Grok 4 Wows, The Bitter Lesson, Elon’s Third Party, AI Browsers, SCOTUS backs POTUS on RIFs

1h 30m

(0:00) The Besties welcome Travis Kalanick and Keith Rabois!

(3:02) Travis on Pony.ai / Uber rumors and the state of Cloud Kitchens

(18:51) xAI launches Grok 4, learning "The Bitter Lesson" in AI

(40:36) How Grok can catch ChatGPT in usage, OpenAI's product excellence

(46:27) Perplexity and OpenAI building AI-native browsers and taking on Chrome

(58:01) Elon's "America Party": is now the right time for a third party, and could he make an impact in 2026?

(1:13:12) SCOTUS backs Trump over federal government RIF plans

Follow Keith:

https://x.com/rabois

Follow Travis:

https://x.com/travisk

Get The Besties All-In Tequila:

https://tequila.allin.com

Join us at the All-In Summit:

https://allin.com/summit

Summit scholarship application:

http://bit.ly/4kyZqFJ

Follow the besties:

https://x.com/chamath

https://x.com/Jason

https://x.com/DavidSacks

https://x.com/friedberg

Follow on X:

https://x.com/theallinpod

Follow on Instagram:

https://www.instagram.com/theallinpod

Follow on TikTok:

https://www.tiktok.com/@theallinpod

Follow on LinkedIn:

https://www.linkedin.com/company/allinpod

Intro Music Credit:

https://rb.gy/tppkzl

https://x.com/yung_spielburg

Intro Video Credit:

https://x.com/TheZachEffect

Referenced in the show:

https://www.nytimes.com/2025/06/26/technology/uber-travis-kalanick-self-driving-car-deal.html

https://www.youtube.com/watch?v=ZW5fJikPmfM

https://grok.com

https://www.youtube.com/watch?v=_wTA90BYo30

https://techcrunch.com/2025/01/08/elon-musk-agrees-that-weve-exhausted-ai-training-data

https://x.com/ArtificialAnlys/status/1943166841150644622

https://x.com/elonmusk/status/1943192643439337753

http://www.incompleteideas.net/IncIdeas/BitterLesson.html

https://x.com/chamath/status/1943177837956968499

https://techcrunch.com/2025/07/09/perplexity-launches-comet-an-ai-powered-web-browser

https://x.com/perplexity_ai/status/1942969263305671143

https://x.com/elonmusk/status/1941584569523732930

https://polymarket.com/event/will-elon-register-the-america-party-by

https://ropercenter.cornell.edu/presidential-approval/highslows

https://news.gallup.com/poll/651278/support-third-political-party-dips.aspx

https://www.whitehouse.gov/presidential-actions/2025/02/implementing-the-presidents-department-of-government-efficiency-workforce-optimization-initiative

https://www.scotusblog.com/2025/07/supreme-court-allows-trump-administration-to-implement-plans-to-significantly-reduce-the-federal-workforce

https://www.afge.org/article/summary-of-afge-lawsuits-against-trump--how-litigation-works

https://cei.org/publication/10kc-2025-numbers-of-rules

https://www.netflix.com/tudum/articles/american-manhunt-osama-bin-laden-release-date-news

Listen and follow along

Transcript

I have a very funny story to tell you, Jason.

Where have you been?

I've been trying to text you.

You've been offline.

What's going on?

Where have you been?

I've been working feverishly, but yesterday I had to

go to prepare for some meetings that I have on Sunday, which I can't tell you about.

But

Nat and I went to Pasalaqua, which is in Lake Como, which is an I mean, it's stunning.

The grounds are stunning.

The hotel is stunning.

If you have a chance to go to Lake Como,

anyways, this is us at Pasalaqua.

Who's the beautiful woman there?

Is that the woman that owns it?

Is that the question?

But the best part is, we had such a good time.

You know how they have like a registry book to leave a message?

Sure.

So I left a message.

Here we go.

What a truly magnificent place.

Above and beyond any expectation we had.

Go below that stuff for me.

Thank you.

We took everything the free.

we took everything the free birds.

Jason, the hangers, okay, everything,

the laundry bags.

Did you get to space the robes, the slippers,

everything absolutely fantastic.

You're gonna have to send a bill to the free birds at

the end.

Let your winners ride

Rainman David Saturn.

And it said, We open source it to the fans, and they've just gone crazy with it.

Love you, Bessie.

Queen of Kinwa.

All right, listen, we've got a great panel this week.

It's the summer, things are slow.

Some people are busy.

I think

our prince of panic attacks, our dear sultan of science, is he's at the beep.

Sachs is busy.

Couldn't make it this week.

In his place, another brilliant PayPal alumni and

dare I say, GOP supporter, Heath Raboy.

How are you, sir?

Pleasure to be with you again.

Nice to see you.

And I'm assuming you're in gorgeous Florida or somewhere in Italy.

Yeah.

I'm actually in New York.

Oh, my hometown.

Is it safe?

Is it okay?

Mom Dami chasing you down the street?

Not yet, but it's safe to see.

Did you seize your assets?

Seize your assets?

See?

Yeah, it's safe right now.

We'll see you on November 4th.

You know, as you probably heard, on July 4th was the first time in recorded history that there were no shootings or no murders in New York on that day.

So right now, things are in pretty good shape, but we

may be leaving New York quickly.

Yeah, you're going to probably want to sell that place if you got one there because Mondami is going to seize it and turn it into a drugstore for you.

Yes, it's going to be the Mondami drugstore.

Travis Kalinik is back with us.

How are you doing, Bestie?

Pretty good.

Pretty good.

Yeah.

Second appearance here on the roundtable and third time on the show.

Of course, you spoke at the summit.

You've been busy with Cloud Kitchens.

Yeah.

Lots of exciting things going on.

Oh, lots of stuff.

Lots of stuff.

The robots are taking over.

We're rolling out.

We're rolling out robots.

TK, can you tell us what you're doing with this Pony AI thing or not?

That's speculation.

Look, you know, obviously is autonomy

as we, you know, in the U.S., we have, of course.

Wait, do you want to just frame for people that don't, that may not be up to speed what was announced or at least

Pony AI is an autonomous company doing self-driving.

It's one of the few players that actually have cars on the road.

They're based in China.

They've got a lot of operations in the Middle East.

They've got a deal with a livery company called Uber, which you might be familiar with.

Okay, so

look.

Well, the deal was basically that you would partner with Uber, license in the pony technology, and essentially start a competitor, I guess, to Waymo and Tesla.

Let me work on this one.

Okay, so

in the U.S., we have Waymo.

We see the Waymos in San Francisco, Los Angeles, Boston, coming soon to Miami, coming soon to Atlanta, coming soon to D.C.

They're even talking about New York.

Tesla is sort of like the, you know, they're doing it the hard way, you know, classic Elon style, like let's

do this sort of in a fundamental, holy shit, let's go all the way kind of approach.

And it's unclear when it gets over the line.

Of course, he, he launched sort of a

semi-pilot of sorts in Austin recently.

But there's no other alternatives.

So what happens is, is some of the folks who are interested in making sure there are alternatives have reached out.

They've reached out to me and there are different discussions that get going because they're like, Travis, you did autonomy way back in the day.

got the Uber Autonomous stuff going in 2014.

Maybe there's something to do here to create optionality.

Now, it may be like, I'm, of course, very interested on the food side.

I talk about autonomous burritos being a big deal because if you can automate the kitchen, the production of food, and then you can automate the sort of logistics around food, you take huge amounts of costs out of the food,

out of what's going on in food.

And that's, of course, near and dear to my heart.

There's folks that, of course, that want to see autonomy and mobility.

That's a real thing.

It may be that, or I would say if you get the autonomy problem right,

you can use it to apply to both problems.

So there's a lot of folks interested in

moving things, moving food, moving people.

And if

there is some kind of autonomous technology that maybe I get involved in, it might apply to a bunch of different things.

And so I've got some inbound.

Let's just put it that way.

There's no real deal right now, but there is definitely some inbound.

And I think there is some news about some of that inbound that may or may not be occurring.

That's probably the best way to put it.

It was long-winded.

I'll try to tighten that up next time.

No, no, I think it's great to get the overview here first and

all in.

Thank you for sharing it with us.

And everybody knows you have been doing a ball builder,

Lab 37, I think it's called.

And we throw it up on the screen.

I'm not sure what the status of it is.

And then I'll let you go, Jamath, with your follow-up question.

But I think there's a pretty interesting concept here of the bowl getting built and then put into a self-driving car.

That machine looks huge, but it's actually 60 square feet.

That picture makes it look monstrous.

It's a 60 square foot machine.

Like imagine running like a sweet green-like brand or a Chipotle-like brand and just making it so it comes to life for people who,

you know, are like, hey, what is this thing?

Imagine you just order online exactly the kind of bowl you want.

And actually, this machine could run run like many brands at the same time and does.

You build the bowl you want, whatever ingredients.

It sort of,

if you look at that bottom, you see those little white bricks at the bottom?

That's what carries the bowl underneath dispensers.

It fills up.

The machine puts a, it sauces the bowl, then it puts a lid on it.

It takes the bowl, puts it in a bag.

puts utensils in the bag, seals the bag, and the bag goes down a conveyor belt where then another machine,

what we would call an AGB, takes the bowl to the front of house.

