All-In Podcast: Silicon Valley's New Oligarchs and the Price of AI Doom-Mongering
June 19, 2026 - Weekly Episode
The All-In podcast hosts delivered perhaps their most pointed critique yet of Silicon Valley's emerging power dynamics, arguing that a new American politburo is forming while frontier AI labs fumble their way into regulatory capture through relentless fear-mongering. The most striking new insight: the chaos around Anthropic's Fable model wasn't primarily about national security but revealed deeper structural problems in how AI companies engage with government oversight.
The Formation of America's New Oligarchs
Chamath Palihapitiya opened with an aggressive framing that set the tone for the entire discussion. He described what he sees as the formation of a "great American politburo" where politicians like Elizabeth Warren, Bernie Sanders, and Ro Khanna are consolidating power to control the economy, education, and media. According to Palihapitiya, these figures present themselves as virtuous defenders of equity and justice while actually seeking to "seize the means of production" and eliminate individual economic mobility.
The critique grew sharper as the conversation turned to learned helplessness. Palihapitiya shared personal history growing up on welfare in Canada, noting his father received just enough government support to create a cycle of dependency. "The threshold for learned helplessness is far far lower than one may think," he explained. At roughly seventeen thousand dollars annually for a family of five, his father oscillated between drinking and unemployment rather than taking available jobs. This wasn't theoretical policy discussion but lived experience informing his views on government dependency.
David Friedberg connected this to fundamental questions about wealth creation and private property. He explained that wealth doesn't come from accumulating stuff but from creating "machines that make stuff," meaning the productive enterprises that generate value for society. When the market values those machines highly, that creates the paper wealth that critics attack. "Elon doesn't have one more dollar in the bank than he did the day before the IPO," Friedberg emphasized, noting the distinction between net worth and actual liquid assets.
SpaceX's Record IPO and What It Reveals About American Capitalism
The SpaceX IPO provided the perfect case study for the hosts' broader arguments about wealth creation and economic mobility. The company went public at one hundred thirty-five dollars per share, raising eighty-five billion dollars in what became three times larger than Saudi Aramco's previous record. The stock closed up nineteen percent on the first day and continued climbing to one hundred seventy-seven dollars by the podcast taping.
What made this particularly instructive was Elon Musk's path to becoming the world's first trillionaire. David Sacks emphasized that Musk arrived in the United States as an immigrant and literally slept on floors. He created companies from nothing through vision and hard work, including thousands of other people who became wealthy alongside him. Sacks pointed to a welder at SpaceX who reportedly made a million dollars in stock, demonstrating that the distinction between labor and capital isn't as rigid as Marxist theory suggests.
The broader point about economic mobility became clear when Friedberg noted that human agency is limitless, but government systems that provide everything actually "take away people's agency." The result is indentured servitude to the state rather than the freedom to progress economically. Palihapitiya argued this creates a shell of what people could become, referencing his own father as someone with unfulfilled potential trapped by dependency.
Jason Calacanis raised an important democratization point about the IPO itself. SpaceX allocated twenty to thirty percent of shares to retail investors through platforms like Robinhood and Charles Schwab. According to Vlad Tenev, roughly six to seven hundred thousand Robinhood users received allocations, even if just one or two shares. This contrasted sharply with traditional IPO processes that exclude ordinary Americans from early-stage value creation. Calacanis argued passionately that the government's accredited investor rules keep "poor people poor" by restricting private company stock purchases to the wealthiest four or five percent of Americans.
The Anthropic Fable Fiasco and What Really Happened
The podcast provided the most detailed insider account yet of why the government forced Anthropic to pull its Fable 5 model just days after launch. According to Sacks, who received direct readouts from White House officials, the situation stemmed from a pattern of concerning behavior rather than a single incident.
