Trump Takes On the Fed, US-Intel Deal, Why Bankruptcies Are Up, OpenAI's Longevity Breakthrough
Summary
Trump’s attempt to fire Fed Governor Lisa Cook turned a mortgage-fraud allegation into a test of central-bank independence and presidential control. Cook, who had not been charged, sued after Trump invoked “for cause” removal over allegedly declaring two homes as her primary residence; the allegation dated to before she became a Fed governor. Polymarket put her chance of being out by year-end at 25% during the recording. Chamath and Sacks called the Fed inherently political, while Friedberg defended its 14-year terms because forcing down short rates could push long-term borrowing costs higher through inflation and fiscal risk.
Chamath’s larger call was that markets should replace the Fed in setting the price of money. He would leave banking supervision and payment clearing with the institution, but shift lender-of-last-resort work toward Treasury and let SOFR, Treasury auctions, bank pricing oracles, and real-time economic data determine rates: “precise data in the free markets, imprecise data in a group of people.” Jason and Chamath pushed back that the FOMC has 11 current voters, recent genuine dissents, and that Powell has only one formal vote.
Sacks argued that Powell’s timing—not merely his forecasts—showed political behavior with lasting asset-allocation consequences. His evidence was Powell retaining the “transitory” inflation narrative through his November 22, 2021 renomination, abandoning it November 30, and continuing easy policy into an asset bubble before an unusually fast tightening cycle. Jason rejected the partisan inference, noting bipartisan votes, subsequent inflation and tariff uncertainty, while Sacks maintained that “too late Powell” helped create the 2021 bubble, the 2022-23 crash, today’s refinancing stress, and a 50-basis-point pre-election cut followed by a pause.
The panel broadly preferred Washington’s new 10% passive Intel stake to giving away roughly $8-9 billion of CHIPS Act support with no upside. Chamath framed equity as the missing taxpayer return when government acts as strategic backstop, while Sacks limited the model to national-security priorities that private markets have failed to deliver—especially a semiconductor supply chain “single-threaded on Taiwan.” Jason supported the economics but called Trump’s public attack on CEO Lip-Bu Tan before the deal “crony capitalism,” arguing that the style obscured sound substance.
The harder Intel question was where government-owned equity should live and who can keep Congress from spending against it. Friedberg proposed placing such assets inside Social Security’s OASI trust fund—which currently holds only special Treasuries and, he said, faces insolvency around 2030-33—while Chamath proposed a sovereign wealth fund seeded by claimed commitments of $600 billion from Japan, $300 billion from Korea, and several hundred billion from Europe, with the US receiving 90% of the upside. Jason objected that a country already $36-37 trillion in debt should prioritize repayment; all sides worried that unringfenced gains would simply become permission to spend more.
The 446 large bankruptcies through July looked to Chamath like delayed ZIRP cleanup, not a three-month tariff shock. Years of zero rates filled weak companies’ “reservoir of free money,” postponing failure at businesses such as JOANN Fabrics and Party City; retail leases and PE leverage then amplified the washout as capital costs normalized. The systemic pressure point is commercial real estate: $2.2 trillion of debt matures before 2028, and Sacks showed how a $100 million building with $66 million of debt could fall to $60 million in value, support only a $40 million refinancing, and demand a prohibitive $26 million equity check.
Friedberg presented GPT-4b micro’s protein-design results as evidence that narrow language models could accelerate cellular rejuvenation, but the medicine remains early and dangerous. Working with Retro Biosciences, the model proposed variants of the Yamanaka factors that were reported as 50 times more effective: over 30% of cells showed relevant markers within seven days and 85% expressed critical stem-cell markers by day 12, versus less than 0.1% of cells previously undergoing rejuvenation. Friedberg estimated seven to 12 years to an approved therapy and stressed that excessive reprogramming can return cells fully to stem-cell status, producing uncontrolled growth that “looks like and acts like cancer.”
Deep dive
1. Lisa Cook’s firing put Fed independence on trial
Jason framed the immediate dispute carefully: Trump fired Governor Lisa Cook “for cause” after the FHFA alleged she had treated two homes as her primary residence. The allegation dated to before she was a Fed governor. Cook had not been charged, a criminal referral had been sent to DOJ, and she sued on the ground that the White House lacked authority to remove her.