The bowl gets put into a locker.

The courier, be a DoorDash who breeds courier, will wave their app in front of a camera and it will open up the locker that has the food that they're supposed to pick up.

So it just takes out a lot of what we would call the cost of assembly, which is

reduces mistakes, right?

It makes sense.

We know exactly how many grams of every ingredient are put in.

That's exactly what you're supposed to get.

And

so you get a higher quality product.

It takes a lot of the cost out.

You imagine ultimately that's going to be, there are going to be couriers with that as well.

That, you know, I like to say autonomous burritos.

Like, is a Waymo going to carry a burrito?

Or is Tesla going to have a machine that carries food?

Or, you know, is there another

company that ends up doing sort of the things,

the autonomous delivery of things.

And the point is, is, well, where we are right now is we've got customers.

And so those customers are starting to deploy this quarter.

And it's pretty interesting.

I mean, you can see

in our delivery kitchens,

the cost of labor is about 30% of revenue.

That's what the successful guy, let's say 30%, 35% of revenue.

In

a brick and mortar restaurants, it's even higher.

When they're running our machine, it's between seven and ten percent of revenue.

Amazing.

Then you take out the cost of the delivery, you know, and now it's becoming everybody can have a private check, which was your original vision for Uber.

It was, people don't know the original tagline, but it was your private, everybody has a private driver.

Everyone's private driver was the original for Uber.

Basically, the infrastructure was already there.

And I said this on, you know, one of your recent,

I think it was at the all-in summit, Jason, but like

in the mobility, cars, you know,

transport space, the roads were already there.

The cars were already built.

People weren't using their cars 98% of the day.

So the infrastructure is already there to get people around,

to do this as a service and do it very efficiently and conveniently.

With food, the infrastructure is not there.

Like, yes, restaurants have excess capacity.

That's what Uber Eats utilizes.

But to go and say, like, let's make 30% of all meals in a city sort of

prepared and delivered by a service, the infrastructure is not there.

So you have to build it.

So our company,

the mission is infrastructure for better food.

So that's real estate, that's software and robotics for the production and delivery of food in a super efficient way.

All right.

Keith, what are your thoughts?

Any questions for you?

Well, he's not here, but isn't this what David David Freeberg tried to do a few years ago?

Yeah, this came up on the last all-in, yeah, or the last one I was at.

Yeah.

Yeah.

Itza.

It's a.

The problem was, I told Freeberg, people don't want to eat quinoa.

You got to put a little steak in there, maybe a piece of salmon.

But he was kind of really, I think eventually he relented and let people have a little bit of protein.

But yeah, so it's such a great vision.

Wait, he died as a vegan martyr?

I think the business died as a vegan martyr.

Well,

that was the hill he was led to.

There's a lot of people that died on that hill.

But the bottom line is if you're going to get into automation, you have to, it has to be end-to-end automation.

And what I mean by that is like, there are pizza, there are pizza companies that have come and gone, automated pizza companies, where it's like, we have a pizza machine and everybody's like, yeah, this is amazing.

And you have a guy, you have a million dollar pizza machine.

And then on the left, you have a guy feeding ingredients into the pizza machine.

And on the right, you have a guy taking the pizza out and then putting it in a box and doing all this.

So instead of one guy making pizzas, I have a million-dollar machine and two guys making pizza.

And so when you look at these

robotic food production machines or food assembly machines, you have to look at the full stack and say, does it work with the ecosystem that exists in a restaurant?

And does it go full stack from, you know, like we have this thing, that machine we saw earlier.

The staff preps the food, they put the food in the machine, and then they leave.

They're gone.

This restaurant runs itself for many hours without anybody there.

But this could be McDonald's, Burger King, and Taco Bell.

Nobody would know.

That right there, that machine is

an assembly machine, right?

The food is prepped by humans and then assembled by this machine.

For a Chipotle or a sweet green, this is like a majority of their labor, right?

You go up to a chipotle, there's like 10 guys at lunch, and you're still in line.

That machine right there does 300 bowls an hour, right?

And so

you go, okay, that's the this is what's called

like the assembly line.

It's just that front line where you basically assemble things.

I think sometimes I'll call it the make line.

What will happen over time is you'll have perpendicular lines going into it where you're producing food,

right?

So you'll have a production or make line going into an assembly line here.

And then you go, oh, wow.

So you have something that dispenses burgers on guns.

That's the dispenser.

That's the assembly.

Right.

But it's like tutorial on steroids, basically.

Yeah.

And then it's like, how do you cook that burger?

That's what I call, that's what we call state change.

So state change is the, is the cooking of the food.

Assembly is the like, how do I put it together and plate it?

Doesn't this collapse, Like, for example, if you have a yield of 300 per hour, you said, out of that one machine,

very quickly you can impute the value of having a smaller footprint store with five of these things in a faceless warehouse with drone delivery or cars.

You don't need the physical infrastructure.

So then don't you create a wasteland of real estate or how do you repurpose all the real estate?

Well, the way to think about it is like 90%.

Well, it's probably a little lower than that right now.

Let's say 85% of all meals in the US are at home.

They just are.

And a vast majority of those meals are cooked at home.

So,

you know, like Uber Eats and DoorDash, they represent like 1.8% or 2% of all meals right now.

It's very tiny, right?

So what you're doing is you're using real estate to and infrastructure to prepare and deliver meals to people at their homes.

And so it's not, restaurants still exist.

We're still going to want to go to restaurants.

We're still going to want to go outside.

We learned that during COVID.

We knew it before.

We definitely know it after.

And so I don't, it's not really like a decimating real estate situation.

It's taking a thing we used to do for ourselves and creating a service that does it higher quality.

Sort of, I like to say, you don't have to be wealthy to be healthy

and just infrastructure to get that cost down.

And so you're doing something as a service that we used to do at home.

I think in the super long run, you're like, what, where's the story on grocery stores?

If you go to like in 20 years, I think everybody agrees

you will have machines making very high quality, very personalized meals for everybody.

This will be good for Keith because he measures stuff down to like five calories based on his Instagram.

Keith,

what's your body fat?

Like 7%, 8%?

It's like Karen.

Just open his Instagram.

So he posted four

times today about his body fat.

He's like so disgusted with himself at 10%.

It's like bad at 10.

But I actually think the vision of this, actually, the natural implication, and maybe the home run version of this is everybody has a private chef in their house, a robot in their house that actually does this personalized because people do want to cook at home, but they don't have the time.

You know, more space space

and infrastructure.

But man, these delivery services are charging, rich people do this all the time, right?

They do these crazy meal delivery services for 200 bucks a day.

And this is just going to abstract it down to everybody.

And man, people get creative when there's that empty space, to your point, Jamaf, about what happens to all this space.

When I lived in New York in the 80s and 90s, It was common to, in Tribeca, in West Chelsea, where I lived, to take storefronts, put your little architect's office in the front and live in the back.

And many people were hacking real estate.

We still need five, 10 million homes in this country, and they're already doing this with malls.

I keep seeing malls being turned into colleges and creative spaces.

One of them in Boston, they turned like the second and third floor into studio apartments for artists.

So, you know, where there's a will, there's a way.

We could use the space.

I mean, where this goes, what Jamas saying and where the real estate goes is we call it the internet food court, where, you know, you're on Amazon, right?

It's the everything store.

Now, Now, imagine that for food.

And then imagine you have an 8,000 square foot facility where basically anything can be made.

Anything can be made.

Because if you have that machine you saw has 18 sort of dispensers for food, 10 different sauces, you get the idea.

Now, what about when it's 50 or 100 dispensers for food?

What if you have multiple machines with 100 dispensers for food?

That's crazy.

You can, the combinatorial math in terms of what's possible, what can be made,

sort of, you know, goes exponential.

And so

the Internet Food Court is sort of the vision for where this all goes.

Another example of the bitter lesson.

The bitter, yeah, we're going to get to that, I guess, today.

In a very full docket, before we get to that, just a little bit of housekeeping here.

September 7th, 8th, 9th in Los Angeles, the all-in summit again, all-in.com slash yada, yada, yada.

The lineup is stacked.

And we're going to start announcing the speakers.

People have been begging us to announce the speakers.

I don't know.

You got to hold some back.

Careful, careful.

Hold a couple back, but we got some really nice speakers lined up.

It is going to be extraordinary.

It is the best one yet.

I mean, well done.

Every year

we have this.

Yeah, yeah.

Every year we have this little bit of panic.

Like,

you know, we're going to get great speakers.

And man, they started flowing in this week.

It's going to be extraordinary.

Almost as extraordinary as this delicious tequila behind my head here.

Get the oil in tequila, tequila.com.

Deliveries begin late summer.

It's moving to the side.

You can't even talk about it.

How's that?

Summer back here.

Oh, yeah.

All right, listen.

Oh, wow.

Lots to discuss this week.

Obviously, AI is continuing to be the big story in our industry.

And for good reason, our bestie, Elon, released Grok 4 Wednesday night.

Two versions, base model and a heavy model, 30 bucks a month for the base, $300 a month for this heavy model, which has a very unique feature.

You can have a multi-agent feature.

I got to see this actually when I visited XAI a couple of weeks ago, where multiple agents work on the same problems and

they do that simultaneously, obviously, and then compare each other's work.

