Key Moments
Anthropic's Fable Backlash, Nationalizing AI, Inflation Heats Up & California’s Broken Elections
All-In PodcastWant to know something specific about what's covered?
We've already dissected every moment. Ask and we will deliver (with timestamps).
Key Moments
Anthropic's Fable 5 model promises advanced AI capabilities but faces backlash over prompted surveillance and undisclosed model downgrades, raising privacy and anti-competitive concerns.
Key Insights
Anthropic's Fable 5 model requires users to agree to 30-day prompt data retention, which applies even to enterprise customers with prior zero data retention agreements.
Anthropic has been downgrading users who it deems are conducting 'frontier AI research' (i.e., developing competitive models) without disclosure until recently, when it agreed to disclose downgrades.
The widespread restrictions by companies like Anthropic are forcing organizations to adopt open-source models, which are currently predominantly Chinese, potentially giving China a competitive advantage.
Senator Bernie Sanders proposes a 50% tax on major AI companies to fund an 'American AI Sovereign Wealth Fund,' aiming to redistribute wealth generated by AI trained on collective human intelligence.
Inflation data shows CPI at 4.2% year-over-year and PPI at 6.5% year-over-year, both at multi-year highs, primarily driven by energy prices and excess government spending.
California's election laws, including unlimited ballot harvesting and universal mail-in ballots without strict ID verification, are criticized for creating an environment where elections can become 'appointments' rather than democratic choices.
Anthropic's Fable 5 model triggers developer backlash over data retention and model downgrades
Anthropic's new "Fable 5" AI model, promising top-tier performance, has ignited significant backlash. A major point of contention is the mandatory 30-day retention of all prompt and output data, a policy even impacting enterprise clients who previously had zero data retention agreements. This extensive data collection raises privacy concerns, especially as Anthropic's stated purpose is to profile users and determine their access to model capabilities. Further fueling outrage is the practice of 'nerfing' or downgrading the model's performance for users deemed to be conducting 'frontier AI research'—essentially creating competitive AI models—without explicit notification. While Anthropic has partially walked back this policy by agreeing to disclose downgrades for certain research areas, the core practice of limiting access based on subjective criteria and undisclosed surveillance remains a significant issue. This has led to accusations of anti-competitive behavior and has prompted businesses to consider alternatives, including open-source models.
The shift towards open-source AI and its geopolitical implications
The restrictions and privacy concerns surrounding proprietary AI models like Anthropic's Fable 5 are driving a migration towards open-source alternatives. However, a critical concern highlighted is that the most advanced open-source models currently tend to be Chinese. This shift forces companies, including those in the US, to rely on AI developed in China, potentially ceding technological and economic advantages. The podcast discusses how governmental or corporate restrictions on AI development and access in the US could inadvertently strengthen China's position in the global AI race. This dynamic raises national security and economic competitiveness questions, as a dependence on foreign-developed AI could create vulnerabilities.
Bernie Sanders' proposal for an AI sovereign wealth fund
Senator Bernie Sanders has proposed the 'American AI Sovereign Wealth Fund Act,' advocating for a one-time 50% tax on the equity of major AI companies like OpenAI and Anthropic. The collected shares would form a government-managed fund, intended to redistribute the wealth generated by AI trained on collective human intelligence. Sanders argues that foundational AI models are built upon humanity's shared knowledge and creativity, which was essentially 'stolen' by a few wealthy individuals and companies. This proposal has found unlikely bipartisan sympathy, with proponents suggesting that AI companies, by acknowledging potential job losses and having been trained on publicly available data, have indeed 'asked for it.' The idea is framed as a way for the public to participate in the economic benefits of AI, mirroring concepts of sovereign wealth funds common in other nations.
Reforming Social Security through AI equity investment
David Sacks suggests a radical restructuring of the US Social Security system, proposing it be transformed into an actively managed sovereign wealth fund. Instead of holding only US Treasury bonds, the fund would invest in equities, including those of AI companies. This approach aims to enhance returns and provide American citizens with direct ownership in the companies that are shaping the future. This contrasts with the current defined contribution model prevalent in private sectors and advocates for a more robust, equity-based system that could potentially shore up the financially strained Social Security Trust Fund and provide a more substantial retirement benefit.
Inflationary pressures surge with high CPI and PPI numbers
Recent Consumer Price Index (CPI) and Producer Price Index (PPI) data for May revealed a significant uptick in inflation, with CPI hitting 4.2% year-over-year and PPI reaching 6.5% year-over-year – both marking multi-year highs. The increases are attributed to a combination of factors, including energy price shocks potentially linked to the Iran war and broader concerns about excessive government spending. This inflationary surge has surprised economists and dashed consumer expectations for a cooling cost of living. The PPI, in particular, is seen as an alarm bell, suggesting potential downstream price increases for everyday goods and services. The discussion touches on the possibility of further interest rate hikes by the Federal Reserve in response to these persistent inflationary trends.
