This episode of the All-In Podcast features the original quartet: host Jason Calacanis, David Sachs (venture capitalist and former White House AI & Crypto Czar), Chamath Palihapitiya (Social Capital CEO), and David Friedberg (CEO of Ohalo and The Production Board). The conversation spans capitalism, AI regulation, and geopolitics across a packed week of news.

The hosts open with a wide-ranging debate on economic mobility, government dependency, and the nature of wealth creation — sparked by SpaceX's record-breaking IPO. They then pivot to the Anthropic-White House crisis over the Mythos and Fable AI models, with Sachs providing an insider readout. Chamath and Friedberg offer sharp critiques of Frontier Lab leadership, and Friedberg references Dario Amodei's essays Machines of Loving Grace and AI Policy on the Exponential as the basis for a Claude-generated psychological profile of Amodei. The episode closes with analysis of the US-Iran ceasefire MOU, debating whether the deal's terms — including uranium destruction and sanctions relief — justify the conflict.

SpaceX IPO: The Biggest Raise in History

SpaceX IPO priced at $135/share, raised $85 billion — 3x Saudi Aramco's 2019 record — closed up 19% at $161, briefly surpassing Amazon and Microsoft to become the 4th most valuable company globally.

At time of taping, shares had climbed further to $177.

SpaceX's 2025 revenue is ~$19 billion vs. Amazon's $717 billion and Microsoft's $282 billion, meaning the valuation reflects massive future expectations.

Sachs emphasized that Elon's trillionaire status is entirely paper wealth: 'He doesn't have one more dollar in the bank than he did the day before the IPO. He doesn't have more stuff, doesn't have more houses.'

Elon is under a one-year lockup; Sachs predicts he will hold far longer as SpaceX is his life's work.

The company has been 25 years in the making with multiple near-bankruptcy moments.

Friedberg framed wealth creation as owning the machines that make things, not the things themselves: 'The reason why humans are more prosperous is because of all the machinery we've created to make stuff well into the future.'

Calacanis highlighted the retail democratization angle: roughly 600,000–700,000 Robinhood users received IPO allocations, with 20–30% of the offering going to retail — a model he argues should extend to private companies.

'It's fucking crazy that we still have this rule keeping poor people poor' — referring to accredited investor laws that restrict private company investment to the top ~5% of earners.

Sachs invoked Joseph Schumpeter's theory of the intelligentsia to explain resentment toward Elon: an intellectual class that produces words rather than things becomes hostile toward those who create machines and amass wealth.

Friedberg reframed the societal divide as makers vs. takers — not rich vs. poor: 'Artists, plumbers, electricians, woodworkers, computer scientists... are the makers. The takers are the analysts, the espousers, the armchair mechanics, the critics, the politicians.'

Cursor Acquisition: Elon's $15 Billion Masterstroke

SpaceX exercised its option to acquire Cursor at ~$60 billion on a stock-for-stock basis. Cursor is doing $4 billion in revenue, implying a 15x revenue multiple vs. SpaceX's 60–70x.

Cursor originally used Anthropic's Claude as its backend; Anthropic secretly built a competing coding agent (Claude Code) while telling Cursor it would never release one publicly.

Chamath had predicted the deal structure on a prior episode: 'If it's $60 billion in tomorrow dollars, effectively Elon's gotten a 50% discount' — and with SpaceX's valuation doubling, the effective cost was closer to $15 billion.

'His business intellect is off the charts. That is an incredible deal.' — Chamath, on Elon's Cursor acquisition timing and structure.

Learned Helplessness, Welfare, and the Limits of Government Support

Chamath shared a personal account of growing up on welfare in Canada: his family received ~$17,000–$19,000/year, just enough that his father cycled between drinking and not working rather than seeking employment.

'The threshold for that helplessness is far lower than one may think' — Chamath.

His father eventually got a low-level government clerical job near the end of Chamath's high school years, which transformed him: he drank less, was more regulated, and maintained the job until he passed away.

Calacanis explained the psychology of learned helplessness: giving people unsolvable problems causes them to stop trying even on easy ones, leading to depression they internalize — making agency and graduated challenge essential to human flourishing.

Friedberg argued human agency is the one resource with unlimited upside: 'No human being that's ever been alive has ever lived up to their true capacity as a human' — and government dependency systematically destroys that capacity.

Chamath warned that stacking government benefits to $30,000–$80,000/year equivalent will 'reduce people to a shell of what their potential is' and is fiscally unsustainable — pointing to Illinois's new property tax legislation as a live example.

Anthropic's Mythos Crisis: Jailbreaks, Foreign Access, and Trust Collapse

Sachs provided an insider readout: Anthropic expanded the Mythos trusted-partners preview to 50+ companies without White House approval, including at least one company — reportedly SK Telecom — that national security officials flagged as having potential China-linked ties.

The White House had previously ordered Anthropic to revoke SK Telecom's access to Mythos before the broader incident.

According to The Washington Post, Anthropic did not initially disclose it had given Mythos to SK Telecom.

Amazon (Anthropic's largest shareholder and cloud partner) reportedly discovered a jailbreak in the public Fable 5 release and escalated directly to the White House after Anthropic did not act — functioning effectively as a whistleblower.