And it gives you kind of like a study group, the best answer by consensus.

Really interesting.

According to artificial analysis benchmarks, you can pull that up, Nick.

Rox4 base model has surpassed OpenAI's 03 Pro, Google Gemini's 2.5 Pro as the most intelligent model.

This includes like seven different industry standard evaluation tests.

You can look it up, but reasoning, math, coding, all that kind of stuff.

This is, you know, book smarts, not necessarily street smarts.

So it doesn't mean that these things can reason.

And obviously there was a little

There was a little kerfluffle on

X, formerly known as Twitter, where XAI got a little frisky and and was saying all kinds of crazy stuff and needed to maybe be red-teamed a little bit more decisively.

Many of you know Grok4 was trained on Colossus.

That's that giant data center that Elon's been building.

And we showed the chart here.

Chamath,

you sent us a link to the bitter lesson by Rich Sutton in the group chat.

That's the 2019 blog post.

We'll pull it up here for people to take a look at and put it in the show notes.

Maybe just generally

your reaction to both how quickly Elon has, and that chart showed, and how quickly Elon has caught up.

And I don't think people expected him to take the lead, but here we are.

Before we start, Nick, can you please show Elon's tweet about how they did on the AGI benchmark?

It's absolutely incredible.

Two things.

One is

how quickly, starting in March of 2023, so we're talking about less than two and a half years, what this team has accomplished,

and how far ahead they are of everybody else, as demonstrated by this.

But the second is a fundamental architectural decision that Elon made, which I think we didn't fully appreciate until now.

And it maps to an architectural decision.

he made at Tesla as well.

And for all we know, we'll figure out that he made an equivalent decision at SpaceX.

And that decision is really well encapsulated by this essay, The Bitter Lesson, by Rich Sutton.

And Nick, if you can just throw this up here.

But just to summarize what this says, it basically says in a nutshell that you're always better off when you're trying to solve an AI problem taking a general learning approach that can scale with computation because it ultimately proves to be the most effective.

And the alternative would be something that's much more human labored and human involved that requires human knowledge.

And so the first method, what it essentially allows you to do is view any problem

as an endless scalable search or learning task.

And as it's turned out, whether it's chess or go or speech recognition or computer vision, Whenever there was two competing approaches, one that used general computation and one that used human knowledge.

The general computation problem always won.

And so it creates this bitter lesson for humans that want to think that we are at the center of all of this critical learning and all of these leaps.

In more AI-specific language, what it means is that a lot of these systems create these embeddings that are just not understandable by humans at all, but it yields incredible results.

So why is this crazy?

Well, he made this huge bet on this 100,000 GPU cluster.

People thought, wow, that's a lot.

Is it going to bear fruit?

Then he said, no, actually, I'm scaling it up to 250,000.

Then he said it's going to scale up to a million.

And what these results show is a general computational approach that doesn't require as much human labeling, can actually get to the answer and better answers faster.

That has huge implications because if you think about all these other companies,

what has Lama been doing?

They just spent 15 billion to buy 49% of scale AI.

That's exactly a bet on human knowledge.

What is Gemini doing?

What is OpenAI doing?

What is Anthropic doing?

So all these things come into question.

And then the last thing I'll say is: if you look back, he made this bet once before, which was Tesla FSD versus Waymo.

And Tesla FSD only had cameras.

It didn't have LIDAR.

But the bet was, I'll just collect billions and billions of driving miles before anybody else does, and apply general compute, and it'll get to autonomy faster than the other more laborious and very expensive approach.

So, I just think it's an incredible moment in technology where we see so many examples.

Travis is another one, what he's just talked about.

You know, the bitter lesson is you could believe that, you know, food is this immutable thing that's made meticulously by hand by these individuals, or you can take this general-purpose compute approach, which is what he took, waited for these cost cures to come into play, and now you can scale food to every human on earth.

I just think

it's so profoundly important.

One thing I'll throw out there, Chimov, is

the Tesla approach for autonomy is taking human knowledge.

In fact, the whole idea is to approximate

human driving, right?

That is the whole damn thing.

Now,

depending on your approach and the technology, you can do like what's called an end-to-end approach, or you can look at, okay, perception, prediction, planning, and control, which are like these four modules that sort of, you, you, you sort of engineer, if that makes sense.

But it's approximating human driving to do it.

The difference is that,

you know, I think Elon's taken

almost a more human approach, which is like, I've got two eyes.

Why can't my car?

Why can't my car do it like a human?

Like, I don't have any LIDAR spinning around on my head as a human.

Why can't my car?

So it's kind of interesting.

He's sort of taking what you're saying, Chamoth, on the computation side, because Hardware 5 is coming out on Tesla probably next year, which is going to make a big difference in what FSD can do.

That's the compute side you're talking about, but then he is approximating human.

Yeah, I just meant that, you know, other than the first versions of FSD, which I think Andre talked about, Andre Karpathy talked about, you know, they're not really so

reliant anymore on human labeling per se, right?

So that's that

interference.

And then

the other crazy thing that he said, subsequent versions of Grok

are not going to be trained on any traditional data set that exists in the wild.

The cumulative sum of human knowledge has been exhausted in AI training.

That happened basically last year.

And so the only way to then supplement that is with synthetic data where the AI creates, it'll sort of write an essay or it'll come up with a thesis and then it will grade itself and

sort of go through this process of self-learning with synthetic data.

He said that he's going to have agents creating synthetic data from scratch that then drive all the training, which I just think is, it's crazy.

Just explain this concept one more time in the bitter lesson.

Hand coding heuristics into the computer and saying, hey, here are specific openings in chess.

Yeah, use chess, right?

Yes, you're hand coding specific examples of openings in there, end games, et cetera, versus just saying, play every possible game and here's every game we have.

So here's youth.

Yeah, so the two approaches would be, let's say, like Travis and I were building competing versions of a chess solver.

And Travis's approach would say, I'm just going to define the chessboard.

I'm going to give the players certain boundaries in which they can move, right?

So the bishop can only move diagonally and there's a couple of boundary conditions.

And I'm going to create a reward function and I'm just going to let the thing self-learn and self-play.

That's his version.

And then what happens is when you map out every single permutation,

when you go and play Keith, who's the best chess player in the world, what you're doing at that point is saying, okay, Keith.

made this move.

So you search for what Keith's move is and you have a distribution of the best moves that you could make in response or vice versa.

That was the cutting-edge approach.

The different approach, which is more, you know, what people would think is more quote-unquote elegant and less brute force, would be, Jason, for you and I to sit there and say, okay, if Keith moves here, we should do this.

We should do this specific variation of the Sicilian defense.

And it's too much human knowledge.

And I think what it turned out was there was a psychological need for humans to believe we were part of the answer.

But what this is showing is because of Moore's Law and because of general computation, it's just not necessary.

You just have to let go, give up control.

And that's very hard for some people.

And for others, it's also very hard in some circumstances where a car is driving down the road and it's learning in that process, which is why you need a safety driver.

And I think Elon made the right decision to put one in there.

Keith, your thoughts.

Yeah, a couple of points.

It's not quite that binary, Chamoff.

I generally agree with your arc, but like if you think about LLMs being the most important unlock in AI, LLMs are all trained on human writing.

So someone wrote every piece of data that every LLM used, a human wrote at some point in history.

So yes, it's true that they've shocked everybody, including OpenAI's original team, on the implications, the broad implications, the general applicability to almost every problem.

But it's not like there was some tablets floating in space that weren't drafted by humans that we've trained on.

As you get in non-LLM-based models, you may be totally right, but almost no one's really using non-LLM based models at scale.

On driving specifically, Travis is totally right that humans are actually really good drivers, except when they get distracted.

They get distracted by drugs or alcohol.

They get distracted by being tired.

They get distracted by turning the radio.

They get distracted by chatting with their passenger.

So trading against human behaviors actually turned out to be a great decision because for whatever sort of Darwinistic reasons, humans are pretty ideal drivers.

And so you don't have to reason from first principles.

This is a much better path.

And I think, again, there may be a broad

sort of lesson there.

The most important thing, I think, as a VC that you said is we've been debating for years, should we invest in companies like Scale or Mercor or any of these surge?

The truth is, I think there's a very short half-life on human label data.

And so everybody who's investing in these companies, just looking at revenue traction,

really didn't understand that there may be a year, two years,

three years max when anybody uses human label data for maybe anything.

Because we hit the end of human knowledge or just the collection of it.

99% done.

Or you train on, you train on it so well that you don't need to label anymore.

Like the machines know how to label as good or better than a human.

And so like we're seeing this in the self-driving space is labeling was huge, right?

You would have a three-dimensional sort of scene that's created by video plus LiDAR, let's say.

Okay, I have to label all of these essentially what become boxes, like I've identified objects.

You're, you're, some of the players in the, in the autonomous software space, autonomous vehicle software space, are no longer doing any labeling because the machines are doing it all,

just broadly.

It'll just be built into the chipset that this is a stop sign.

Like it's like, we know what a stop sign is.

We don't need the millionth time for somebody to make captchas.

Like you're like, find the stop sign or what's the traffic light?

And eventually the machines are just way better than humans at identifying these things.

For you to be very practical, when you see a stop sign, you don't have to identify that it's a stop sign.

You just see that every human, when they encounter a stop sign, 99.9% of the time, they hit a break.