California's election integrity and the 'legal fraud' debate
Significant concerns have been raised regarding the integrity of elections in California, particularly following the LA mayoral primary. The discussion highlights statistical anomalies, such as the shift in vote counts after Election Day, and points to specific California laws like unlimited ballot harvesting, universal mail-in balloting, and relaxed voter registration requirements (e.g., no citizenship proof, use of gym cards for ID). While these practices are legal, critics argue they create an environment ripe for 'legal fraud'—where the system's design facilitates exploitation, leading to outcomes that may not reflect the true will of the voters. The concentration of late-arriving votes for one candidate in specific areas like Skid Row, coupled with a lack of robust audit trails and penalties, fuels suspicion that elections are becoming 'appointments' rather than genuine democratic selections. This debate centers on whether the laws themselves are flawed or if illegal fraudulent activities are occurring within the legal framework.
The AI jobs debate: Productivity gains vs. job displacement
A central theme in the AI discussion is its impact on employment. While some AI leaders like Dario Amodei predict significant job losses (up to 50% for knowledge workers in the near term), others, like Chamath Palihapitiya, argue that AI's primary impact will be on the revenue side of businesses, boosting productivity and enabling the creation of new products and services. This increased productivity, they contend, will lead to job *creation* rather than destruction, as companies expand their capabilities and demand for skilled workers. The argument against widespread job loss is bolstered by recent strong employment reports. The podcast suggests that the narrative of AI-driven job destruction, often amplified by media, may be misleading and that AI is more likely to augment human capabilities and drive economic growth.
The need for critical infrastructure and control in the AI economy
The discussion emphasizes the immense capital required to build and maintain the infrastructure necessary for AI, particularly high-performance computing. The cost of a gigawatt of data center capacity is now estimated at $100 billion, creating a significant capital moat. This high barrier to entry raises concerns that a few dominant companies could control access to AI capabilities, leading to monopolies or duopolies. This scarcity of resources and the high incremental cost of AI operations (unlike the near-zero marginal cost of internet users) support the argument for government or public investment in these critical AI infrastructures. The parallel is drawn to the federal government building interstate highways, suggesting a similar role in developing foundational AI infrastructure that companies then utilize.
Mentioned in This Episode
●Products
●Software & Apps
●Companies
●Organizations
●Drugs & Medications
●Concepts
●People Referenced
LA Mayoral Primary Vote Breakdown by Ballot Type
Data extracted from this episode
| Candidate | In-person Voting (Election Day) | Mail-in Ballots (Pre-Election Day) | Mail-in Ballots (Post-Election Day) |
|---|---|---|---|
| Spencer Pratt | 35% | 28% | 19% |
| Karen Bass | 29% | 38% | 35% |
| Nithya Raman | 26% | 20% | 37% |
Odds of Reaching Higher Valuations for Companies
Data extracted from this episode
| From Valuation | To Valuation | Odds of Success |
|---|---|---|
| Unicorn ($1B) | Decacorn ($10B) | 8% |
| Decacorn ($10B) | Centicorn ($100B) | 13% |
| Centicorn ($100B) | Trillycorn ($1T) | 31% |
Common Questions
Anthropic's Fable 5 model is criticized for its high token costs, mandatory storage of all prompt data for at least 30 days (raising privacy concerns), and secretly downgrading users engaging in frontier AI research without their knowledge, which is seen as anti-competitive.
Topics
Mentioned in this video
An AI company criticized for its Fable 5 model's data retention, anti-competitive practices, and advocacy for AI regulation while expanding its own capabilities.
A cybersecurity company whose CEO used Anthropic's Mythos model to seal up vulnerabilities within their own systems.
David Friedberg's company that performs proprietary genomics work, finding AI models invaluable for gene editing and prediction but is now constrained by Anthropic's restrictions.
Mentioned as a search engine that became dominant due to its consumer-friendly, non-paid inclusion model, contrasting with early search engines that engaged in biased deals.
A pharmaceutical company mentioned in a hypothetical scenario where AI providers could favor specific corporations for competitive drug information.
A pharmaceutical company mentioned in a hypothetical scenario against Novartis, highlighting the potential for AI providers to favor specific corporations.
A financial company mentioned hypothetically to illustrate how AI providers might favor certain businesses by manipulating information access.
A financial company mentioned hypothetically alongside JPMorgan to show how AI providers could create informational biases.
The company that 'fumbled' its roll-out of the open-source LLaMA model, missing an opportunity to promote competition in AI.
An AI company included in Bernie Sanders' proposed wealth fund, and whose founder, Sam Altman, is discussed regarding his views on AI's impact on jobs.
An AI company founded by Elon Musk, listed as one of the 'largest AI companies' targeted by Bernie Sanders' proposed sovereign wealth fund.
Elon Musk's aerospace company, mentioned as a potential asset for a sovereign wealth fund and an option for investment.
