're just in development

now, but basically they use technology

to uh like a head-up display, show you

what you need to do. So in this case,

instead of using your mobile phone as a

driver to scan the parcels, you simply

look at them and work out which parcel

needs to go where. But then when you

head out to deliver, it gives you actual

information about the place you're

delivering. It give you warning about

dogs and things and shows you exactly

where to leave it. And it's all done.

Even the photos are taken and you never

need to use the mobile phone. So cool

technology, very much like Meta's Ray-B

bands or maybe Apple Vision from Amazon.

>> Okay, so this is what I think is going

on. It's this is put forward as we're

going to help our drivers you know keep

them away from you know barking dogs and

help them you know do this with

hands-free delivery. I think this is a

mechanism by which Amazon uses the

drivers to collect a lot of information

to train their delivery robots. This is

just like uh like Tesla uh with its

cameras training its its full you know

self-driving models. Dave, what do you

think?

>> Yeah, you're exactly right. And and it

shows you how the technologies interact,

too, because the glasses will be

profitable instantaneously within their

internal use case. They can perfect them

and then they can decide later. You

remember there was a Kindle phone,

Kindle Fire phone. It didn't succeed,

but they they've tried before to compete

with Apple and and anthrop or or Android

in the device warfare, you know, game.

So this is a great stepping stone for

them to make money and perfect the

device while gathering all the data

which will then feed their robotics

initiative but also the consumer glasses

initiative which will come later. So

you're you're exactly right.

>> Yeah. Do you want to add anything Alex?

>> I'll just add I I think this

functionality can generalize well to

non-dely functions as well. I think this

is the tip of the iceberg for for using

wearables to automate and even before we

get to automation to capture telemetry

and training data for the entire

services economy. So I I think that this

we're going to see this across many many

other verticals, healthcare, energy,

hospitality. Expect smart glasses and

wearables for building training data

sets and post-training data sets across

every possible

>> also construction. Construction, you

know, we're doing the biggest

construction buildout in the history of

of America and certainly probably the

world and it's all, you know,

electricity and plumbing and buildings

and everything. But because those are AI

forward projects like you know Chase

Lock Miller at Cruso and and Project

Stargate they're going to be early

adopters of exactly the same thing

you're talking about for construction.

So that'll that'll be and construction

is a huge fraction of the global

economy. Uh so that'll be a really fun

>> and for me this the most important thing

for me uh for a aging population is

going to be memory augmentation right

using these glasses to remember you know

who you're talking to the last

conversation you had I mean personally I

can't wait I meet so many people and I

love being able to you know remember the

details but sometimes it's just a

challenge. All right, we're going to go

into a subject we covered on the last

pod with Emod in particular and Eric

Bolier, but I cannot wait to hear the

take that Alex you have on this and Dave

you have. It's again this is Google's

quantum breakthrough nears realworld

use. So this is the Willow quantum chip.

Uh a friend in Santa Barbara Hartmoot

Nevin who heads the Google quantum team.

Congratulations.

Uh but at the end of the day, Alex, what

does this mean? Well, f first maybe a

little bit of the background. So, I read

the the core nature paper behind this

announcement. Very interesting. Uh this

was the the Google team.

>> And by the way, Alex, I have to say I

really appreciate the fact that you dig

in

>> on everyone's podcast to go and read the

actual science, you know.

>> Well, it's difficult to comment on it if

I haven't read it, but thank you.

>> I understand that.

>> Well, everybody else on the planet is

commenting on it without reading it.

You're the only one doing it. And I

>> I've heard I've seen these I've seen

these uh these comments in uh on YouTube

that Alex is an AI. I've seen him

glitch. Uh you know, God knows.

>> We want to use this as our cold open.

>> I'm not going to I'm not going to

disclose any any details. Uh but maybe

we'll see you in live. Anyway, dive in

please. You read the paper. What does it

say?