Dario publicly argued the jailbreak was 'narrow' and not serious, publishing a blog post distinguishing minor from major jailbreaks — which Sachs called 'very off-brand' for a self-described AI safety company.

'They tried to call Dario and there were a number of conversations and the TL;DR of it is Dario refused... the White House was really just bewildered and confused by that' — Sachs, on the breakdown that led to the export control letter.

The Treasury Secretary reportedly called Dario personally; Sachs said it should have been a 5-minute call ending in immediate compliance.

Chamath argued the episode is handing hyperscalers a monopoly argument: they can now credibly offer governments KYC-wrapped, auditable AI access that Frontier Labs cannot match — potentially locking out neoscalers entirely.

'Can you imagine the cost and the complexity if you asked a neoscaler to build the same robust KYC or the same VPC infrastructure that Amazon and Microsoft and Google have spent decades investing trillions of dollars in? It's an impossibility.' — Chamath.

Sachs noted Anthropic had effectively captured the Biden administration's AI policy apparatus, with multiple senior Biden AI officials — from the NSC, AI Safety Institute, and the AISR role — going to work at Anthropic after the administration ended.

Sachs called for industry self-certification modeled on the MPAA and video game rating systems: 'The industry should have an industry certification... We should just self-certify. It's the simplest thing in the world to do.'

Dario Amodei's Psychology: Claude Analyzes Its Own Creator

Chamath fed Dario Amodei's two long-form essays — Machines of Loving Grace and AI Policy on the Exponential — into Claude and asked for a psychological analysis through the lens of the EA/EAC movement and the Mythos controversy.

Claude's analysis identified a pattern it called 'epistemic exceptionalism': 'Not that I am superior, but more my reasoning is the load-bearing one, and the failure of others to reach my conclusions is evidence of their corruption or slowness, not of my error.'

The analysis noted Dario distrusts other labs, authoritarian states, markets, institutions, and — post-Mythos — the government itself: 'That's a very long list of untrustworthy actors. The list of trustworthy ones is conspicuously short.'

The tell identified: Anthropic's use of the word 'misunderstanding' to describe the White House conflict — 'that word choice under stress... presumes that if everyone simply understood correctly, they'd agree with him.'

Sachs connected the psychological profile to Anthropic's business model: 'They believe that AI is super dangerous and only they are virtuous enough to basically control the negatives of it... They believe they will bring balance to the universe.'

Sachs argued Anthropic's framing of AI competition as a 'dangerous race condition' is self-serving centralization dressed as safety: 'They effectively want to create a cartel... if you see it from the point of view of what's really happening, it's incredibly self-serving.'

AI Regulation: IBM Parallels and the Disaggregation Thesis

Friedberg drew a historical parallel to IBM's mainframe monopoly in the 1960s: IBM controlled chips, hardware, OS, and software until the government forced disaggregation of the software layer — which launched the independent software vendor era.

The same disaggregation then happened with desktop PCs in the '80s, enabling the productivity boom and proliferation of the software stack.

Friedberg predicts the AI stack will similarly fragment: multiple chip vendors, multiple clouds, multiple models, local inference, and purpose-built application layers.

Friedberg pushed back on AGI existential narratives by citing 1961 Newsweek and New York Times articles predicting the mainframe computer would eliminate the entire workforce — a fear that proved completely wrong as productivity multiplied.

The federal government spent hundreds of millions in early '80s workforce retraining programs for the desktop PC transition: 'You know what ended up happening? People figured shit out.'

'Everyone needs to chill the fuck out and just let technology do what it does.' — Friedberg.

Sachs stated he does not want the Anthropic export control letter to become precedential policy: 'I don't want policy being made out of these types of emergencies... I hope this does not become the basis for policy.'

US-Iran Ceasefire MOU: Deal Terms and Unanswered Questions

A ceasefire MOU was announced after 110 days of conflict, with formal signing set for June 19th in Geneva, mediated by Pakistan. The deal extends the ceasefire 60 days and includes a Lebanon ceasefire.

Conflict began February 28th; the US is removing forces from the region after the final deal is signed.

Key US wins per Sachs: Strait of Hormuz reopening, Iran commits to not developing nuclear weapons, Iran agrees to destroy its enriched uranium stockpile under IAEA supervision, and the US pays nothing toward the $300 billion reconstruction fund.

Chamath flagged the uranium destruction as the most significant element: 'If all that enriched uranium comes out of the country, I think it's a very big deal' — even a restarted program would take a decade or more to reconstitute weapons-grade material.

Major unresolved items: Israel's sign-on to the deal, Iran's nuclear enrichment going forward, and Iran's ballistic missile program — all deferred from the MOU.

Sachs rejected neocon calls for ground troops or continued bombing: 'Iran is basically a mountain fortress. I've heard estimates it would take over a million troops, and even then it might not be successful. We sent half a million to Iraq, and Iran is 3 times bigger.'

Calacanis argued the war was unnecessary — Israel's periodic strike strategy ('trimming the grass') had been working — but acknowledged Trump may have had classified information justifying the approach that the public cannot evaluate.