Jul 11, 2025 · 1h 30m · allin
Grok 4 Wows, The Bitter Lesson, Third Party, AI Browsers, SCOTUS backs POTUS on RIFs
Jason Calacanis · 23m spoken Travis Kalanick · 22m spoken Chamath Palihapitiya · 17m spoken Keith Rabois · 14m spoken
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the All-In Podcast, guest panelists Travis Kalanick and Keith Rabois join Chamath Palihapitiya and Jason Calacanis to discuss food automation robotics, artificial intelligence scaling, agentic web browsers, political initiatives like Elon Musk's America Party, and Supreme Court rulings on executive federal restructuring.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 51.4% of the talking time here. How this is scored →
The hosts as informed peer 4.5 Guest teaching 3.3 Guest disagreement 2.9 The hosts pushing back 2.7
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:00:36 Keith dismisses Elon's third-party proposal
Keith bluntly characterizes Elon's third-party initiative as completely absurd for any human, comparing him to Michael Jordan attempting to play baseball.
Hardest push from the hosts ▶ 52:15 Chamath rejects agentic browsersChamath directly refuses Jason's positive framing of Perplexity's browser, calling building a web browser in 2025 the dumbest capital allocation decision possible.
Biggest teaching moment ▶ 1:00:36 Keith educates on approval ratings and political scienceKeith corrects the host's polling charts by explaining the flaw of averages, pointing out that Trump holds a record 95% approval among Republicans and that political science shows smart parties absorb third-party ideas.
The host holds their own ▶ 20:39 Chamath articulates The Bitter LessonChamath demonstrates deep domain context by breaking down Rich Sutton's architectural thesis to explain why general computation scales faster than human-labeled data strategies.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Lake Como Guestbook Prank on David Friedberg | 1 | 0 | 1 | 0 | Chamath and Jason kick off the episode with playful banter about leaving a fake guestbook entry blaming David Friedberg for stolen hotel items in Lake Como. Keith Rabois joins in casually, making a mild joke about NYC safety. | |
| Travis Kalanick on Autonomous Mobility & Pony AI Rumors | 2 | 3 | 2 | 2 | Jason introduces Pony AI while Chamath asks Travis Kalanick directly about potential competitive moves against Waymo and Tesla. Travis clarifies that there is no formal deal in place, reframing the news as inbound interest from partners seeking operational optionality. | |
| Lab37 Robotic Food Assembly & Kitchen Economics | 2 | 5 | 1 | 1 | Travis details the mechanics and unit economics of Lab37's 60-square-foot robotic bowl assembly machine. He educates the hosts on how labor costs fall from 30% to under 10% of revenue while eliminating order prep errors. | |
| Full-Stack Food Automation & The Internet Food Court | 3 | 4 | 2 | 2 | Chamath asks if fully automated kitchens will create a wasteland of commercial real estate. Travis reframes the problem by pointing out that 85% of meals are prepared at home, meaning automated food courts augment home meal preparation rather than replacing physical dining out. | |
| All-In Summit Announcement & All-In Tequila Launch | 8 | 2 | 3 | 4 | Chamath delivers an extensive breakdown of Rich Sutton's 'The Bitter Lesson' to explain xAI's rapid leap with Grok 4 over rivals reliant on human data labeling. Travis pushes back gently, noting that Tesla's autonomous approach still relies heavily on approximating human driver behavior. | |
| Synthetic Data, Human Data Limits & AI Scientific Discoveries | 6 | 5 | 4 | 3 | Keith directly corrects Chamath's binary split between human data and compute, reminding him that LLMs fundamentally depend on human writing. Chamath challenges the panel on venture capital strategy, asking how non-Grok companies survive if human labeling creates lag. | |
| Judo Flipping OpenAI and Elon Musk's Manufacturing Scale | 6 | 3 | 2 | 3 | Chamath asks how Grok can judo-flip OpenAI's massive active user base. Jason highlights Elon Musk's unique advantage in physical infrastructure and manufacturing speed, citing the rapid deployment of the Colossus GPU cluster. | |
| Open Sourcing Self-Driving Data and AI Vertical Integration | 5 | 5 | 4 | 3 | Jason suggests open-sourcing self-driving data would be a killer disruptive move. Keith forcefully counters that vertical integration—combining hardware, custom design, and software like Apple—is the true moat that competitors cannot easily copy. | |
| Applying Rich Sutton's The Bitter Lesson to Physical AI | 8 | 3 | 3 | 6 | Jason demonstrates Perplexity's new Comet agentic browser. Chamath forcefully rejects the premise, declaring that building a web browser in 2025 is a wasteful capital allocation and arguing Perplexity should focus on replacing the Bloomberg terminal instead. | |
