Thumbnail for Anthropic's Fable Backlash, Nationalizing AI, Inflation Heats Up & California’s Broken Elections

Anthropic's Fable Backlash, Nationalizing AI, Inflation Heats Up & California’s Broken Elections


Episode Details
Channel

All-In Podcast

Published

6/12/2026

Episode Summary

The All-In Podcast episode reunites the core host quartet of Jason Calacanis, Chamath Palihapitiya, David Sacks, and David Friedberg for an extensive discussion spanning technology, finance, and politics. They begin by unpacking the backlash surrounding anthropic and its newly released Fable 5 AI model. With CEO Dario Amodei back to blogging, the hosts analyze the model's performance—which exceeds the older Opus 4.8—and its rigorous safety safeguards. Though praised for its capabilities (having been validated by the CEO of Palo Alto Networks), the model's 30-day data retention and aggressive monitoring for frontier AI research have sparked major concerns over AI Censorship. Chamath Palihapitiya notes the significant business risk of accidental censorship. In response, David Friedberg highlights how companies are being forced to adopt Open source AI, noting that currently the best models are coming from China. The hosts discuss the hardware implications, touching on the value of running models locally on Apple Silicon. They also reference Meta's competitive fumble and subsequent recovery with Llama, which stands as a prime example of accessible AI. Debating the necessity of guardrails, the hosts compare the evolution of AI to the dual-use nature of the Manhattan Project and question why basic KYC/AML checks aren't used instead of opaque algorithmic downgrading. The conversation then shifts to the political response to AI's rapid ascent. They discuss a controversial proposal by Senator Bernie Sanders to seize 50 percent equity from leading AI companies to create a Sovereign Wealth Fund. This fund would theoretically tax leaders like OpenAI and xAI, aiming to redistribute wealth amidst fears of AI-driven job displacement. While David Sacks fundamentally opposes property confiscation, he understands the populist reaction, noting that prominent CEOs like Sam Altman and Elon Musk (who runs xAI and SpaceX) have fueled fears by predicting massive job losses. Even Donald Trump has previously favored the concept of sovereign wealth funds. Looking at alternative ways to build national wealth without seizures, David Friedberg proposes revamping the US Social Security trust fund by allowing it to invest in equities rather than exclusively holding US Treasuries. Reflecting on their own recent ecosystem events, the hosts recount the tremendous success of the All-In Liquidity Conference. They highlight a standout presentation by Thomas Laffont on the power laws of Venture Capital, specifically noting the statistical odds of unicorns evolving into decacorns, centicorns, and eventually Trillicorns. The conference also featured exclusive networking events, including a speaker dinner at the prestigious French Laundry hosted by the New York Stock Exchange, and a meeting hub sponsored by EY. Notable attendees included influencer and investor Jake Paul. Addressing the macroeconomy, the hosts react to unexpectedly hot numbers for both CPI (Consumer Price Index) and PPI (Producer Price Index). They attribute this stubborn inflation largely to global supply shocks, particularly the ongoing War with Iran, though they note that China's economic strategy of smoothing energy consumption has prevented a complete blowout in oil prices. Looking ahead, they assess the likelihood of rate hikes by the Federal Reserve—potentially influenced by a future leader like Kevin Warsh—and contrast this with the recent rate cuts by the European Central Bank. Finally, the podcast delves into the controversy surrounding recent elections in Los Angeles. The hosts examine the delayed vote count in the mayoral primary involving Spencer Pratt, Karen Bass, and Nithya Raman, where massive late swings heavily favored Raman. They attribute this statistical anomaly not necessarily to illegal fraud, but to legalized structural loopholes such as California Assembly Bill 1921, which institutionalized unlimited Ballot harvesting. They criticize Governor Gavin Newsom for signing laws that severely restrict election audits, creating a system that critics, including organizations like the Heritage Foundation, argue acts more like an appointment process than a democratic election.

Investment Ideas
4 ideas
3 high confidence
1 medium confidence

This episode highlights critical shifts in the AI landscape, marked by developer backlash against closed-source model censorship and data retention, alongside a growing corporate pivot toward open-source alternatives. Additionally, stubborn inflation prints (CPI and PPI) driven by geopolitical friction point to a higher-for-longer interest rate environment, while venture capital data reveals distinct power laws favoring late-stage mega-cap investments.

Portfolio lens: This set of ideas represents a macro hedge and technology infrastructure basket designed to navigate persistent inflation, rising capital costs, and the structural shift toward local, open-source enterprise AI.