>> Right. So, so uh I I read the core

nature paper behind this announcement.

It's very interesting. The the the

premise is that there's a a certain

physical quantity and in the case of of

this announcement, it's called a second

order out of time order correlation. And

it this is basically a measure of

quantum chaos. It it measures how

chaotic a given quantum system is. and

and the Google and collaborator team

showed that it would be very challenging

for a classical computer, which was to

say a nonquantum computer, to be able to

to compute it. So, I I think it's it's

very interesting. It's nice progress in

terms of demonstrating quantum speedups

or quantum advantages versus classical

computers. What I'm still waiting for

though, I if if I got my wish, is a more

call it economically transformative

quantum algorithm. What I'm waiting for,

what I'm hoping for is that sometime in

the next few years, we will achieve a

definitive breakthrough speed up for

quantum acceleration of AI. I I think

applications like this where there are

applications in quantum simulation,

quantum chemistry, simulating materials,

optimizing molecules, I think it's

great. I don't think it is necessarily

worldchanging. And the the worldchanging

use case for for quantum acceleration if

if the physics of our universe are so

kind as to allow them would be I think

something like being able to achieve

orders of magnitude speed up in training

or inference for a frontier model. I

think that would be utterly

game-changing. Amazing. The term quantum

advantage was coined a few years ago as

the point in which a quantum computer

demonstrates the ability to do a real

world thing better than any classical

computer, right? With ones and zeros.

And so people have been chasing this

idea of a quantum advantage really to

rationalize the massive investments and

to actually get traction. Uh now we have

a number of public quantum companies and

you know wanting to get revenues. Uh I

think one of the other important things

to note here is uh the concept of error

rates in quantum computers. Um and uh

how do we get to logical cubits and how

do we reduce the error rate so we

actually uh have something that's going

to be useful. But let me ask you a

different question here Alex. How big is

quantum computation as compared to AI?

How big a a you know relative is it

larger many times larger what are your

thoughts? Well, I I want to answer I I

want to bisect the question into

nowshortterm versus long-term at the

moment and in the short term the the

actual applications are relatively

pedestrian prosaic not economically

transformative. the the the best

applications I I think that that I've

seen anywhere close to to being useful

in the short term are for quantum

simulation leveraging the fact that it's

relatively straightforward as Richard

Fineman who are arguably helped to to

create the entire field of quantum

computing pointed out you can use one

quantum system to simulate another

quantum system relatively easily. I but

these aren't economically transformative

not in the same way as AI that is just

turning our service economy as we were

discussing earlier and just automating

it. Quantum doesn't have that capability

in the short term. In the long term I

would hope quantum will enable us to to

build much faster AI systems. So in the

long term holding out hope that quantum

in the end there's almost an angle

you'll forgive me for this. There's

almost a redemption arc that I'm hoping

for of quantum information systems

because so many of the problems right

now that AI is is solving grand

challenges like protein folding. Do do

you remember 10 20 years ago there was a

sizable community that thought protein

folding would require quantum computers

to solve. That did not happen. We were

able to solve it with just AI on top of

classical computing. So there's there

there's almost a who moved my cheese

angle to to the the sense like the the

the grand challenges that quantum was

supposed to be the the the great white

knight and solve for us keep getting

devoured by AI instead. I I'd love to

see a bit of turnaround sometime in the

next 10 years on

>> fascinating my my favorite science

fiction books all have uh digital super

intelligent AIs conscious AIs uh doing

so on the backs of quantum clusters. So

>> there would be certain advantages like

>> yeah go ahead.

>> So potential advantages like energy

efficiency if if we could build a fully

reversible AI supercomput that that

would are probably be have some sort of

quantum coherent foundation that would

be transformative. We wouldn't need to

we wouldn't need to to build all these

SMRs and uh and vision plants and NATG

gas collocation facilities if if we had

fully reversible quantum computer based

foundation models everywhere. But we're

not there yet.

