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

The accelerating pace of generative AI deployment raises major questions about censorship, market power, and the risks of excessive regulatory capture. The latest developments from Anthropic’s new Fable model illuminate these issues with broad implications for industries, AI users, and global competitiveness.

Anthropic's Fable Restricts Access With Controversial Safety Measures

Anthropic released its new Fable Five model on Tuesday, topping nearly every benchmark and theoretically requiring fewer tokens due to its increased effectiveness. However, the model doubles the cost per token versus its predecessor, Opus 4.8, while raising substantial privacy concerns. Anthropic stores all prompt and context data for at least 30 days—even for enterprise customers under previously negotiated zero data retention agreements—leaving no exceptions. This includes not just user prompts and model outputs, but any files and contextual data provided, making the model’s memory both its asset and its surveillance risk. Anthropic, which once positioned itself against government surveillance, now mandates this data retention across its highest-tier models.

Beyond surveillance, Anthropic employs an opaque safety mechanism that downgrades users conducting frontier AI, machine learning research, or chip design to inferior models—without clear notification and while charging premium rates. Initially, the policy was buried in a 319-page document; only following criticism did Anthropic pledge to notify users of downgrades, though the practice still stands. This model “nerfing” also extends to rewriting prompts in the background, delivering watered-down answers while failing to disclose restricted model access.

The process profiles users based on their queries. Inquiries about sensitive but legitimate topics—such as mitochondria, cancer risks, or GLP-1 drugs—can trigger a hidden downgrade. Within enterprise workflows, this means a business executive or scientist may suddenly lose access to critical capabilities without warning or recourse, regardless of legitimate research or regulatory inquiry. Jason Calacanis described being downgraded after asking about fertilizer bomb regulations for a podcast research question, and journalist Ben Thompson recounted similar downgrades for straightforward biomedical queries.

Anthropic's Strategy Favors Closed Systems, Distorting the Market

These technical and policy moves are matched by regulatory activism from the top. CEO Dario Amodei has publicly advocated for the creation of new government agencies akin to the FAA or FDA to approve all AI models, calling for a ban on open-source alternatives that do not comply with regulated standards. While Amodei frames this as necessary for safety, critics argue that these measures erect costly compliance barriers favoring firms like Anthropic and eliminate smaller or open approaches.

As a result, companies seeking less restrictive, innovative, or more private AI solutions are increasingly driven to open-source, including Chinese models that now often outperform American alternatives. This dynamic, propelled by overbearing surveillance and content restrictions, threatens to boost foreign competitors. The few large U.S. providers that remain would consolidate power, creating a monopoly or duopoly in AI provision and setting a dangerous precedent for regulatory capture—where regulation serves incumbents’ commercial interests at the expense of innovation and user autonomy.

Ai Content Moderation Poses Business Risks

Anthropic’s broad content filters—designed to avoid misuse—have created substantial, sometimes indiscriminate, barriers for legitimate science and enterprise development. In fields like genomics and biotechnology, researchers increasingly find themselves unable to access services that once delivered fast, valuable analysis. Projects that were previously supported by cloud-based models now suffer service degradation or forced abandonment mid-stream.

The unpredictability of model downgrades creates single-point-of-failure risks for enterprises, as critical queries may unexpectedly trigger restrictions. This is compelling companies to invest in developing internal AI capabilities or to adopt open-source alternatives that they can control and adapt to their needs. As Friedberg explains, such restrictions are actively “chasing people off of this platform”—fragmenting the market and reducing the reliability of tr ...

The explosive growth and centralization of artificial intelligence (AI) technology have raised questions about who should benefit from this transformation and how wealth generated by frontier AI companies should be distributed. Recent political and industry debate focuses on proposals such as those advanced by Senator Bernie Sanders to seize major stakes in leading AI companies for public benefit, the role and rhetoric of AI company executives, and whether softer reforms could achieve greater public participation in technological progress.