Generated with gemini-3.5-flash on 6/25/2026, 8:50:26 AM. For research only. Not financial advice.
Enterprise Pivot to Open-Source AI and Edge Execution
high confidence
Sector Theme
Time horizon: medium (1-3 years) as enterprises re-architect their AI pipelines and upgrade local hardware
Thesis

Enterprise backlash against closed-source AI censorship, data retention, and silent downgrades will accelerate the adoption of open-source models run locally on edge hardware.

Rationale

Anthropic's Fable 5 release sparked outrage due to mandatory 30-day data retention and silent downgrading of users conducting frontier research. This is forcing companies to transition to open-source models run locally on proprietary data to maintain control and avoid single-point-of-failure risks.

Evidence
  • Anthropic stores all prompt data for at least 30 days and downgrades users doing frontier AI research without explicit notification.
  • David Friedberg noted that bioweapon-related restrictions are locking out legitimate genomic research, forcing his company to transition to open-source models run locally.
Catalysts
  • Release of highly competitive open-source models from US or Chinese developers.
  • Increased enterprise deployment of local hardware, such as Apple Silicon, optimized for LLM execution.
Risks
  • Regulatory capture or government bans targeting open-source AI distribution under the guise of safety.
  • Aggressive price cuts by closed-source providers making local execution economically unviable.
Next Diligence
  • Analyze enterprise adoption rates of open-source APIs versus closed-source models.
  • Evaluate performance benchmarks of local LLM execution on M-series Apple Silicon.
Anthropic
OpenAI
Meta
Apple
Llama
Gigawatt-Scale AI Power and Data Center Infrastructure Moats
high confidence
Macro
Time horizon: long (3-5+ years) due to the multi-year timeline of power grid connection and data center construction
Thesis

The massive capital and power requirements of frontier AI models will create a highly concentrated infrastructure bottleneck, heavily favoring zoned land and energy providers.

Rationale

Frontier AI models require massive physical infrastructure. Chamath Palihapitiya notes that a single gigawatt of data center capacity now costs $100 billion, making power access and land zoning the ultimate competitive moat.

Evidence
  • Chamath stated that a gigawatt of capacity now costs $100 billion, up 20x from previous estimates of $4-5 billion.
  • The overwhelming majority of megawatts are still directed to the big closed-source models.
Catalysts
  • Approval of new multi-gigawatt data center sites with secured grid connections.
  • Utility companies signing massive power purchase agreements (PPAs) with technology giants.
Risks
  • High interest rates increasing the cost of capital for $100B+ infrastructure projects.
  • Rapid algorithmic efficiency gains reducing the absolute compute and power required per model run.
Next Diligence
  • Track utility grid capacity queues in key data center markets like Arizona and Virginia.
  • Evaluate capital expenditure plans of major infrastructure asset managers.
Blackstone
Brookfield
Google
xAI
OpenAI
Hedging Persistent Inflation and Higher-for-Longer Interest Rates
medium confidence
Macro
Time horizon: short to medium (6-18 months) based on macroeconomic data prints and geopolitical developments
Thesis

Stubbornly high CPI and PPI, driven by geopolitical supply shocks, will force central banks to keep interest rates elevated, favoring short-duration assets and energy hedges.

Rationale

CPI came in hot at 4.2% and PPI at 6.5%, driven by energy blips from the Iran war. The hosts predict overnight rates could reach 5.5% to 6.0% under a potential Kevin Warsh Fed, contradicting expectations of rapid rate cuts.

Evidence
  • CPI came in at 4.2% year-over-year and PPI came in at 6.5% year-over-year.
  • David Friedberg predicted overnight rates could reach 5.5% to 6% under a Kevin Warsh Fed.
Catalysts
  • Further escalation in the Iran war driving oil prices higher.
  • China returning to the spot market to buy oil, potentially pushing prices to $150-$200 per barrel.
Risks
  • A sudden diplomatic resolution to the Iran war leading to an energy price collapse.
  • A rapid economic slowdown forcing central banks to cut rates despite inflation.
Next Diligence
  • Monitor weekly oil inventory data and Chinese energy import statistics.
  • Track Fed funds futures pricing for potential rate hikes.
Federal Reserve
European Central Bank
China
Iran
Kevin Warsh
Late-Stage Growth Equity and Trillicorn Hunting
high confidence
Private Market
Time horizon: long (3-7 years) to realize late-stage growth equity returns
Thesis

Capital allocation should shift toward late-stage decacorns and centicorns rather than early-stage unicorns, as historical power laws show a significantly higher probability of massive scale-up at later stages.

Rationale

Data presented by Thomas Laffont at the Liquidity Conference shows that the probability of a centicorn reaching a trillion-dollar valuation (31%) is much higher than a unicorn reaching a decacorn (8%).