>> Nice. Uh Dave, let's go to the next

article here and I'd love your your

thoughts on it. Uh which is that

President Trump eyes equity into US

quantum firms. So, you know, this is the

potential beginning of a sovereign style

VC fund for the United States. Uh, he's

targeted IMQ, Regetti, D-Wave, Quantum

Computing Inc., and Atom Computing. Uh,

I mentioned uh on the last pod when we

talked about this that I had taken

D-Wave public through a uh spa um huge,

you know, 8,000x return from the

earliest uh lowest point to where it is

today. Uh Dave, thoughts on this?

>> Yeah. Well, I I love it and I hate it as

a president, but I still love it because

Alex is always pointing out that what

we're what we're doing right now is

unprecedented, except maybe during the

buildup to World War II. And you think

about 1939, we're basically flying

biplanes in the US Air Force. By the end

of 5 years later, we have we have jets.

>> Uh so just incredible amount of

government investment. Yeah. So that's

what's going on right now in AI, and

it's great. It's what we need. So now

that's moving into quantum too. And

you've made the point many times Peter

that our our fun our economy doesn't

function well in these areas that

require you to think more than 5 or 10

years in the future. China works really

well thinking 10 20 30 years in the

future but we don't do that well. So the

government kickstarting quantum is a

great move uh if you believe in it 5 10

15 years in the future. Uh but as a

precedent for government involvement in

the economy, it's terrible. You know,

it's cuz cuz they're going to make

terrible decisions in the long run.

These are very good decisions in the

short run, but that's because all this

incredible talent has gone to Washington

for the first time in my lifetime. But,

you know, that's not sustainable. And

so, I hate it as

>> we see and we see the government

investment triggering huge amounts of

private investment that follow on,

right? So, after the Intel deal, you

know, Intel stock doubled between $20 a

share before and 40 bucks a share, you

know, a day or two ago. Uh, and we're

seeing this again, a 10 to 15% increase

in these quantum stocks after this, uh,

the story got leaked.

>> I wonder where they're going to go next

when they go, you know, I think the

government's be going going into rare

earth metals. We've seen some of that

conversation. Where else might they be

making uh sort of strategic investments?

Well, I I hope they take that, you know,

Alex's World War II analogy and and stay

focused on the things we need in this

very specific race uh to AGI and ASI.

>> So, Rare Earth would fit for sure and

and energy would fit for sure. Quantum

may or may not

>> I'm kind of I'm kind of shocked that the

government hasn't made a move to get

into the fusion companies or the SMR

companies uh really to help accelerate

that because I think that one thing uh

would bring a lot more capital. I mean,

Commonwealth Fusion um is probably the

best funded. You know, I was talking to

some of the fusion companies here at

Visionering and talking about Helion. Uh

interestingly, they said, you know,

Helion is so closed lip. We have

actually no idea what they're doing and

how far they're they're along. You know,

there's public disclosure, some

information. They're claiming 2028

Microsoft, but we don't actually know.

And these were from the top fusion

experts. Uh Commonwealth Fusion, you

know, targeting 2030, but they still

have a lot more development. Alex, do

you have any thoughts on that? Yeah, I'm

not going to second guessess the

commerce department or the executive,

but there is some reporting that there

may have been some money left over from

the chips act and quantum firms might be

interested certainly would be interested

in either obtaining uh equity

investments or my guess is more likely

loans or or warrants or or some other

financial structure. I I think the the

question of how strategically important

quantum is as a technology when you

compare it with more obvious feed stocks

like rare earths or energy or compute or

fabs. I I I think that's that's to be

decided. I don't know.

>> Well, I will say I can't add anything to

Alex's insights on this at all, but I

will say I talked to Frank Wilchuk about

it. He's a Nobel Prize winner in was

winner in physics and you know famous

and spent his whole career in quantum

physics and he said almost exactly the

same thing Alex said. So there's two

data points.