And they never, so nobody actually knows it's a stop sign.

It's just that hit a break when you see something that looks like this object.

It's just a vibe.

Yeah, it's a vibe.

I would just say that that's like intuited knowledge versus like the expressly labeled human knowledge.

The question for me is, if everybody was so reliant on human labeling initially, if you're an investor now,

when you see these Grok 4 results,

how do you make an investment decision that's not purely levered to just computation?

So if you look at these results, does it mean that

there's 300 to 1,000 basis points of lag

between just letting the computers vibe itself to the answer versus interjecting ourselves?

If interjecting ourselves slows us down by 300 to 1,000 basis points per successive iteration, then over two or three iterations, you've totally lost.

So, what does it mean for everybody that's not Grok

when they wake up today and they have to decide, how do I change my strategy or double down?

I think,

look, I'm not in the investment game, but if I were, it would be all about scientific breakthrough.

So I sometimes get in this place where I'm looking, I'm going down a path.

You know, I'll be up at four or five in the morning.

My day hasn't quite started, but I'm not sleeping anymore.

And I'll start going, like, I'll be on Quora and see some cool quantum physics question or something else I'm looking into.

And I'll go down this thread with with GPT or Grok,

and I'll start to get to the edge of what's known in quantum physics.

And then I'm doing the equivalent of vibe coding, except it's vibe physics.

And we're approaching what's known, and I'm trying to poke and see if there's breakthroughs to be had.

And I've gotten pretty damn close to some interesting breakthroughs just doing that.

And I, you know, I pinged, I pinged Elon at some point.

I'm just like, dude, if I'm, if I'm doing this and I'm super amateur hour

physics enthusiast, like, what about all those PhD students and postdocs that are super legit using this tool?

And this is pre-Grok 4.

Now with Grok4, like, like, there's a lot of mistakes I was seeing Grok make

that then I would correct and we would talk about it.

Grok 4 could be this place where breakthroughs are actually happening, new breakthroughs.

So if I'm investing in this space, I would be like,

who's got the edge on scientific breakthroughs and the application layer on top of these foundational models that orients that direction?

Is your perception that the LLMs are actually starting to get to the reasoning level that they'll come up with a novel concept theory and have that breakthrough or that we're kind of reading into it and it's just trying random stuffs at the margins?

uh, or maybe it doesn't happen.

No, no, no.

So, what I what I've seen, and again, I haven't used Grok4, I tried to use it early this morning, but for some reason, I couldn't do it on my on my app.

But

so, let's say we're talking Grok 3 and existing chat GPT as it is.

No, it cannot come up with the new idea.

These things are so wedded to what is known, and they're so like even when I come up with a new idea, I have to really, it's like pulling a donkey, sort of you see, you're pulling it it because it doesn't want to break conventional wisdom.

It's like really adhering to conventional wisdom.

You're pulling it out and then eventually it goes, oh shit, you got something.

But then when it says that, when it says that, then you have to, you have to go, okay, it said that, but I'm not sure.

Like you have to double and triple check to make sure that you really got something.

To your point, when these models are fully divorced from having to learn on the known world and instead can just learn synthetically.

Yeah.

Then everything gets flips upside down to what is the best hypothesis you have, or what is the best question you could just give it some problem and it would just figure it out.

So, where I go on this one, guys, is it's all about the scientific method,

right?

If you get, if you have an LLM or foundational model of some kind that is the best in the world of the scientific method,

game the F over.

You basically just light up more GPUs and you just got like a thousand more PhD students working for you.

Keith, you're nodding your head here.

I agree with that.

I think that's fantastic because the scientific method also, the faster it is, the more you, when you have a hypothesis, the faster you get a response, you're more likely to dive in and dive in and dive in recursively and recursively.

And every lag, every millisecond lag causes you to like lose your train of thought, sort of, so to speak.

So you get the benefits that Travis alluding to, plus speed, and you go places you know everybody else.

This happens all the time when you run a company and you're doing like analytics and you have a tool that allows you to constantly query quickly, quickly, quickly, double click, triple click.

You get to answers that you never get to, even if there's even a second or two second or three second delay, let alone sending it to a human.

Secondly, where you actually see this today, it's already happening.

If you look at foundational models that just apply to science, There's lots of things about the human body, let's say in health biology, that we humans don't actually understand all the connections.

Like, why do we do X?

Why do some people get cancer?

Why do other people not get cancer?

Why does the brain work this way?

Models trained solely on science tend to expose connections that no human has ever had before.

And that's because like the raw material is there, and we only have a conscious awareness of call it 110%.

But when you apply it to other human domains where you're training on human sort of data, human-produced data, human-produced output, they're limited to that output.

So I think you just take the science and apply it writ large and you're going to wind up finding things that no human has ever thought before.

And it's the thing about science, though, is that it's the hypothesis that you then have to test in the physical world.

So

you're like, okay, have you got this hive mind, this like, you know,

this

computation engine, this brain of sorts.

You wanted to say consciousness, but you stopped yourself.

Yeah, there's an idea.

I was like, how do I describe it?

It's a big C word, consciousness.

But you need to be able to test in the physical world.

So you could imagine a physical lab connected to one of these systems where then you could say, okay, like if it's a chemistry experiment, you could do chemistry experiments or physics.

You get the idea.

What could go wrong?

It would be, it's, yeah, no big deal.

It's going to be fine.

Okay.

So, but, but this is where it goes, because if you have a scientific method machine, you still have to be able to test your hypothesis.

You have to go through the scientific method.

And the verification.

Yeah, exactly.

Yeah.

Wow.

It's kind of mind-blowing.

It reminds me of- It's really mind-blowing.

Remember, I don't know if you guys remember dark matter and like the discovery of it and everything.

And as explained to me by Lisa Randall, you know, the discovery was made not by knowing there was dark matter there and observing it, but observing there was something, you know, gravitational forces around this other matter.

And then they said, well, wait, what's causing that?

And that's where they found dark matter.

So these ideas, you know, the idea that an LLM could actually do that, come up with something so novel is, it doesn't, it feels like we might be right there, right?

Like we're kind of on the cusp of it.

One of the seven most difficult problems in math or the most important problems in math is proving a general solution to this thing called Mavier-Stokes, which is basically like viscous fluid dynamics and conservation of mass.

We use it every day in the design of everything.

You know what?

It hasn't been proved.

Isn't that the craziest thing where you're just like, how is this even possible?

We use it to design airplanes, to design design everything.

It hasn't been proved.

And so you could just point a computer at this thing and you would unlock all these incredible mysteries of the universe.

And we would probably find completely different propulsion systems.

We could probably do things that we didn't think were possible, teleportation.

I mean, who knows what's possible?

But remember, remember, you know, how Elon talks about rock and about AI generally, is about why are we here?

What is the purpose?

Meaning of the universe.

What is the meaning of the universe?

How does it work?

And a sort of fierce truth-seeking mechanism there.

Let me ask you a question, Keith, Travis, Jason.

If you guys were running Grok 4,

that'd be so much fun.

How do you judo flip OpenAI?

Because they

are marching steadfastly towards a billion Mao, then a billion DAO.

It's a juggernaut.

So how do you use the better product

in a moment to judo flip

the less better product?

Look, yeah, I mean, here's the thing, right?

So you do the Elon Way.

So you get a bunch of missionary, like full-on missionary engineers that work twice as hard.

And you have a culture that is ultra fierce truth-seeking.

And

you don't get caught up in politics bureaucracy bs

and you just you go for it and and i think you know that that's where you know and then you go wow scientific breakthrough scientific method like you start winning on truth and that will start i believe that will start to give the product awesomeness of open ai a run for its money

but like the product of open ai the product department those guys are crushing.

They're crushing.

They're really good.

They're not only ahead of the game, but they feel like it just, they're just leading in a lot of different ways.

But if you are better at truth, you will eventually, you'll eventually have an AI product manager.

And on a technical basis, too, people forget how good Elon is at factories and physical, real world things.

What he did standing up Colossus made like Jensen Juan was like, how is this possible that you did this?

right?

So pressing that, his ability to build factories, and he said many times, like the factory is the product of Tesla.

It's not the cars that come out of the factory or the batteries.

It's the factory itself.

So if he can keep solving the energy problem with solar on one side and batteries and standing up, you know, Colossus 2, 3, 4, 5, he's going to have a massive advantage there.

on top of Travis, you know, the missionary individuals, which by the way, was what he backed before Sam Altman corrupted the original missionary basis of open AI and made it closed AI.

And

nothing derogatory towards him, but he did hoodwink and stab Elon in the back.

It's not nothing personal.

I mean, he just screwed him over.

Would you say he bamboozled him?

He bamboozled him, screwed him, hoodwinked him.

You know,

pick your term here, but he did it.

He didn't dirty.

The original mission was to be a missionary to open source all this content.

That's the other piece I think is a wild card.

And then I'll measure Sitting Keith's position, but open sourcing some of this could have profound ramifications.

I think open sourcing the self-driving data could have a really profound impact.

Elon wanted to do something really disruptive, like he open sources patents for

charging.

If he open sourced the data set and self-driving, does anybody have the ability to produce robo-taxis at the scale he can do it?

I don't think so.

Well, if Travis's hypothesis is true, then everybody will.