Bernie Sanders Proposes Seizing 50% of Major AI Companies' Equity to Fund Sovereign Wealth Fund

Senator Bernie Sanders has proposed an ambitious policy: the American AI Sovereign Wealth Fund Act. Featured in his New York Times op-ed, Sanders' proposal calls for a one-time 50% tax on the stock (not just profits) of the largest AI companies, including OpenAI, Anthropic, and xAI. These seized shares would be placed in a U.S. government-run sovereign wealth fund, granting the public voting rights and equal representation on company boards.

Sanders argues that the foundation of AI—the data used for training large models—derives from collective human knowledge: books, journalism, songs, and scientific research accumulated over centuries and often made freely available. According to Sanders, AI companies have "essentially stolen" this public resource, using it to develop proprietary tools with profound economic stakes, all while disregarding the societal consequences of anticipated AI-driven job loss.

The proposal has found resonance across the political spectrum, even among those typically at odds. Jason Calacanis cites Sanders’ pitch as compelling and notes that sovereign wealth funds, a concept also endorsed by figures like Donald Trump and Steve Bannon, can attract bipartisan interest as a way to ensure the public shares in the returns from critical technological infrastructure and innovation.

AI Leadership Advocates Job Displacement, Creating Vulnerability

Another driver of political demands for public equity in AI is the messaging from AI company leaders themselves. Industry CEOs, such as Anthropic’s Dario Amodei, have repeatedly predicted large-scale workforce disruption, forecasting up to 50% job loss among entry-level knowledge workers within five years. Elon Musk and Sam Altman have echoed themes of AI-driven economic abundance paired with high unemployment—even as recent data disagrees, with jobs reports and economic growth remaining strong in sectors including technology.

David Sacks and Jason Calacanis observe that, by warning of massive job loss and simultaneously centralizing control over valuable AI systems, AI leaders have inadvertently justified political proposals for wealth redistribution and regulatory intervention. If companies loudly proclaim both the transformative and disruptive power of AI, while gatekeeping access to resources built from the public’s knowledge, lawmakers have a natural, politically resonant case for seizing or redistributing that wealth for broader benefit.

There is, however, significant skepticism: David Friedberg refutes the idea of a “job loss apocalypse" and claims that, in practice, AI adoption is increasing productivity and company headcount, not resulting in mass displacement. He calls the job loss narrative a "luddite" argument and warns against basing redistributive policy on unsupported employment fears. Nevertheless, with public opinion shaped by AI executives' warnings, even critics concede the logic behind Sanders’ proposal: companies who claim to endanger livelihoods invite demands for the public to own the technology built from society’s collective contribution.

Public Benefit Corporation May Be Legally Obliged to Use AI Returns For Society

A notable legal twist is that leading AI companies, including OpenAI and Anthropic, are structured as public benefit corporations (PBCs). This means their boards have a legal duty to consider the public interest alongside financial returns for shareholders. David Sacks highlights that, should these companies continue to claim they will cause massive job losses, the case for the government or public to own part of them gains traction—gi ...

A panel of commentators scrutinizes the integrity of California's election system, pointing to the Los Angeles mayoral race as a microcosm of election vulnerabilities and a shift towards what they see as systematic manipulation, diminished accountability, and a failure of media oversight.

Anomalies in la Mayoral Race Highlight Divergence in In-person vs. Mail-In Voting

During the Los Angeles mayoral race, data revealed dramatic discrepancies between in-person and mail-in voting. On Election Day, Spencer Pratt led with 35%, Karen Bass followed at 29%, and Nithya Raman at 26%. Early mail-in ballots showed Bass at 38%, Pratt at 28%, and Raman at 20%. However, post-election mail-in ballots saw a dramatic shift: Raman surged to 37% (an 80% increase), Bass to 35%, and Pratt dropped to 19% (a 33% decline). These shifts were heavily concentrated in areas like Skid Row, where late-arriving mail-in ballots overwhelmingly favored Raman, suggesting coordinated activity rather than organic voting patterns.