>> All right, let's jump into energy. A few

different articles here. Uh this one's

interesting. Uh in particular's a chart

showing us the increasing price for US

construction of nuclear reactors versus

China. And here's the quote.

Construction costs for nuclear reactors

in the United States have risen roughly

a th000% since 1970s while China's costs

have steadily declined. Um, that's not

good news. Alex, do you want to weigh in

on this?

>> Yeah, I I think there is an alternative

history where the US never basically

stopped building nuclear plants in the

late 1970s. And if you're familiar with

all of the the microeconomics around

experience curves, costs, unit costs

tend to collapse the more you make of a

given item. And as a country, the US

basically stopped making nuclear power

plants decades ago. And we're going to I

I think if if we're going to feed the

voracious energy appetite of of these AI

data centers, we we need as a country to

relearn how to build lots of next

generation nuclear plants. And the good

news is the demand signal is being sent

by the AI data center companies. But I I

think there will be all of these

knock-on benefits, not just for AI data

centers, but for everyday life if we

live again in a a truly powerrich

society.

>> Well, Alex, it's worse than it's worse

than that sounds too because it's not

just about unit costs. If you look at

the actual construction of a nuclear

facility in the US, it's mostly

overhead, regulatory, political garbage,

cost.

>> It's regulation, it's litigation, it's

loss of manufacturing expertise, all of

these things. said, "We've done it to

ourselves." All right. Uh, next article

here is fascinating. US is offering

nuclear energy companies access to

weaponsgrade plutonium. So, this comes

out of energy secretary Chris Wright.

The US Department of Energy will let

private firms use 19 tons of plutonium

from old warheads to fuel their next

generation reactors. Um, the moving the

move is boosting domestic nuclear

supply, reducing reliance on Russian

uranium. Uh I find this as a fascinating

move. I mean talk about you know sort of

removing the shackles uh and giving

entrepreneurs access to feed stock.

Who wants to take it?

>> Well, everybody everybody probably knows

this, but the cost of the fuel in a

nuclear reactor is tiny. It's it's a

rounding error. And so everyone's been

buying their fuel from Russia for a long

time. Opening up the US supply doesn't

really change anything. It's a rounding

error in the overall costs anyway, but

you know, if you're going to buy it from

Russia anyway, what's what's what's the

harm in using our surplus plutonium? So,

it's not it's not changing the math one

iota.

>> Alex, take I I'd also comment maybe even

more broadly on nuclear engineering as a

vibrant discipline. There there was

maybe a bit of a hot take uh but there

was a period of time for a few decades

when nuclear engineering unless it was

for say some biomedical application was

positively unfashionable uh to to study.

Uh, and I I I think that um I I don't

want to call it a nuclear winter for

obvious reasons, but that there there

there was a I think that period of time

we're coming out of that now. And as a

society, speaking particularly of the

US, but the the West in general is is

entering an era when we need to

refamiliarize ourselves with with the

nuclear fuel cycle and get comfortable

with nuclear fuel cycles in general.

It's part of the future. uh in

particular part of the future is fusion

uh and so the US has put forward a new

roadmap for fusion energy. The DOE road

map touts commercial fusion by the mid

2030s

actual aim to for public infrastructure

uh in the 2030s to scale up. Uh

interestingly this has zero dollars of

federal funding behind it uh and $9

billion of private investment. Alex, you

found this particular timeline. Talk to

us about it. What does it mean?

>> Yeah. No, I I I I enjoyed reading the

road map. I I thought it was delightful

in in some respects. So, the the road

map calls for three stages of

advancement in in fusion energy in the

US. The the first stage, call it the

short term over the next 2 to 3 years,

calls for early stage price

demonstrations. So, that takes us

through 2027 2028. the the second stage

medium-term calls for early stage fusion

pilot plants uh between 2028 and 2030.