Well, everybody will what?

Sorry, everybody will what, Shimon?

If you have access to the money that buys the compute, everyone could solve that problem.

What's the hardware piece I'm talking about?

Which problem?

He said if he published all the FSD data, could somebody build an autonomous vehicle?

Well, yes, but could somebody produce 100 million robo-taxis from a factory with batteries in them?

Okay, no, that's a different, that's a different

thing I'm saying.

And not really, because last time I was a guest on, you know, all, and we talked about vertical integration.

Products really require vertical integration.

So ultimately, you have a self-driving something that is custom built for knowing it's going to be self-driving.

And it interacts differently.

The cost structure is different.

The controls are different.

The seating is different.

Everything you build a product taking advantage of where in the staff you have the most competitive advantage, but then you leverage that and it reinforces.

It's still why, like, Apple, despite missing the AI wave, still a pretty good company from any empirical standpoint.

I mean, like their performance is absolutely miserable on the most important technology through the last 70 years, but the company is still alive and still worth trillions of dollars because it's vertically integrated.

Open AI, per your point, they do have a good product team and they need to stay ahead of the product level because they can't compete on the factory level.

The way to stay ahead of the product level is shipping a device.

got to ship the device.

It's got to be good.

It's got to be right.

It's got to be the right form factor.

It's got to do things for humans that are unexpected.

But then if they do that, they're like Apple Plus AI.

Shamath, what's the paper you were talking about before?

What was the name of it again?

The Bitter Lesson.

It could apply to autonomous driving is right now.

It's still like, hey, how do I drive like a human?

We talked about that.

But the leapfrog moment here could be like, hey, drive a car, make sure it's efficient.

Don't hit anybody.

And just simulate that quadrillion times and it's all good.

Right.

But right now, we're still trying to drive like humans because we don't have enough data and therefore can't do enough compute.

That's the global lesson, by the way.

Tramath, you're totally right.

Conceptual, you know, the blog post is right, but that's only true when you have enough data.

And depending upon the use case, the level of data you need may not be possible for years, decades, and you may need to hack your way there through human interactions.

Physical world AI is

lacking in data.

And so you just try to approximate humans.

I don't know if you guys have seen this.

In related news, news, OpenAI and Perplexity are going after the browser.

Perplexity launched Comet for their $200 a month tier.

I actually downloaded it.

I'll show it to you in a second.

But this is a really interesting category.

It's something developers can do already, and they do it all the time.

But having your browser connected to agents.

lets you do really interesting things.

I'll show you an example here that I just fired off while we're talking.

So I just asked it, hey, give me the best flights from United Airlines and

business class from New York City, from San Francisco to New York City.

It does some searches, but what you see here is it's popped up a browser window and it's actually doing that work.

And you can see the steps it's using.

And then I can actually open that browser window and watch it do that.

This is just a screenshot of it.

And it will open multiple of these.

So you could, I was doing a search the other day saying, hey, tell me all the autobiographies I haven't bought on Amazon, put them into my

shopping cart and summarize each of them because I like biographies and like doing it here.

And when it did this last time, it put my flight into

like, and I was logged in under my account, and it basically put it into my account in the checkout.

So again, this isn't like if you're a developer, you do this all day long.

But this really seems to be a new product category.

I'm curious if you guys have played with it yet.

And then what your thoughts are on having an Agentic browser like this available to you to be doing these tasks

in real time.

You can also connect, obviously, your Gmail, your calendar to it.

So I did a

search, tell me every restaurant I've been to, and then put it by city.

And then I was going to open my open table and then pull that data as well.

What's interesting about this,

Keith, and I know you're a product guy and done a lot of product work.

I'm curious your thoughts on it, is

you don't have to do this in the cloud you're authenticated already into a lot of your accounts nor do you have to worry about being blocked by these services because it doesn't look like a scraper or a bot it just it's your browser doing the work your thoughts on this have you played with that at all yeah i think it's a great hail mary attempt by perplexity i think opting something like this perplexity is toast like for the stat about chat gpt going to a billion users like it's becoming the verb you know that the way you describe using ai for a normal consumer there's nothing left of perplexity if they can't pull this off.

So it's a great idea because like the history of like consumer technology companies is whoever's up has uphill ground, like in a military sense, whoever is first has a lot of control.

This is actually what Google should be doing, truthfully.

Like I think Google is also Google's search quality search is toast.

And since they have Chrome and they theoretically have a quality team in Gemini, they should be putting these two things together and hoping to compete with ChatGPT.

They're going to lose the search game.

Like assets that are best at Google right now have nothing to do with search.

It's every other product is the only thing that's going to save that company if they can figure out how to use them.

Travis, your thoughts on this category, anything come to mind for you in terms of,

you know, feature sets that would be extraordinary here?

I know you like to think about products and the consumer experience.

It's really interesting.

So, you know, I've been spending, as you guys know, I've been spending my time on real estate and construction and robotics.

And so I've been out of this kind of consumer software game for a long time.

But it's super interesting over the last six months, there have been a number of consumer software CEOs.

Like when I hang out with them or whatever, they're like,

yo, how are we going to, how are we going to keep doing what we do when the agents take over?

Yeah.

The paradigm shift is so profound that the idea that you would visit a web page goes away and you're just in a chat dying.

You know, you have an agent that's just taking care of your flights for you.

So I, I kind of, I think there's a leapfrog over that.

I think

it's just like you tell something, yo, I want to go to New York.

Can you, you know, I'm sort of looking at this time range.

Can you just go find something I'm probably going to like and give me a couple of options?

Yeah.

And it's just a whole, you have an interface and then,

you know,

is perplex, is this thing that you just showed on perplexity, is that the interface?

Or do I just have an agent that just goes and does everything for me?

And is this the start of that?

You know, I just haven't spent enough time.

I do know that every consumer software CEO

that has an app in the app store is tripping.

They're tripping right now.

And I mean big boys.

I mean guys with real stuff.

And sometimes I'm doing like almost like therapy sessions with them.

I'm like, it's going to be fine.

You actually have stuff.

You have a moat.

You have real stuff that's of value.

They can't replace it with an agent.

So you're lying to them.

You're doing hospice care and you're telling them everything's going to be okay, but the patient's not just analyzing options on Robinhood.

He's like, yeah, yeah, tell me more.

Tell me more.

There's certain things that are protected and there's certain things that aren't.

That's all.

Well, let's talk about that because you and I are old enough to remember General Magic.

This vision was out there a long time ago with personal digital assistants, and you would just talk to an agent, it would go do this for you.

This feels like a step to that, where it does all the work for you, presents you the final moment and says, approve.

So look like a concierge or a butler.

I think what you're describing is what we want, but I think more specifically for today, Keith and Travis totally nail it.

Look, I think building a browser is an absolutely stupid capital allocation decision.

Just totally stupid and unjustifiable in 2025.

Specifically for Perplexity, I think their path to building a legacy business is to replace Bloomberg.

Everything that they've done in financial information and financial data in going beyond the model has been excellent.

As somebody who's paid $25,000 to Bloomberg for many years,

The terminal is atrocious.

It's terrible.

It's not very good.

It's very limited.

And

anybody that could build a better product would take over a $100 billion enterprise because I think it's there for the taking.

I wish that perplexity would double and triple down on that.

And so when you see this kind of random sprawl.

Let's do it, Jamas.

Let's just go do it.

When you do the random sprawl, I think it doesn't work, but I just want to say, like,

A browser is like the dumbest thing to build in 2025 because in a world of agents, what is a browser?

It's a glorified markup reader.

It's like handling HTML.

It's handling CSS and JavaScript.

It's doing some networking.

It's doing some security.

It's doing some rendering.

But it's like, this is all under the water type stuff.

I get it that we had to deal with all that nonsense in 1998 to try Lycos or Google for the first time.

But in 2025,

there's something that you just speak to, and eventually there's probably something that's in your brain, which you just think, and it just doesn't.

You're thinking,

I need a flight to JFK.

Or at the maximum today, in a very elegant, beautiful search bar, you type in, get me a flight, and it already knows what to do.

Keith, in some ways, this is a step towards that ultimate vision.

So you'd think it's worth it to, you know, sort of perplexity to make this waypoint, perhaps, if you look at it as a waypoint between the ultimate vision, which is a command line, an earpiece, a hotel.

How do you get distribution, Jason, for the 19th web browser in 2025?

Well, yeah, that is a challenge.

And I think most people are speculating Apple, which has a lot of users, might buy Perplexity or do a deal with Perplexity and give them that distribution because of the Justice Department case against Google.

So there's been a lot of speculation about that.

But Keith, what do you think?

Well, I don't think they'd buy anything worth it.

Like, what is Apple going to get?

I mean, if you continue this failed strategy of Apple,

Apple has missed every possible window on AI and continues to miss it.

And it has cultural,

I think the CEO has challenges.

I think culturally they have challenges.

I think they have infrastructure challenges.

So it's not an easy fix, but buying complexity is not going to help.

Like Java strategy is actually a pretty coherent one for perplexity, quote, perplexity.

So I think that

pick a vertical and own a strategy.

Not a bad idea, especially because you need unique data sources.

Some of those data sources may or may not license their data to OpenAI.

So you can do some clever things there.