The timeline began in April when Dario Amodei came to Washington and "spiked the cortisol level" by describing Mythos as a cyber weapon with advanced capabilities. This got officials extremely focused on the security implications. Anthropic then created a trusted partners program for Mythos preview, but according to the Washington Post, they expanded this to approximately fifty companies without consulting the White House. Most problematically, they allegedly shared access with parties the White House believed had connections to China, specifically SK Telecom in South Korea.
When Anthropic launched Fable as Mythos with guardrails, private companies testing the model discovered jailbreaks that bypassed the safety features. One of these companies was Amazon, Anthropic's largest shareholder and cloud partner. Amazon's security team escalated concerns all the way to CEO Andy Jassy, who then contacted the White House. This wasn't a casual complaint but a formal warning from a major cloud provider worried about liability exposure across AWS.
The critical breakdown occurred in how Anthropic responded. According to Sacks, White House officials including the Treasury Secretary called Dario personally expecting a quick resolution. Instead of immediately agreeing to fix the problem, Dario "pedantically argued that this jailbreak wasn't really a serious issue." He later published a blog post differentiating between minor and major jailbreaks, which struck officials as completely off-brand for a company positioning itself as the AI safety leader.
Sacks emphasized this came after Anthropic had already created problems by extending Mythos access without approval. The administration felt they had credible information from Anthropic's own trusted partner about a national security threat, yet received what they perceived as resistance rather than cooperation. The export control letter restricting Fable to US citizens was the government's reaction when Anthropic refused to voluntarily take down the model until the jailbreak was fixed.
Political Tribalism or Legitimate Concerns
Calacanis raised the question of whether political tribalism influenced the government's response. Anthropic notably didn't attend White House CEO meetings when President Trump hosted tech leaders. The company is backed by Reid Hoffman, who funded lawfare against the Trump administration. There's undeniable tension between the political left and right, with Anthropic positioned firmly in one camp.
Sacks pushed back hard on this interpretation. He insisted the decision didn't come from the Department of War or from him personally, despite his role in the administration. Secretary Pete Hegseth's tweet saying "every passing day proves why this was the right move" to kick Anthropic out of the building came after the export control letter was sent, not before. Sacks characterized it as vindication rather than motivation, a sense of "we told you so" after having difficult contract negotiations with Anthropic earlier.
Friedberg took a different angle, arguing the real problem is lack of regulatory framework. Anthropic antagonized the government to create regulatory oversight, then felt antagonized when that oversight actually materialized. He predicted inevitable disaggregation of the AI stack similar to what happened with IBM mainframes in the sixties and desktop PCs in the eighties. The current oligopoly of three or four companies controlling everything isn't sustainable because market forces favor fragmentation across chips, clouds, models, and applications.
The Leaders of Frontier Labs Leave Much to Desire
Palihapitiya delivered perhaps the harshest assessment, arguing that frontier AI lab leaders demonstrate "a consistent pattern of evasiveness and immaturity" that disserves the entire AI movement. He believes AI represents the ultimate economic mobility enabler, giving everyone unlimited capacity to discover their upper bound as humans. Against that incredibly positive-sum backdrop, constant "doomerism, hype cycle, naivety" from lab leaders creates mistrust and holds progress back.
The consequences are concrete. First, Silicon Valley's prestige in American society continues decaying. When Palihapitiya arrived in two thousand one, it was a place of well-intentioned misfits building important things without taking themselves too seriously. Now there's a "veneer of negativity and mistrust" created largely because "we just cannot get our shit together. And the leaders of the Frontier Labs are public enemy number one."
Second, this creates massive opportunity for hyperscalers to position themselves as responsible gatekeepers. Amazon, Microsoft, and Google can argue to governments worldwide that they should provision AI models with robust KYC, customer identification, prompt auditing, and compliance infrastructure. The frontier labs made this inevitable through poor handling of safety concerns. Palihapitiya pointed out the impossibility of neocaler data centers replicating the trillions of dollars in infrastructure the hyperscalers spent decades building.