And the the third quote unquote

long-term calls for actual operation at

production of generation power plants

between 2030 and 2035. So this is

actually a very I I think some would say

it's a very ambitious timeline at least

by historic standards where fusion was

always 30 to 50 years out. Now it it's

basically in in our short term. Uh and

it also I think aligns with some of the

public announcements that Helion on the

one hand and Commonwealth Fusion on on

the other hand have made regarding

actual test facilities being in

operation between 2028 and 2030. So I I

think in in short this road map is more

a reflection or at least I interpreted

as more a reflection of some of the the

most ambitious private sector players

and their actual plans. Yeah, Dave,

we're going to be having uh dinner with

uh Bob Mumgard on Wednesday night in

Riad. We have our abundance dinner that

we're co-hosting uh with uh Amjad from

Replet uh and Link Ventures. A lot of

incredible people are going to be there.

So, I look forward to asking him more

about this.

>> Yeah, me too. Yeah,

>> I mean the head of Commonwealth Fusion,

he's done extraordinary work. Uh I'm

excited to see where they're going to

go. All right, continuing on the energy

theme. Amazon bets big on nextg nuclear.

So this is uh the state of of SMR, small

uh modular reactors. This one is with X

energy. We've talked about X energy

before. its initial 320 megawatt output

that can scale to nearly a gigawatt uh

which can power data centers obviously

carbon-f free um you know I love SMRs

and I love the Gen 4 nuclear reactors uh

you know we unfortunately shut we talked

about this we've shut down our ability

to manufacture these and so this has

become an entrepreneurial effort but one

of the things that I find fascinating is

while we have the designs we have

permissions the timelines for getting

these SMRs out. Um they're not like 26,

27, 28. They're 2030s.

>> Um which is concerning. Why can't we get

these going going faster, right?

>> No, the timelines are timelines are

really interesting to track and it'll

come up at FII next week in a big way.

But uh you know, a gigawatt you we you

know, Eric Schmidt said we need 100

gigawatts by 2030. And that's just a

that's just a fact. You know, it can't

go up or down because that's the number

of GPUs we'll be making. they're going

to go into production one way or

another. And so you need to find 100

gigawatts um by 2030. That's only about

a 10% expansion of the US power supply.

So it's not it's not insurmountable. But

then 2031 2032

the the GPU the new fabs will be online

and the GPU production will go way up in

2031 2032. And so then you need some

massive you know the 100 gigawatts is a

stepping stone to something much bigger

just a few years later. So if the fusion

comes online in 2029 2030 it's massively

important but if it's just 5 years late

like where's that power going to come

from then suddenly you're launching them

into space and so these completely

different ideas you know and the modular

reactors here they're fision so that's

the third option plus renewable is a

fourth so all those things are racing

against this 2030 clock

>> I have to imagine by 2030 we're going to

have figured out more energyefficient

compute um 10x or 100x more efficient.

And you know, Alex, I'd love to hear

your thoughts on that.

>> The intelligence of the AI between here

and there is going to be like,

>> yeah, but also like I I I I have to

invoke Jeban's paradox. We're we're

going to have presumably much more

demand for it as well, even though cost

per computer algorithmic advances are

going to to 5x to 10x every year. may be

uh optimistically the the amount of

energy uh the energy reduction that we

need in any given year. So I I I don't

know when or if there will be a turning

point where we need less energy. I I

will point out though with the SMRs I I

think it's striking no cooling towers.

This is a totally new form factor.

Decades of acculturation people being

trained to look for those iconic

cylindrical cooling towers. No cooling

towers. These can be put in so many more

locations. They are compact. They can be

put into novel sites that otherwise

might never been might never have been

on the table for for some of the first

generation nuclear power sites. So even

if there is a sequencing issue and even

if the the the first boatloads of SMRs

start arriving circa 2030, I I do think

they they're very likely to end up being

an important part of the overall power

mix for AI data centers and otherwise.