Evidence
  • The odds of a unicorn getting to a decacorn is 8%, a decacorn to a centicorn is 13%, and a centicorn to a trillion-dollar market cap is 31%.
Catalysts
  • IPO window reopening for mega-cap private companies like SpaceX or OpenAI.
  • Increased secondary market liquidity for late-stage private equities.
Risks
  • High interest rates compressing late-stage valuation multiples.
  • Regulatory antitrust actions blocking mega-mergers or massive scale-ups.
Next Diligence
  • Review secondary market pricing and volume for top-tier centicorns.
  • Analyze historical IPO performance of companies valued over $50 billion.
SpaceX
OpenAI
Anthropic
Thomas Laffont
Watchlist
  • Apple Silicon M-series performance benchmarks
  • Anthropic Fable 5 API usage metrics
  • CPI and PPI monthly prints
  • Global crude oil spot prices
  • SpaceX and OpenAI secondary market valuations
  • California voter ID ballot measure progress
Open Questions
  • Will US regulatory agencies successfully implement restrictions on open-source AI models under the guise of national security?
  • How long can China continue to smooth global energy consumption before being forced back into the spot oil market?
  • Will the US Social Security trust fund ever be legally permitted to invest in public equities?
  • What are the exact data-retention and user-downgrading policies of competing closed-source LLM providers?
Key Topics & People

Host of the All-In Podcast conducting the interview with Ryan Cohen.

US President who announced the brokering of a comprehensive Middle Eastern peace deal.

OpenAI
OpenAI
Organization

Leading AI company mentioned in contrast to Anthropic's political alignments.

Freely available AI models that are seen as a vital counterbalance to closed, regulated AI monopolies.

CEO of Anthropic, heavily critiqued for his political maneuvering and ideology regarding AI safety.

anthropic
Organization

Leading AI frontier lab that faced government pushback and was forced to shut down its Fable 5 model.

Fable 5
Technology

Anthropic's latest AI model, pulled offline following government intervention over jailbreak vulnerabilities.

SpaceX
SpaceX
Organization

Aerospace company that completed a massive IPO pushing its valuation over $2 trillion.

Elon Musk
Elon Musk
Person

CEO of SpaceX, recently labeled the world's first trillionaire based on the private aerospace company's soaring valuation.

US Senator grouped with the 'new oligarchs' trying to coalesce power and control the American economy.

Venture capitalist and podcast host who criticizes the behavior of frontier AI labs.

Entrepreneur and host of the podcast, known for his political, geopolitical, and venture capital insights.

Host of the All-In Podcast, referred to as Bestie or JCal, who moderates the discussion.

The restriction and shaping of AI outputs by model providers.

Opus 4.8
Technology

An earlier AI model compared to Fable 5.

Historical project used as an analogy for the development of powerful AI.

Companies reaching a trillion-dollar market capitalization.

Famous restaurant that hosted an event for the Liquidity conference.

A measure of inflation reflecting wholesale goods and raw materials.

EY
Organization

Professional services firm that sponsored the meeting hub at Liquidity.

Jake Paul
Jake Paul
Person

Influencer and investor who attended the Liquidity conference.

A California law legalizing the practice of ballot harvesting.

The gathering and submitting of absentee or mail-in voter ballots.

The central bank for the euro area.

Conservative think tank tracking election integrity.

Governor of California who signed laws regarding election audits.

Candidate in the Los Angeles mayoral election.

Candidate in the Los Angeles mayoral election.

Candidate in the Los Angeles mayoral election.

Los Angeles
Los Angeles
Location

City in California holding a disputed mayoral election.

A potential candidate for the Federal Reserve.

Federal Reserve
Federal Reserve
Organization

The central banking system of the United States.

China's handling of energy consumption and global market impact.

Ongoing geopolitical conflict impacting global inflation.

A measure of inflation reflecting consumer prices.

Stock exchange that hosted a dinner at the Liquidity conference.

Investment funding provided to startups and early-stage companies.

Investor who presented at the Liquidity conference.

A conference hosted by the podcast focusing on capital allocation.

Bonds issued by the US government.

US government retirement fund proposed for equity investments.

CEO of OpenAI.

xAI
Organization

An AI company founded by Elon Musk.

A proposed government fund holding equity in major AI companies.

KYC/AML
Topic

Know Your Customer and Anti-Money Laundering verification protocols.

Llama
Llama
Technology

An open source AI model developed by Meta.

Meta
Meta
Organization

Technology company that developed the Llama AI model.

Apple Silicon
Apple Silicon
Technology

Apple's proprietary hardware used to run local AI models.

China
China
PoliticalEntity

Country producing competitive open source AI models.

Palo Alto Networks
Palo Alto Networks
Organization

Cybersecurity company whose CEO tested Anthropic's models.