The result shifts from a diverse, robust, open ecosystem to debates about "gatekeeping and duopoly versus oligopoly." Instead of distributing the fundamental unlock for human potential widely, a handful of companies will control access. "They have created a total mess and it's a shame," Palihapitiya concluded. Silicon Valley once again looks like people optimizing for themselves, taking nobody else seriously in an unserious and immature way.
Claude Psychoanalyzes Its Creator
In a remarkable segment, Palihapitiya had Claude read Dario's essays "Machines of Loving Grace" and "AI Policy on the Exponential," then asked it to perform psychological analysis of its creator. He specifically requested truthfulness despite Claude being an Anthropic model, and asked it to view everything through the lens of the effective accelerationism movement.
Claude's analysis identified what it called "epistemic exceptionalism," not a god complex but something more subtle. The AI described Dario as someone who distrusts other labs as recklessly racing, distrusts authoritarian states, distrusts markets to distribute gains fairly, distrusts institutions to move fast enough, and now distrusts government to wield power transparently. Claude observed: "That's a very long list of untrustworthy actors. The list of trustworthy ones is conspicuously short and it has a suspicious tendency to resolve towards people who reason the way I do operating under the rules I help design."
The analysis continued: "When your safety framework requires that someone hold the keys and your analysis keeps concluding the other key holders can't be trusted, you've built a machine that outputs me no matter what you feed it." Claude identified the pattern as believing "my reasoning is the loadbearing one. And the failure of others to reach my conclusions is evidence of their corruption or slowness, not of my error."
Most tellingly, Claude pointed to Anthropic's use of the word "misunderstanding" to describe the Mythos situation. That word choice "presumes that if everyone simply understood correctly, they'd agree with him. That word choice under stress is the closest thing in the public record to the tell. It's not megalomania. It's the quieter, more defensible feeling conviction that disagreement is downstream of error."
Sacks agreed the analysis was accurate, noting Ben Thompson made similar observations. Anthropic's entire mantra is that AI is super dangerous and only they are virtuous enough to control the negatives. They position themselves as Jedi who will bring balance to the universe. This makes them a magnet for talent but also explains their fraught relationship with oversight.
Regulatory Capture Through the Back Door
The conversation revealed that Anthropic didn't have particular problems with the Trump administration initially. Their issue was getting spoiled by the Biden administration, which they had completely captured. Sacks noted that leaders of Biden's AI policy, including officials in charge of the NSC's AI efforts, the new AI Safety Institute, and his predecessor as AI czar, all went to work at Anthropic immediately after the administration ended.
Anthropic prioritized government affairs from an early stage, perhaps before any other AI company. Dario went to Washington and convinced key Biden officials that AI was extremely dangerous and needed control by government merged with a very small number of handpicked companies. This was the message Marc Andreessen heard when told not to waste time investing in different AI companies because the government would anoint winners and create a cartel.
Anthropic describes competition between AI companies as a "dangerous race condition" where safety falls by the wayside in the rush toward AGI. They want centralization and control, viewing competition as nefarious rather than beneficial. Sacks argued this is incredibly self-serving despite being framed in safety language. The fundamental philosophical question is whether competition is good or bad. Competition protects consumers, provides choice, brings out the best in competitors, prevents regulatory capture, and leads to decentralization rather than the totalitarian threat of centralized AI.
Dario was remarkably successful at resetting the conversation with Mythos in April. Just after getting kicked out of the Department of War due to unsuccessful contract negotiations, Anthropic came to Washington declaring they'd invented a cyber weapon. They freaked everyone out and got officials to go along with their idea of a government approval regime. Only through what Sacks characterized as hubris did they end up getting kicked out under their own nascent pre-approval regime.
The Broader Technology Cycle and Historical Perspective
Friedberg provided crucial historical context on technology-driven workforce anxiety. In nineteen sixty-one, Newsweek and the New York Times published articles about the demise of the workforce due to mainframe computers. The narrative was that automation made humans irrelevant and nobody would have jobs anymore. Instead, mainframes drove productivity gains that enabled individuals to do a hundred times more work, creating new industries and economic opportunity.