Yeah, keep in mind the vast majority of

the data centers don't need to be near

population centers.

>> And that's a big difference. You know,

the those iconic cooling towers that

Alex was mentioning, people hate them

when they're on the beach in front of

your house.

>> But these these SMRs can be, you know,

Wyoming and and Texas and Nevada in the

middle of, you know, very unpopulated

areas. That's a great place to put some

of these really large scale data

centers. So, this will happen for sure.

>> They look like they look like normal

buildings. That that's what's most

striking to me. You you would never at

least with the eyes of 2025 today look

at the the building that you're sharing

and say, "Aha, that that's that's

obviously a fision site. It looks like a

normal building."

>> Amazing. So, you know, we're going to

see a continued mix. Um, I sure hope

that the government does start backing

solar and backing SMRs and backing

fusion more. We need to accelerate our

energy production uh beyond just natural

gas and coal and other areas. Uh, I'm

going to end this with what I'm going to

call uh a weird science article. So,

let's uh let's end on something that um

doesn't normally enter our conversation

in the exponential world. Alex, you

found this one. Uh it's called butt

breathing, a real medical option. Do you

want do you want

>> Sure. Sure. Peter, I I'll take the hit

for ending on a low note. So, so but but

in all seriousness, um this is a

transformative breakthrough or at least

the beginnings of a transformative

breakthrough for people suffering from

severe respiratory failure who can't

breathe through their lungs. Uh and if

folks have seen the abyss, the the

science fiction movie where there's a a

famous scene where uh a character is is

consuming uh oxygenated or I should say

an oxygen substitute liquid. So

breathing liquid basically uh deep

underwater that they'll have some

familiarity with novel forms of of

respiration and blood oxygenation. This

was also the subject of last year's

Ignobbel Prize u for for discovering

that non-human animals could oxygenate

their blood supply by uh by consuming

oxygen uh via the other end as it were.

So only only so many euphemisms I can

use here, but

>> well the intestines are a very bloodrich

large surface area part of your body.

And so if you're able to put sort of a

uh hyper oxygenated fluid enema, let's

call it that. Uh then you can perhaps

oxygenate your blood supply and get

enough uh enough of your uh your red

blood cells oxygenated to get to your

brain. Seriously,

>> I mean,

>> it's like it's

>> but but to elevate just a little bit,

>> we've lost our entire audience on this

particular art. The only reason I like

this story and I wanted it in the

podcast is because every time Salem says

something in the future, we have the

option to say, "Oh, he's butt breathing

to to to maybe just to to try to elevate

a little bit that there's been interest

over the decades in uh nanoobots that

would help with uh with oxygenating the

blood, so-called respirites.

And to the extent that it's possible,

and I I should add also parenthetically,

sci-fi scenarios like enabling humans to

be able to hold their breath underwater

for hours on end. So there there's been

like persistent sci-fi pressure to

discover new ways to oxygenate the blood

in uh in environments that are call them

uh less than hospitable. So to the

extent

>> you're you're really really reinforcing

the theory that you're an AI

So I you know all of this all of this

materializes on the backside of

nanotechnology and one of these times

you know I really want to dive into not

wet nanotechnology where we're using DNA

origami but you know drexlerian uh you

know assemblers that just opens up

everything and respites are fantastic

you know uh literally BCI enabled

through uh it's through nanobots in in

the brain um I can't Wait. So, you know,

I'm going to get Ray Kerszswwell on our

podcast uh so we can have the

conversations with him. Uh Ray's been a

dear friend and a mentor for so many

years. At the end of the day, you know,

his prediction is nanobots by uh the

early mid 2030s, so 2033. And that's

going to unlock uh you know high

bandwidth BCI but unlocks basically

longevity escape velocity or I don't

like using the term immortality cuz it

sort of hits so many different negative

buttons but if you can repair on a

cellular and subcellular level all parts

your body that is an incredible future.