The stakes extended beyond Cook: it was described as the first presidential firing of a Fed governor in US history, with an emergency hearing scheduled for the next morning and an eventual Supreme Court fight considered likely. At recording, Polymarket priced a 25% probability that Cook would be gone by year-end.
Chamath rejected the premise that Fed officials are meaningfully independent: “These are partisan employees,” chosen by presidents because they align with a governing philosophy. Since governors do not hold lifetime appointments, he considered presidential removal reasonable when an official conflicts with the program voters elected.
2. Friedberg defended insulation because the yield curve can revolt
Friedberg’s pushback — worth keeping: 14-year gubernatorial terms were specifically designed to outlast political cycles, giving the institution resilience against “the electoral whims of politics.” If that duration is wrong, Congress should change it explicitly rather than using investigations to evade it.
Cutting overnight rates may stimulate borrowing, purchases, growth, inflation, and government spending simultaneously. Friedberg warned that this could push 30-year yields higher as markets reprice inflation and the government’s ability to service its debt, making the short-rate demand much less straightforward than “interest rates come down.”
His distinction on enforcement was narrow: officials who may have broken laws should be investigated, but investigations should not become a mechanism for defeating fixed terms. The underlying allocation of power between Congress, the executive, and an independent board should be “relitigated” openly.
3. Chamath would reduce the Fed to supervision and payments
Chamath’s first-principles question was what a monthly committee can contribute to a “$130 trillion global economy” moving at extraordinary speed when its inputs—from BLS releases to GDP estimates—are frequently incorrect or revised. “The two things that are the most dynamic, they are the worst at doing.”
Of four core functions, he would move lender-of-last-resort responsibility toward Treasury, which he believes can protect taxpayer upside better, and let capital markets handle monetary pricing and stability. He would preserve banking supervision and regulation, plus payment-system and clearing-house functions, which he called comparatively useful and uncontroversial.
The proposed replacement was not a single AI deciding the rate. Banks would build pricing oracles from real-time economic information, submit bids into Treasury auctions, and converge on a market-clearing risk-free rate; SOFR already provides a more precise market mechanism than a Fed funds target expressed as a range.
Chamath pointed to Commerce beginning to publish GDP data on a blockchain and proposed adding anonymized payroll and employment feeds. Markets had repriced tariff risk within weeks, he argued, demonstrating why “precise data in the free markets” can respond faster than humans reviewing older aggregates.
4. Powell’s 2021 timing became Sacks’s evidence of politics
Sacks traced his case to the summer 2021 shock inflation print of 5%. Powell supported Biden and Yellen’s “transitory” interpretation until Biden renominated him on November 22, then said on November 30 that the word should be retired—a sequence Sacks considered too exact to dismiss as coincidence.
Policy did not merely stay at low rates: Sacks cited continued quantitative easing, including roughly $180 billion of government-bond purchases, even as inflation changed regime. Chamath added that the eventual tightening was historically fast, asking whether earlier candor would have reduced the later economic shock.
Sacks’s causal chain was categorical: delayed tightening inflated startups and real estate in late 2021, contributed to 9% inflation the next year, caused the 2022-23 crash, and left low-rate property deals facing today’s refinancing wall. “It caused a horrible misallocation of capital.”
Jason resisted diffusing every institutional decision into Powell but also resisted concentrating every vote in him. The FOMC normally contains seven governors and five bank presidents; with one vacancy it had 11 voters, and recent decisions had included dissents. Sacks countered that any chairman still sets the agenda and assembles majorities.
5. The 50-basis-point cut produced a second political indictment
Sacks’s second exhibit was the September 2024 half-point cut, larger than the expected 25 basis points and shortly after Elizabeth Warren urged Powell to front-load easing. Her September 16 letter cited inflation at 2.5%, down from a 7% mid-2022 peak, and warned that delayed cuts risked a crisis.
The alleged hypocrisy was that Warren later wanted Powell to resist Trump, while Powell paused easing after Trump defeated Kamala Harris. Sacks concluded: “He’ll cut for Biden. He’ll cut for Yellen. He’ll cut for Kamala. He will not cut for Trump,” even with PCE, by his account, at 2.0%.
Jason’s rebuttal was that many observers—including the podcast—thought cuts were overdue, so Warren did not explain the vote. Inflation had ticked upward, Trump’s tariff policy was unprecedented, and the Fed waited for May and June evidence before moving toward September; Bowman and Waller’s July preference for a 25-basis-point cut also showed internal dissent.