Elon Musk's electric vehicle company, mentioned as a potential investment for a sovereign wealth fund.
A semiconductor company, mentioned as a company whose stock could be part of a sovereign wealth fund's portfolio.
CEO of Anthropic, described as having 'AI psychosis' with hyperbolic views on AI dangers, advocating for regulation while his company engages in restrictive practices.
Philanthropists who funded research and the output of the open-source genome language model from the ARC Institute.
A technology analyst whose experience with Fable 5 was cited, having been downgraded for asking about cancer risks and GLP-1s, and who pointed out Anthropic's hypocrisy.
A venture capitalist whose presentation on VC trends and company valuations was highly praised, providing data on "unicorn" to "trillycorn" growth.
A former Federal Reserve Governor, whose potential influence on monetary policy is mentioned in the context of rising interest rates.
Former US President, whose early political career and election to the Senate is attributed to ballot fraud.
The President of China, mentioned in the context of potential understandings with the US President regarding China's energy supply and its impact on global oil prices.
Former US Vice President, mentioned as a political figure who filed lawsuits over election results, similar to Trump.
A US Senator who proposed the 'American AI Sovereign Wealth Fund Act,' suggesting a 50% tax on the stock of large AI companies to create public ownership.
Former US President, whose views on sovereign wealth funds are noted as aligning with Bernie Sanders' proposal, and whose administration's job report numbers are cited.
A candidate in the LA mayoral primary whose voting numbers showed a suspicious decline in post-election mail-in ballots, suggesting potential election irregularities.
Governor of California, who signed a law making it virtually impossible to audit election processes, further exacerbating concerns about voter fraud.
A journalist whose video of going to Skid Row and alleging payment for votes was cited as evidence of potential election fraud.
CEO of XAI and Tesla, who has discussed AI's potential to create abundance and eliminate the need for work, a more nuanced view than Dario Amodei's.
CEO of OpenAI, initially worried about job loss from AI but later revised his stance, noting the data doesn't support widespread job loss.
Chief Business Officer of Athena, who played golf and gambled with Chamath.
A candidate in the LA mayoral primary who saw a significant increase in her vote count from post-election mail-in ballots, particularly from a specific area, raising questions about election integrity.
A political figure proposed as a candidate who could declare a state of emergency in California to address systemic issues and corruption.
Former US President, whose 1960 election against Kennedy is cited in the context of historical ballot fraud.
A candidate in the LA mayoral primary whose vote count remained relatively flat, contrasting with the suspicious shifts seen in other candidates' post-election ballots.
An individual who pled guilty in 2026 to paying people to vote, cited as proof of actual fraudulent activity.
Former US President, whose election against Nixon in 1960 is cited as an example of historical ballot fraud involving the mob.
Anthropic's latest AI model, which 'tops every benchmark' but has been criticized for high token costs, mandatory prompt data storage for 30 days, and secretly downgrading users doing frontier AI research.
An AI model, used as a point of comparison for Fable 5's token cost, suggesting Fable 5 is more expensive.
An earlier Anthropic model not publicly released due to strong hacking capabilities, later used by Palo Alto Networks to identify vulnerabilities.
A publication cited for reporting on Anthropic's changes to Fable 5 safeguards.
An open-source AI model released by Meta, whose initial handling was criticized for not maximizing its potential to compete with closed models.
A partner company that sponsored the podcast and offers a month free of service.
Elon Musk's AI system, mentioned as an alternative for engineers who want libertarian, free-market aligned work.
An institute that ingested genomic data to create a genome language model, supported by the Collison brothers, which is a powerful open-source tool for plant breeding.
A newspaper that published Bernie Sanders' op-ed titled 'AI is a public resource, you should own half of it.'
A conservative think tank mentioned as having spent millions of dollars tracking election fraud and related lawsuits.
A US government entity that manages Social Security benefits, currently holding only US Treasury certificates and proposed for reform to own equities and become an account-based system.
A regulatory agency mentioned as an example of a potential new regulatory body for AI models, suggested by Dario Amodei.
A regulatory agency mentioned as an example of a potential new regulatory body for AI models, suggested by Dario Amodei.
A central bank that raised rates by a quarter point in May, their first hike since September 2023.
More from All-In Podcast
View all 423 summaries
68 minAll-In's Best Ideas Pitch Competition: 4 Investors Present Their Top Trades Live
25 minDan Dreyfus: The Next AI Bottleneck is Copper
44 minSenators John Fetterman and Dave McCormick: Bipartisanship, Money in DC, Datacenters, Graham Platner
29 minBill Maris: How Google Could Crush AI Competitors, Why Small Funds Win, and AI's Atari Stage
Ask anything from this episode.
Save it, chat with it, and connect it to Claude or ChatGPT. Get cited answers from the actual content — and build your own knowledge base of every podcast and video you care about.
Get Started Free