>> Well if you get Ray and Alex on the same

podcast that podcast could also be

immortal. That would be something I

would kill to see. Well, we'll we'll do

that uh for sure. And uh again, to all

of our friends listening, I hope you've

enjoyed this episode of WTF. Uh if

you're not a subscriber, please join us.

We'll let you know. You know, it's

interesting. We're putting out news as

it breaks. So, while we try and do this

once a week, sometimes it comes out

twice a week. And uh you'll get a notice

of that. Uh we hope that uh other than

butt breathing that this helps you

understand how fast the world is

changing. uh and that you know we're

living this extraordinary time uh where

we can solve any grand challenge.

Congratulations to the visionering

x-priseze teams uh for winning

visionering and to the entire x-prise

organization for really accelerating

these grand challenges. I'd love to know

in uh in the notes here if you have an

X-P prize that you'd love to see in the

future. Let us know what it is. Uh Dave,

I'm heading to the airport in I think

two hours to head to Saudi.

>> Crazy.

>> It's gonna be fun. I'll see I'll see you

and Immod and Sem there. Alex, we will

miss you.

>> You'll be there either digitally or in

spirit, but we have quite quite the week

lined up uh meeting with the top CEOs

from all the AI and tech companies. Um

it's going to be fun. Any favorite

meetings you're looking forward to,

Dave?

>> Well, you know, you're kicking it off

with the the big shots. So, you know,

you've got Eric Schmidt, Larry Fank,

just like the the big big money people

and the big vision people. So, that's

the that's going to be such a fast

start.

>> But then backstage, it's like God, it's

just like a who's who of incredible

people. So, I'll be backstage the whole

time.

>> Uh it's Yeah, it's going to be wild.

>> So, thanks all. Uh

>> yeah, no, a pleasure. I chair FII out of

Saudi. It's the future investment

initiative and I'm on the board there

and I chair their AI activities. You

know, one of the things that's going to

be interesting this year is we have uh I

think 20 something heads of state and

I'm going to be co-chairing uh a

conclave with uh and uh an Mida from

A16Z

uh and we're going to be talking about

how to use AI to accelerate governance

for countries. You know, one of the

biggest challenges we have, we'll talk

about this when we come back, is that

the speed of change is so extraordinary

and so disruptive in terms of AI and

humanoid robots and longevity that

countries out there are having a

difficult time trying to understand what

policies do they put in place? How do

they, you know, what do they do best for

their for their nation state, for their

citizenry? And so, we're going to be uh

announcing a program uh called Sovereign

AI governance engine. We'll talk about

what that means, but it's really to help

people around the world deal with the

disruptive change and disruptive

opportunity uh at the speed of AI versus

the speed of uh governing governments

and PDFs.

Yeah,

>> it's going to be good. Yeah. And the re

the reason that's coming out in Saudi

Arabia at RI in Riyad is because uh the

deployment rate of ideas like that can

be very very very fast in those

countries because you know they make

decisions kind of in a in a very

tight-knit little very very fastmoving

group. And so that'll be a huge bell

weather for Western democracies because

it'll it'll happen there long before it

happens in the US and Europe.

>> Alex, what's the week like for you

buddy?

It's in some sense the the same as every

week for me, which is trying to

accelerate and smooth out the gentle

singularity.

>> Yes, I love that. By the way, our

episode on the singularity is now has

just done incredibly well. People um I

mean I've had people telling me uh f you

know faculty at UCLA and others saying

I've assigned this to all of my students

to listen to

>> to that podcast. Yeah. No,

extraordinary. It's It's really done. Uh

it's gone viral. So, if you haven't

heard that episode, the Singularity is

now. Go listen to it. Uh it's the

Moonshot Mates at their best. Love you

guys. Uh see you on the other side of

the pond. Dave, Alex, see you in a week

when we're back.

>> All right.

>> Sounds great.

>> All right.

>> Take care all.