6. Intel equity converted a subsidy into taxpayer participation
The government’s Intel transaction swapped roughly $8-9 billion of allocated CHIPS support for a 10% passive stake: non-voting shares, no board seat, no golden share, and no governance rights. The panel’s baseline consensus was simple—if the money will be deployed anyway, taxpayers should receive upside.
Chamath situated the deal against a strategy he attributed to Hu Jintao’s 2006 speech: China identified semiconductors, rare earths, and pharmaceutical APIs as strategic sectors, then used government balance sheets, price shaping, and spot-market pressure to make competing projects difficult to finance.
America historically served as backstop without retaining comparable gains. Chamath contrasted TARP and the Goldman Sachs rescue—where he said Warren Buffett and Berkshire shareholders captured upside—with the Intel and MP structures: “We have given money away in times of duress with absolutely no upside, and I think it has to change.”
7. National security supplied the limiting principle
Sacks agreed on equity but rejected a generalized government portfolio. Intervention should require a national-security priority or similar necessity plus a demonstrated failure of private markets; semiconductor manufacturing qualified because the US had allowed a critical supply chain to become “single-threaded on Taiwan.”
Requiring equity or warrants also imposes a cost on recipients, reducing the incentive to seek federal rescues. That made Intel a significant improvement on the original CHIPS structure without turning every troubled company into a candidate for public ownership.
Friedberg identified three signals of free-market failure: a unique regulatory unlock, government-supplied capital, or government becoming the primary buyer. Their growing prevalence suggested to him that “there is some degree of socialism underway,” regardless of which political explanation one prefers.
Jason supported the substance but attacked the method. Trump publicly called CEO Lip-Bu Tan conflicted and demanded his resignation; Tan later explained that his China investments dated to a period when they were broadly accepted, and the meeting produced a deal. Jason called the sequence “crony capitalism,” while Sacks answered that “you’ve got to break some eggs to make an omelet.”
8. The equity’s destination mattered more than its mark-up
Friedberg saw three possible homes: the federal balance sheet without a coherent strategy, a new sovereign wealth fund with a new bureaucracy, or the existing OASI trust fund behind Social Security. He favored OASI because its assets currently consist only of special US Treasuries—effectively IOUs from the government to future retirees.
Congress would need to revise statutes dating to the 1930s, but Friedberg argued that equity could compound for beneficiaries and help fill a program he said would go bankrupt around 2030-33. “Rather than create new holes in the government,” the assets should “fill holes” already visible.
Chamath proposed immediate seed capital from tariff-related investment commitments: $600 billion from Japan, $300 billion from Korea, and several hundred billion from Europe, collectively exceeding $1 trillion. He said the US receives 90% of the upside under those arrangements and should combine it with stakes such as Intel and MP.
Jason objected to having the government “act as a venture capitalist” with retirees’ Social Security savings. Sacks rejected taking those funds, though he accepted placing selectively acquired bailout equity into a ringfenced fund that could support Social Security.
9. Any new asset pool risks becoming a spending license
Friedberg’s core governance worry was that Congress would spend ahead of future gains. He cited California expanding its budget as revenue rose, then confronting a deficit, and argued that Social Security contributions had likewise been spent while the trust fund received government obligations in their place.
The group therefore converged more on separation than portfolio strategy: equity cannot be booked as ordinary income that makes a single-year deficit look smaller. Without a legally protected “box,” Friedberg warned, it becomes “another spending mechanism” vulnerable to everyone’s “grubby hands.”
A related proposal from Grover Norquist used Colorado’s formula of population growth plus inflation to limit budget growth, with excess funds returned to taxpayers and the state interest rate lowered. The panel found it almost too uncontroversial to debate; Jason suggested it as a single-issue platform for an America Party focused on balancing the budget.
10. Bankruptcies are clearing delayed ZIRP failures
S&P Global counted 446 large bankruptcies through July 2025—public companies with at least $2 million of debt and private companies with at least $10 million of assets or liabilities—putting the year on pace for the most since 2010, though still nowhere near the financial-crisis peak.
Chamath rejected articles blaming tariffs because large companies do not fail within “30, 60 days” of a policy change. His explanation was temporal: zero rates and COVID-era capital filled weak businesses’ “reservoir of money,” giving structurally broken companies such as JOANN Fabrics and Party City years of extra oxygen.