But I don't think there's any residual value that Apple would get out of perplexity, except there's some product taste, but what are you going to spend like a billion dollars for product taste?

I mean, Mark's spending hundreds of millions of dollars, hundreds of billions of dollars or whatever he's spending these days.

And, you know, Grok, if anything, Grok force shows that Mark really doesn't need to just spend money to build a whole new team because everything they've done in AI has also missed the boat.

Well, I mean, Keith, the way you phrase it there almost makes it worth it for Apple to throw a Hail Mary, have a team with some taste, because that's how they tend to do things, is something that is elegant.

And why not just throw your search to it, throw 10 billion at least?

What's elegant would be if there's a bunch of agents and just a chat box.

Seeing a bunch of visual diarrhea is not elegant.

It's lazy.

Shamash on our little Bloomberg clone.

I'll give you naming rights.

So you can call it.

Polyhapatia.

So, hey,

can somebody bring up the polyhapatia?

You know what's so funny?

TK, listen, we were trying to do a screen of companies,

and it maxes out at five companies on a specific type of screen where you're like, you're trying to compare stock price to EBITDA, and you're like, okay, I can only choose five, I guess.

So, which five should I choose?

What font was on, right?

Like two episodes ago, he was like, I can't pull this up.

It's limited to six companies.

Dude,

so what do people use, Blue?

They use it for the messaging.

Now, like my team has traded huge positions via text message on Bloomberg.

So there is something very valuable there.

But the core usability and the core UI of that company has not evolved.

I have my contribution.

And Perplexity is very good at that, by the way.

They do a very good job.

I got a new domain name, Travis.

Let this one just sink in here.

This is my way to weasel my way into the deal.

Begin.com.

Begin.com.

You own that, don't you?

I do.

I'm I'm just a little, I sniped some good ones once in a while.

I got begin.com and I got annotated.com.

Those are my two little.

You're like one of these old people that show up at those.

Oh, like the road show and then they almost like that.

The antique road show.

And you're like, oh, I have this thing that I bought in 1845.

Guys, Jason, Jason is, Jason is the daddy and GoDaddy.

Okay.

That's what it is.

Who's your daddy?

Hey, speaking of daddy, let's go on to our next story.

it's now the right time for a third party elon seems to think so last week he announced that x he would be creating a new political party i'll let you decide who daddy is in this one

he said quote when it comes to bankrupting our country with waste and graft we live in a one-party system not a democracy he's not yet outlined a uh a platform for the American party.

We talked about it here last week.

I listed four core values, which seem to get a good reaction on X.

Fiscal responsibility/slash Doge, sustainable energy and dominance in that.

Manufacturing in the U.S., which Elon has done single-handedly here.

Pronatalism, which I think is a passion project for him.

And Shabbat, you punched it up with the fifth technological excellence.

According to Polymarket, 55% chance that Elon registers the American Party by the end of the year.

And, you know, one thing I was trying to figure out is just how unpopular are these candidates

and these political parties.

This is a very interesting chart that I think we can have a great conversation around.

It turns out we used to love our presidents.

If you look here from Kennedy at 83%, his highest approval rating, his lowest was 56%.

That was his lowest approval rating.

So he operated in a very high band.

Look at Bush 2 during after 9-11.

92% was his peak.

His lowest was 19, right?

Wartime president.

But then you get to Trump 1, Biden, Biden, and Trump 2, historically low, high approval.

They're high watermark, 49 for Trump 1, 63 for Biden, one of one, and then 47 for Trump 2.

And their lowest 29, 31, 40.

So maybe it is time for a third-party candidate.

Let's discuss it, boys.

I have no idea how to read this graph.

This is the worst.

I'm like, what is happening here?

This is the worst formatted chart.

This is a confusing chart, but well, the reason I'm putting it up is for debate.

So you should be saying thank you for debating that it's creating great debate.

Why did you put it up?

Here's another one.

Gallipol Americans desire for a viable third party, 63% in 2023.

So

it's bumping along an all-time high.

Okay, I'm really concentrating on this one.

Okay.

Anyway, I'm going to stop there.

What's the gray?

I'm going to let you

different presents during that time period and how popular the party is.

Let's stop here.

This is a good, this is a good place to stop.

I just blew a GPU.

Yeah, a couple points.

Yes.

The idea of Elon creating a third party is for any other human being, like absolutely absurd and ridiculous.

Elon has obviously done incredible things.

So dismissing anything he's touching is a bad idea.

However, I think the best metaphor I've seen is it's a little bit like Michael Jordan tried to play baseball.

He became a replacement level baseball player, which actually really hard to do, by the way.

Elon is probably a replacement level politician.

um he's michael jordan for entrepreneurial stuff but the third party stuff is not going to work first of all um

there that chart is misleading it's a flaw of average it was badly designed and it's a flaw of average of policy trump is incredibly popular among republicans he actually has the highest approval rate of any republican ever measured in recorded history it's 95 reagan was peaked out at 93

It's just Democrats don't like him, which is perfectly fine.

Being polarizing is an ingredient to being successful, including with people on the show.

Like the point of accomplishing things in the world is you don't really care what half the world thinks.

You need to make sure that there's a lot of people who like you and really approve and are enthusiastic about what you do.

And Trump is about as popular with his party as anybody's ever been ever, period.

No exceptions.

Secondly, there's MAGA has kind of already

changed the Republican Party.

Trump is sort of like a third-party takeover of the Republican Party.

And so it's kind of already happened.

And maybe you can do this every 20 years or 30 years.

I don't think you can have like this kind of transformation on one party within a too compressed period of time for a lot of reasons.

Third is really smart parties absorb.

The lesson of political science.

Unfortunately, I studied political science.

I wasted kind of my college years.

And instead of saying CS and, you know, maybe then I'd be coding stuff and doing physics like Travis.

But one thing I did learn is smart parties absorb the best ideas of third parties.

So the oxygen is usually not there because there's a Darmanistic evolution.

If you get traction on an idea, it's really easy to conscript some of those ideas and take away the momentum.

No third party candidate that's a true like third party has won a Senate seat since 1970.

And that's actually Bill Buckley's brother.

And so he had some name ID.

The other thing, Elon, I think, is missing and the proponents of what he's doing is people vote not just for ideas, they vote for people.

It's a combination.

The product is

what do you believe?

And who are you?

And you can't divorce the two.

Trump is a person, and that generates a lot of enthusiasm.

And it's one of the reasons why he has challenges in midterms because he's not on the ballot.

His ideas may be on the ballot, but he is not specifically on the ballot.

So, unless, because Elon can't be the figurehead of the party, he literally can't constitutionally, you need a face that's a person, Obama, a Clinton.

Like, there's reasons why people resonate with Reagan.

Without

that personality, specific ideas just are not going to galvanize the American people.

Okay.

So, the counter to that, and what people believe he's going to try to do is win a couple of seats in the House, Travis.

Win maybe one or two Senate seats.

If you were to do that, those things are pretty affordable to back a couple of million dollars for a House race.

Senate, maybe 25 million.

If Elon puts, I don't know, 250 million to work every two years, which he, I think he put 280 million to work on the last one, he could kind of create the Joe Manchin moment and he could build a caucus, a platform, Grover-Norquist kind of pledge along these lines.

So what do you think of that?

If he's not going to create a viable third-party presidential candidate, could he, Travis, pick off a couple of Senate seats, pick off a couple of congressional seats?

Okay, so first, I have this axiom that I'm making up right now.

Okay.

Okay.

It's called Elon is almost always right.

Okay.

All right.

Elon was right about everything.

Seriously, let's just be real.

And, like,

honestly, the things he's upset about and that he's riled up about, especially when you look at the deficit, like, man, I am right on board that train.

Part one, part two:

we've never had somebody with this kind of capital that can be a quote-unquote party boss outside of the system, right?

And

there's a lot of people that agree with the types of things he's saying.

And he knows how to draw, you know,

Elon in his own right kind of has a populist vibe.

Like he does his thing.

And he's turned X into what it is.

And

he's a big part of X.

And so I think it's the, I think it's great.

And honestly, there's, there's the moves you can make on Senate and House and just having a few folks and then being you being levers then to get the things you want done.

That's part one.

And then part two of that is the threat of that happening can make good things happen separately, even if it doesn't go all the way.

I just love it.

I'm on the email train.

Yeah, I'm in love with this role for Elon more than picking a party because he's picking a very specific platform that I think resonates with folks, which is just balance the budget, don't put us in so much debt, and let's have some sustainable energy.

You know, job done, great job done.

The problem with that is like he's actually wrong about the reason why we have a deficit or debt.

It's not because we're undertaxed, it's we're massively overspending.

If we just no, I think he believes we're overspending, but they should have been supporting the last

beautiful bill.

Because if you just held federal spending to 2019 levels, so 2019 is not like decades ago, literally with our current tax revenues, we would be in a surplus.

500 billion.

Yeah.

So

all we need to do is cut spending.

Now, I admit that there's- Why didn't that happen with the big beer?

Well, though.

So this is where details do matter.

I think there is a willingness and a discipline problem on both parties.

And I think maybe he can help fix that.

The second thing is that we have these arcane rules, particularly in the Senate, that you need 60 votes in many ways to cut things, except through very hacky methods.

And that's a reality.

So the best thing truthfully you could do is help get a Republican Party to 60 votes.