| Analyzing Elon Musk's America Party and Third-Party Viability | 7 | 6 | 7 | 5 | Discussion turns to Elon Musk's proposal for an 'American Party'. Keith dismisses the idea as absurd, comparing Elon to Michael Jordan trying to play baseball, and explains political science principles around party absorption and approval rating distribution. | |
| Supreme Court Backs Executive RIFs and DOGE Restructuring | 7 | 4 | 4 | 4 | The panel analyzes SCOTUS's 8-1 ruling allowing presidential planning for federal workforce reductions. Keith provides constitutional context regarding executive power versus congressional power of the purse and statutory mandates. | |
| Keith Rabois Recommends Netflix's Osama bin Laden Documentary | 2 | 2 | 2 | 1 | Keith recommends Netflix's Osama bin Laden documentary for its surprising depth. Travis recounts purchasing the preeminent backgammon engine XG and getting recognized in the basement of an LAX Hilton hotel tournament. | |
| Planning an All-In Backgammon Tournament and Paypal Mafia Banter | 2 | 1 | 3 | 1 | The group banters about organizing an All-In high-stakes backgammon tournament with custom boards. Keith playfully insults Jason's tequila brand and mocks Reid Hoffman joining Elon's party of political misfits. |
Statements from this episode (21)
Rabois: NYC recorded zero July 4th shootings for the first time ever
“You know, as you probably heard on July fourth was the first time in recorded history that there were no shootings or no murders. In New York on that day.”
Kalanick denies Pony AI deal but admits inbound autonomous vehicle interest
“There's no real deal right now, but there is definitely some inbound.”
Kalanick: Lab37 robotic assembly cuts kitchen labor costs to 7-10% of revenue
“When they're running our machine, it's between seven and 10% of revenue.”
Kalanick: Lab37's food assembly machine produces 300 bowls per hour
“That machine right there does 300 bowls an hour.”
Kalanick: Uber Eats and DoorDash represent only 2% of US meals
“Uber Eats and DoorDash, they represent like 1.8% or two percent of all meals right now.”
Chamath: Meta spent $15B to buy 49% of Scale AI
“They just spent fifteen billion to buy 49% of scale AI.”
Kalanick: Tesla Hardware 5 will probably launch next year
“Hardware five is coming out on Tesla probably next year, which is going to make a big difference in what FSD can do.”
Keith Rabois predicts human-labeled AI data will be obsolete within three years
“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 to get revenue traction really didn't understand that there may be a year, two years, Three years max when anybody uses human lab…”
Travis Kalanick: Autonomous vehicle companies no longer use human data labeling
“You're some of the players in the autonomous software space, autonomous vehicle software space are no longer doing any labeling because the machines are doing it all.”
Rabois: Science-focused AI models uncover biological connections humans never identified
“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 o…”
Calacanis: Sam Altman corrupted OpenAI's mission and betrayed Elon Musk
“Individuals, which by the way, was what he backed before Sam Altman corrupted the original missionary basis of opening. I made a closed AI in a, you know, this is nothing derogatory towards him, but he did hoodwink and stab Elon in the back.”
Rabois: Apple missed the AI wave but thrives through vertical integration
“The 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.”
Rabois: OpenAI must ship a hardware device to maintain product advantage
“The way to stay ahead of the product level is shipping a device. They've got to ship the device.”
Rabois: The Bitter Lesson requires data that may take decades to acquire
“Conceptual, you know, the blog post is right, but that's only true when you have enough data, and depending on 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.”
Rabois predicts Google Search will lose to AI competitors like ChatGPT
“I think Google's also, Google search, cross-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 Chad GPT. They're going to lose the searc…”
Chamath: Building a new web browser in 2025 is a stupid investment
“Look, I think building a browser is an absolutely stupid capital allocation decision. Just totally stupid and unjustifiable in 2025.”
Chamath: Perplexity's path to a legacy business is replacing Bloomberg
“Specifically for perplexity, I think their path to building a legacy business is to replace Bloomberg.”
Rabois: Capping US spending at 2019 levels would yield budget surplus
“Because if you just held federal spending to 2019 levels, 2019 is not like, you know, decades ago. Literally with our current tax revenues, we would be in a surplus.”
Rabois: Musk's America Party can win House seats but not Senate races
“I think you can win a few house races. I think that's doable. I don't think you can win a Senate race.”