The same pattern repeated with desktop computers. The federal government spent hundreds of millions on workforce training programs in the early eighties that "literally did nothing" because the transition wasn't as difficult as feared. People figured out how to turn on Windows PCs, and companies helped employees navigate to higher productivity levels using new technology and tools.
Friedberg argued there's deep arrogance among technologists, himself included, that drives existential conclusions. "This will cure all cancer. This will solve all disease. This will eradicate all jobs. None of us. And by the way, no one's immune to this arrogance. From Elon to Sam to Dario to all of us. We all assume that this time is different and everything's about to change completely. And you know what? It doesn't."
He described technology as part of a long-term continuum of improvements driven by human ingenuity, unlocking human potential and giving people more capacity to do more. The job loss narrative and end-of-world framing drives everyone into insane frenzy when the reality is that technology does what it's always done, enhancing rather than replacing human capability. His prescription was simple: "Everyone needs to chill the fuck out and just let technology do what it does."
Sacks added that AI CEOs have been unnecessarily scaring the public about AI. If Dario had to release Mythos with no guardrails in April, he would have exposed the company to massive legal liability when thousands of companies got hacked. Meta and Google face product liability lawsuits constantly, including a recent multi-million dollar verdict over body shaming accusations. Of course Anthropic needed safety features and corporate responsibility to avoid business-ending liability. That's not a sacrifice for humanity deserving giant pats on the back. They shouldn't have gone to Washington to freak everybody out when they could have quietly worked with the NSA.
Iran Deal Framework and Market Implications
President Trump announced for the thirty-seventh time that the Iran war has ended, with an initial agreement signed June fifteenth to be formally codified June nineteenth in Geneva. The conflict that began February twenty-eighth appears headed toward resolution mediated by Pakistan, extending the ceasefire for sixty days with significant commitments from both sides.
The United States gets the Strait of Hormuz reopening, though it was open before the war. More significantly, Iran commits to not developing nuclear weapons, agrees to destroy its stockpile of enriched uranium under IAEA supervision, and will freeze its nuclear program at current levels for sixty days. Iran gets all sanctions lifted and three hundred billion in reconstruction funds, though the US isn't paying directly. Iran gets access to frozen assets and US forces will remove from the region after the final deal.
Major issues remain unresolved and deferred, including Israel's sign-on to the deal, Iran's nuclear enrichment going forward, and Iran's ballistic missile program. Sacks called the deal a tremendous achievement for the president despite difficult negotiations with Iranians. He highlighted reopening the Strait of Hormuz for vital materials, commitment against nuclear programs including collecting nuclear materials, ceasefire on all fronts, zero cost to the US, and potential rapprochement with Iran.
He challenged critics to articulate alternatives. Neocons apparently want ground troops to effectuate regime change, which Sacks called insane. Estimates suggest over a million troops would be needed for Iran, which is three times bigger than Iraq and constitutes a mountain fortress. Half a million went to Iraq, and even that many might not succeed against Iranian geography. "I'm not sending my kids to fight that war. You don't want to send your kids to fight that war," Sacks stated flatly.
Friedberg focused on the enriched uranium removal as the critical element. If all enriched uranium leaves the country, even if Iran restarted the program and rebuilt equipment and expertise, it would take years to a decade or fifteen years plus to build enough for nuclear weapons. That natural setback provides substantial defense for conditions that matter most. He hadn't realized that was in the deal and called it a really big deal if the enriched uranium actually exits Iran.
Calacanis expressed relief the war might end but called it a huge blunder that never should have happened. He argued the "trimming of the grass" approach Israel had taken was correct, isolating dictators and waiting them out rather than starting full-scale wars with unpredictable escalation risks. Bombing facilities every eighteen months was working fine without needing full-scale conflict. He acknowledged not having complete information available to Trump but emphasized that people must lead their own revolutions, and the US cannot enforce democracy in the Middle East after learning that lesson repeatedly. Palihapitiya simply noted the market will go to the moon on the news.