And then

in theory, he could be absolutely furious if you didn't cut back to 2019 levels.

But it's very tricky, or you can just overrule.

Like this, the filibuster is an artifact of history.

And at some point, some majority leader is just going to say, we're done with the filibuster and just steamroll steamroll through all the cuts at 50 or 51 votes, which you can do.

There's no constitutional right to a filibuster.

It is an artifact of centuries of American history.

And at some point, it's going to go away.

So maybe the time is now.

Maybe we should just fix everything now.

I think you're exactly right.

I think that the filibuster, it's just a matter of time.

I think it's on borrowed time.

And I think in a world.

where it is on borrowed time, Jason, I think your path is probably the one that gives the American Party, if it does come into existence, the most leverage, which is if you control three to five independent candidates, you gain substantial leverage.

I just want to take a step back and just note something.

I don't know if you guys know this, but the only reason we're even having this conversation or this is even possible is because in 2023,

the FEC, Federal Elections Commission, they actually released guidance and they changed a bunch of rules.

And the big change that they made then was it allowed super PACs to do a lot more than just run ads.

Up until that point, all you could do if you were a super PAC is just basically run advertising, television and radio.

I guess online as well.

But what they were allowed to do starting in 23 was they were allowed to fund ground operations.

They were allowed to do things like door knocking, phone banking, you know, get out the vote.

So in other words, what happened was a super PAC became more like a full campaign machine.

And Trump showed the blueprint of using a super PAC, specifically his, to win the presidential election.

So he was able to fund this massive ground game.

He built infrastructure across the swing states.

He was obviously incredibly effective.

And now that playbook can actually be used by other folks.

And so to the extent that Elon decides to use those changed FEC rules, Jason, I think what you said is the only path.

But

I just wanted to double click on Keith's point because it's so important.

I do think the filibuster is going to go away.

And it is because

the arcaneness of these rules, having to do a reconciliation build and needing a supermajority, a veto-proof supermajority in any other case, it just means that nothing gets done.

And I think somebody will eventually get impatient and just steamroll this thing.

We've never had so many people say they feel politically homeless as we did the last two cycles.

And that includes many people on this podcast, people in our friend circle.

And I think just the idea that Elon could create a platform that people could opt into and support, just the existence of that would make the other two parties get their act together.

By the way, I think that's what we need is a little bit of a stick there and a carrot.

Hey, if you don't control spending, there's this third option.

And if Travis and I are in it, and Keith, I know you'll never leave the Republican Party, but Shamath, you know, you're probably set where you want to be right now.

But I can tell you, we go through our top 10, 20 frank list, out of those, 50% will join New Orleans Party.

Well,

the other thing, Jason, that Keith said, which I think is really important, is

if he were to run people,

I think they have to transcend politics and policy.

And I think they need to be straight up bosses, people that have enormous name recognition, so that effectively what you're voting is a name and not an agenda.

Equivalent to, I think, what happened to Schwarzenegger when he ran.

He ran on an enormous amount of name recognition in the Great Davis recall.

He didn't run on the platform, which any of us could mention.

J.D.

Vance had this great book, Capture People's Imagination.

He's an incredible speaker.

He pisses off a third or two-thirds of the country, depending on where you are in the country.

But you can't ignore him.

I think Elon can find 10 J.D.

Vance-type characters and back them fairly easily.

He is a magnet for talent.

People will will line up.

I have been contacted by high-profile people.

I was actually thinking of running.

Can you put me in touch with Elon?

I was thinking more like actors and sports stars, meaning where they just come with their own inbuilt distribution.

Like, I think you almost have to rank X followers and Instagram followers and do a join and say, okay, these are, do you know what I mean?

Like, I think it's a totally different place.

It's painful, guys.

It's painful.

Like, let's not get more celebrities as politicians.

Like, let's get like people who've led large, large efforts, large initiatives, complex things.

Ideally, but they still have to communicate, right, Keith?

They have to be able to communicate on a podcast.

That's the new platform.

If they can't spend two hours, three hours chopping it up on a podcast like this or Joe Roman, you know, that's Kamala's.

The reason she couldn't even contend was because she couldn't hang for two hours in an intellectual discussion.

If you can't hang, you're out in today's political arena.

It'll be interesting to see if he can tune his algorithm for talent, which is epic,

to tune for politics because it's a slightly different audience.

But if you can tune the algorithm and quality, that might work.

I think you can win a few house races.

I think that's doable.

I don't think you can win a Senate race.

Well, there it is, Elon.

Keith doesn't think you can win a Senate race, but he thinks you win a couple of congressional ones.

Thanks for giving him the motivation, Keith.

I appreciate it.

I'm sure he's going to win.

This is the biggest mistake you've ever made.

He's not going to win too.

People in the Republican Party right now are going, oh, no, don't poke the tiger.

Listen, speaking.

But that's how Trump got into politics.

So I don't want to be Obama here.

You just know politics.

Elon, right?

Yeah.

Congratulations.

All right.

Listen, SCODIS made a big decision here.

This is a really important decision.

They've sided with Trump for plans for federal workforce rifts, reductions in workforce, for those of you who don't know.

As you know, Elon, Trump, they wanted to downsize the 3 million people who are federal employees.

This is just federal employees we're talking about.

We're not talking about military and we're not talking about state and city.

That's tens of millions of additional people.

If you remember, Trump issued this executive order back in February when we got in office implementing the president's Doge Workforce Optimization Initiative.

And he asked all the federal agencies, hey, just prepare a RIF for their departments consistent with applicable laws, was part of this EO.

Okay.

In April, the American Federation of Government Employees, AFGE, sued the Trump administration, saying the president must consult Congress on large-scale workforce changes.

This is a key debate because the Congress, as you know, has power of the purse, they set up the money, but the president and the executive branch have to execute on that.

And that's what the key is here.

So they accused Trump of violating the separation of powers under the Constitution Act.

AFGE has 820,000 members.

In May, a San Francisco-based federal judge sided with the unions blocking the executive order.

The judge, who was appointed by Clinton, said any reduction in the federal workforce must be authorized by Congress.

This is a key issue.

And the White House submitted an emergency appeal, yada, yada.

Eight of nine Supreme Court justices sided with the White House in overturning this block.

And so the reasoning, it's very likely the White House will win the argument of the executive order.

They have the right to prepare a RIF.

The question is, can they actually execute on that RIF?

And who has that power, Chamoff?

Does the power reside with the president to make large-scale or riffs, or do they have to consult Congress first?

Your thoughts on this issue?

It's an incredibly important ruling, incredibly right.

I think President Trump should have absolute leeway to decide

how the people that report to him act and do their job.

If you take a step back, Jason, there are more than 2,000 federal agencies.

Employees plus contractors, I think, number almost 3 million people.

If you put 3 million people into 2,000 agencies and then you give them

very poor and outdated technology, which unfortunately most of the government operates on, what are you going to get?

You're going to get incredibly slow processes.

You're going to get

a lot of checking and double checking.

And you're going to ultimately just get a lot of regulations because they're trying to do what they think is the right job.

So since 1993, what have we seen?

Regulations have gotten out of control.

It's like 100,000 new rules per some number of months.

Like it's just crazy.

So eventually we all succumb.

to an infinite number of rules that we all end up violating and not even know it.

So if the CEO of the United States, President Trump, isn't allowed to fire people, then all of that stuff just compounds.

So, I think that this is a really important

thing that just happened.

It allows us to now level set how big should the government be.

But more importantly, the number of people

in the government are also the ones that then direct downstream spend, that make net new rules.

And if you can slow the growth of that down, you're actually doing a lot.

In many ways,

I wish Elon had come in and created Doge now.

Like, could you imagine if Doge was created the day after this Supreme Court ruling?

It would have been a totally different outcome, I think, because with that Supreme Court ruling in hand, these guys probably would have been like a hot knife through butter.

Travis.

So I think it's a big deal.

Except that ruling doesn't happen without Doge, that Doge caused that ruling to occur.

True.

Well, the EO did.

You could have passed.

That was all Doge style, though.

You know what I'm saying?

If they wasn't firing people, yeah, they probably wouldn't have felt the need, to your point, Travis, to actually file this.

But Travis, if you are living in the age of AI, efficiency right now, operations of companies is changing dramatically.

Can you imagine telling somebody you can be CEO, but you can't change personnel?

That's the job.

You get to be CEO, but you just can't change the players on the team.

You can buy the Knicks, but you can't change the coach.

You can grow it.

You just can't shrink it.

It's like running a unionized company, which actually does exist.

Our large unionized companies where you can't do any of these things.

Right.

Do they still exist or are they all gone?

I think they're going quickly.

Yeah, probably.

I think this just gets back to what is actually Congress authorizing when a bill occurs.

And

there's certain things that are specific and certain things that aren't.

And I don't, I'm not sure that in a lot of these bills, bills, it's not very specific about exactly how many people must be hired.

And so if it's,

I'm just doing the common man's sort of approach to this, which is like, if, if the law says you have to hire X number of people, then that is what it is.

If the law says you, here's some money to spend, here are the ways in which to spend it, but it's not specific about how many people you hire, then that is different.

Yeah, it should be outcome-based.

Hey, here's the goal.

Here's the key objectives, right?

Travis is totally right.

There's a variety of different laws, some with incredible specificities, some with very broad mandates.