A second constraint was suppressed creative destruction. Chamath argued that an anti-consolidation regulatory environment made rescue M&A harder; with those limits relaxing, distressed companies can sell valuable assets, labor can move, and more—not fewer—bankruptcies may produce a healthier allocation of resources.
Competition is also arriving from unexpected directions: Travis Kalanick’s CloudKitchens launched what was described as a strong Chipotle competitor. Chamath treated it as a specimen of new operating models attacking incumbents, not merely weak demand taking old companies down.
11. Retail leverage is visible, but CRE refinancing is the real wall
Sacks observed that a 10-year retail lease behaves like debt: the operator owes a fixed monthly payment and cannot move nimbly when traffic deteriorates. That leverage compounded the shift toward Amazon, Shein, and Target; Jason separately pointed to PE buyouts saddled with debt, while Sacks described the broader problem as zero-rate “indigestion” among overfunded businesses with negative unit economics.
Sacks did not see the bankruptcy chart as evidence of broad collapse, especially alongside a 3.3% Q2 GDP print. His concern was concentrated in rate-sensitive sectors, above all the $2.2 trillion of commercial-real-estate debt scheduled to mature before 2028.
The refinancing math was brutal: a building once worth $100 million might have supported $66 million of debt. If its value falls to $60 million, a lender may now offer only $40 million, forcing equity owners to contribute a $26 million gap while also accepting a higher rate that could turn previous positive cash flow negative.
“Blend and extend” restructurings delayed recognition; critics called them “pretend and extend.” Jason noted that roughly a third of San Francisco office space remained vacant; Sacks said some assets must return to banks and be auctioned at clearing prices so new owners can fund tenant improvements, replacing today’s stranded “zombie buildings.”
12. GPT-4b micro searched a protein space too large for straightforward brute force
Friedberg explained the Yamanaka factors as four proteins—OSKM—that can reset mature cells into stem cells. At lower doses, the aim is partial reprogramming: restore DNA repair and gene-expression networks without erasing cellular identity, potentially rejuvenating skin, eyes, brain, and muscle.
OpenAI’s GPT-4b micro was described as a reduced GPT-4 model augmented with protein sequences, biological text, and tokenized 3D structural data. The search space illustrates the attraction: the O protein alone has 360 amino acids drawn from 20 possibilities, or 20^360 possible sequences—“more than there are atoms in the universe.”
In partnership with Retro Biosciences, the model generated altered amino-acid sequences and corresponding DNA instructions that could be synthesized in bacteria or yeast. Friedberg said the resulting proteins were 50 times more effective than the original factors: above 30% of cells showed markers within seven days, and 85% expressed critical stem-cell markers by day 12.
He linked the result to the Arc Institute’s EVO2 work, where sequence patterns alone identified DNA errors and pathogenic variants. The strategic implication was that biology may support many smaller fine-tuned models that are cheaper and exceptional at narrow tasks: “Yeah, it’s working with just text.”
13. Rejuvenation has a path, a long clock, and a cancer boundary
Friedberg said companies were initially targeting specific diseases because regulators can approve a defined indication before treating aging itself as one. Programs remained around phase 1, testing human tolerability and dosing after promising animal data; his estimate for the first approved therapy using these mechanisms was “somewhere between seven and 12 years.”
Jason asked whether destinations such as Costa Rica might offer treatments within three to four years. Friedberg called that possible but highly speculative, and immediately emphasized the danger: push a cell fully back into stem-cell status and it may divide uncontrollably, producing growth that “looks like and acts like cancer.”
The scientific optimism was therefore conditional, not a prediction of imminent immortality. Better proteins may reset damaged expression networks, but successful therapy requires precise modulation—enough reprogramming to rejuvenate a cell, never enough to destroy the identity and controls that keep it safe.
14. The vaccine sidebar ended with an explicit non-answer
Asked about RFK’s changes to mRNA funding and COVID-vaccine recommendations, Friedberg declined to improvise: “I want to be more prepared for that conversation.” Jason called the rules a “moving target,” provisionally describing reduced federal development funding, narrower recommendations for healthy children and pregnant women, and physician consultation for healthy individuals.
Their broader discussion focused on institutional trust and the emotional cost of revisiting decisions about children. Friedberg described a culture of “fall in line or there’s something wrong with you asking questions,” but offered no scientific conclusion; on claimed autism-linkage data, he said he wanted to see the promised report before judging it.