The Constitution clearly says that all executive power resides in the President of the United States, period.

There's no exceptions there.

However, Congress does appropriate money.

And post-Watergate,

many people think Congress has the power to force the president to spend the money.

And you can debate that.

You can debate it on a per statute basis.

And that will be more nuanced and that's going going to get litigated whether the president can refuse to spend money that Congress explicitly instructed him to spend, sometimes called empowerment.

That's a very interesting intellectual debate.

This one's a little bit easier.

It'll get more complicated again.

Like this EO is only approved to allow for the planning.

I think the vote might be closer.

I think there's still a majority on the Supreme Court for the actual implementation, but it may not be 8-1 when there's a specific plan that has to navigate its way through the courts again.

Yeah, it's super fascinating.

Yeah, I wonder if they're going to get to the point where they're going to say in every bill, you need to hire this number of people to hit the public.

Well, I don't know if they can.

Like, that's where it gets borderline unconstitutional, like, where you actually prescribe that the president, in the exercise of his constitutional duties, has to hire a certain number of people.

That feels pretty precarious.

Well,

I'm not sure, Keith.

That's just like they prescribe a whole bunch of other things.

I know, but

you must appropriate money to this specific institution to do this specific work.

But that's not an executive function.

Like, if you said, like, the Secretary of State has to have X number of employees doing something, the Secretary of State is your personal representative to conduct foreign affairs on behalf of the President of the United States.

It gets a little bit more messy as you translate it to people that the President should.

I mean, yes, Congress does set, you know, which people are subject to Senate confirmation and what their salaries and compensation bans are.

So it's never going to be fully binary where the president can do whatever he wants.

And it's never going to, I don't think it'll be constitutional for Congress to mandate and put all kinds of hand costs on the president.

Well, then you also have performance that comes in here.

What if you look at the Department of Education and say, scores have gone down?

We've spent this money.

We're not getting the results.

Therefore, these people are incompetent.

Therefore, I'm firing firing them for cause, and I'm going to hire new people.

How are you going to stop the executive from doing that?

There's been a bunch of litigation, you know, in parallel to this litigation about the president's ability to fire people.

And for the most part, the Supreme Court's basically, with maybe the exception of the Federal Reserve Chair, said that the president can fire pretty much anybody he wants.

I mean, that's the way to go.

It's like, I mean, I hate to be caught.

But if the results aren't there, I think if they're a presidential, yeah, if they're a presidential appointee, the president should be able to fire you at will.

Just like if you were a VP at one of our companies, the CEO should be able to fire you at will.

But what about, Keith, if the whole department sucks?

Hey, you guys were responsible for early education.

You had to put together a plan.

The plan failed.

Everybody's fired.

We're starting over.

Like, you should be allowed to do that.

How are we going to have an efficient government?

Some of these departments were created by congressional statute.

like the Department of Education in 1979.

And you're right.

Every single educational stat has got worse in the United United States since the department was created.

But there is a law on the books that says there shall be a department of education.

So you may have to repeal that.

All right.

Listen, we're at an hour and a half, gentlemen.

Do you want to do the FICO story or should we just wrap, Jama?

And we got plenty of show here.

It's a great episode.

Anything else you want to hear?

I didn't have much to say on the FICO story.

I thought these other topics were really good, though.

We did great today.

This is a great panel.

I'm so excited you guys are here.

Let me just ask you guys,

off-duty stuff that you can share with us, with the audience?

Any recommendations?

Restaurants, hotels, trips,

movies you watch, books you read.

Keith, I know that you are an active guy.

What's on your agenda this summer?

Anything interesting you can share with the audience that you're consuming, conspicuous or otherwise?

Well, I don't want to share any good restaurants or hotels because

you're on had a babysitter, you're not going to select your babysitter.

Yes.

Can I get your nanny now?

But there are, there are things that are, what do you call it, no marginal cost consumption, like Netflix.

So, for example,

you know, this documentary on Osama bin Laden is phenomenal.

Like, I don't know if any of you have seen it.

It's brand new.

And, you know,

I'm a student of this stuff and I thought, you know, I knew the whole story and et cetera.

Watch episode one.

Just start with episode one.

And it just blew me away with new information, new new footage, just absolutely incredible stuff.

So highly, highly recommend it.

What was the big takeaway for you so far?

I don't know if there's any like specific takeaway, but just like so many parts of the story are misunderstood and not really understood.

And how the various confluences of somewhat random things lead to a very catastrophic result.

But

it's like as dramatic as the best movie, but it's a full documentary and you will learn things and absorb things.

I just, I've had friends while I've been recommending it to friends and for a story you think you know, it's incredibly incredibly revealing.

Okay.

Travis, anything you got on your plate there that you're enjoying?

A restaurant, a dish?

I mean, look, you know, I mean, Jason, you know, I go to Austin a lot.

Yes.

Like basically from March till October, I do about 15 weekends.

In Austin, I have a lake house.

Jason's hung out a couple times.

So I love water skiing.

That's my whole thing.

That's my like,

I just love it.

It's just my thing since I'm very sad.

Very sad.

Yeah, and it's lake.

It's, I call it lake life.

So

that's a thing.

And then I recently, this little bit of like a side quest,

I recently purchased the preeminent Batgammon engine.

XG.

XG.

That's right.

It's acronym is it's Extreme Gammon.

And so so the preeminent engine, so all the pros rate themselves based on this.

It was done.

It was built by this amazing entrepreneur, this guy Xavier,

who is just a full-on sort of

ultra, ultra.

I mean, just, what's the word I'm looking for?

It's not

like a savant, essentially.

but hasn't worked on it for many years.

So I'm getting back into it and

making it like taking modern machine learning sort of deep learning techniques and like big compute and saying can we push the game of backgammon forward so super exciting and ultra training apps to get people up to speed quickly i played in my first backgammon tournament in cached

so that was pretty cool no wait yeah okay yeah all due respect you're the founder of uber you're very high profile you go to this backgammon is this like held at the motel eight in like a conference room in the back i mean take a it was amazing

it was at the it was like a month ago or so

there's like a big tournament and it was uh so the the united states back end federation had this big tournament it was i guess it was um

at the los angeles lax at the l ax hilton and it was in the yes it was in the basement of the hilton great

and it was like

next to the dungeons and dragons convention it would it had those kinds of legit vibes.

I love it.

And like, people would, so, so, I went in super low pro, just did my thing, but eventually was recognized.

But I was not recognized as the founder of Uber.

I was recognized as the owner of XG.

Ooh, the owner of XG.

And then there was like a full-on

melee that basically occurred.

They're like, oh, the owner of XG, Travis is here.

Tramath, I feel like we've got a window here to do the all-in backgammon high-end tournament.

We got to lock this down now.

We got to lock down the all-in backgammon set.

I get the co-branding rights on this, okay?

Absolutely.

XG, XG.

Well, no, the all-in XG.

You know, like, because I love a great backgammon set.

If we could make like a $10,000 one, Shamath, we could kill turtles or white rhinos, all the animals that, you know, Freeberg's trying to protect.

We could murder them and then make.

That would be so great.

Yes.

Like maybe the white could be, you know, rhinos and then they could take something else, elephant skin, something, you know, just really tragic, and then eat the meat and make

the backhand set for you.

I love backhand.

And honestly, like, if I wasn't attempting to be like an expert poker player,

that is the game.

I mean, if you're talking about a Pandora's box where once you open it, oh my God, you can go down the rabbit.

Let's do that.

Backhand is a beautiful, beautiful, beautiful game.

I love the vibes of sitting.

Travis and I sat.

I got some cigars out.

You know, we pour a little of the all-in tequila, tequila.allin.com.

We get that going.

A couple of the all-in cigars, and then we have the all-in back.

It's a wonderful hang.

Yeah.

Keith, would you consider giving us some of your money playing back?

Absolutely.

We got to get some of that

money on the table because you don't play poker with us.

I don't play poker, but backyamon, yeah, that sounds great.

And I'll bring better tequila.

I'll have better tequila.

Well, like we're going to upgrade.

We'll do a little taste off.

Yeah.

So you've insulted now Elon with the Senate seats and Sachs with his uh

my Tiquilo is much better, trust me.

Wait, why don't you?

Who is left in the PayPal Bafia?

You'd like to insult before the Senate.

Reed Hoffman

or Peter.

Anything about Peter?

Reed can join Elon's party.

He's collecting a bunch of misfits, so we might as well take Reed too.

All right, listen, this has been another amazing episode of the number one podcast in the world, the only podcast for your Sultan of Science who couldn't make it today.

He's at the

conference, so we don't mention.

And David Sachs, who is out making America safe in AI and crypto.

Shamath Polyhapatiya, world's greatest,

Travis.

Keith, thanks to the campaign.

Thank you for coming.

Thanks for appreciating.

You guys were great today.

What a panel.

See you all next time.

Bye-bye.

We'll let your winners ride.

Brain Man David Sachs.

And it said, we open source it to the fans and they've just gone crazy with it.

Love you bestie.

I'm the queen of Kino.

Besties are gone.

Oh, man.

My avatasher will meet me at the end.

We should all just get a room and just have one big huge orchief because they're all just useless.

It's like this like sexual tension that they just need to release somehow.

We need to get mercy.

I'm going all in.