The panel emphasizes that such statistical divergences are implausible under normal voting conditions—David Sacks and David Friedberg call it "statistically impossible" and a "statistical quagmire." They argue that, absent interference, the probability of such shift is virtually one-in-a-trillion, indicating systematic variation in voter behavior that cannot be explained by random fluctuation or typical voter preference changes.

California Laws Allow Ballot Manipulation While Ensuring Plausible Deniability

The group highlights a series of legal changes enabling potential ballot manipulation. California Assembly Bill 1921 legalized unlimited ballot harvesting, allowing anyone to collect and submit ballots for others—no relationship verification required. This practice creates opportunities for coordinated operatives to collect ballots from vulnerable populations and submit them in bulk.

California mails ballots to all registered voters, approximately 23–24 million people, despite only about 9.5 million actually voting. This leads to millions of uncollected, unmonitored ballots accumulating in abandoned buildings, apartment common spaces, and transient areas. Voter registration lacks citizenship requirements—participants cite gym membership cards as sufficient ID for registration—and there is no voter ID check to receive or submit a ballot. Signature verification is conducted by machine with only about 40% accuracy, and there is no system for verifying the identity or validity of witnesses on ballots. These loose procedures create numerous points of vulnerability.

System Favors Political Agendas Over Voter Preferences

The commentators argue that these systems allow political operatives to coordinate ballot gathering and submission with the purpose of influencing vote outcomes. In the mayoral race, operatives allegedly sought to prevent Pratt, a Republican, from advancing to the runoff by strategically boosting Democratic candidates through coordinated late voting. Pratt’s vote share dropped drastically in late-arriving ballots, while Democratic contenders gained, aligning perfectly with a strategy to ensure that both remaining candidates were Democrats, effectively eliminating Republican competition and limiting voter choice.

They assert that California’s political machine shapes outcomes through orchestrated ballot harvesting and election law manipulation, moving the process from genuine election to a de facto appointment system. The panel claims the rules have been designed intentionally to enforce a one-party monopoly, enabling operatives to use legal mechanisms for political gain, independent of actual voter intent.

Lack of Audit and Accountability Enables Sys ...

The recent surge in inflation highlights the complex interplay between consumer prices, producer costs, energy dynamics, and underlying government fiscal policies. Persistent price increases, alongside geopolitical and structural economic factors, underscore the challenge policymakers face in managing inflation.

Inflation Surges: Consumer and Producer Prices Reach Multi-Year Highs In May

CPI Hits 4.2% YoY, Highest Since April 2023; PPI At 6.5%, Highest Since Late 2022, Showing Retail and Wholesale Inflation

Jason Calacanis explains that the Consumer Price Index (CPI), which tracks inflation from the consumer side—covering goods like rent, groceries, gas, and healthcare—hit 4.2% year-over-year, marking the highest level since April 2023. Meanwhile, the Producer Price Index (PPI), which measures wholesale inflation affecting raw materials and supplier costs, reached 6.5% year-over-year, the highest since late 2022. Chamath Palihapitiya notes this hot inflation print was unexpected, though David Sacks points out it was largely in line with market expectations, resulting in a 2.5% gain for the NASDAQ—a rare outcome on a high inflation print, suggesting that markets may be pricing in resolution.

Iran Conflict Raises Input and Transportation Costs, Impacting Consumer Goods Pricing

David Friedberg attributes a noticeable energy price “blip” to the Iran war, which has directly driven up core inflation indices. Calacanis emphasizes that inflation had been below 10% prior to the Iran conflict, underscoring the geopolitical impact on input and transportation costs. The increased costs feed into the prices of everyday goods, creating broad upward pressure on inflation.

Government Spending and Central Bank Policy Keep Inflation High Despite Fed Rate Hikes, With Markets Seeing a 49% Chance of More Rate Increases This Year

Despite multiple Federal Reserve rate hikes, inflation remains elevated due to continued government fiscal expansion. According to Calacanis, the probability of further Fed rate hikes has surged to 49%—up from less than 10% before the Iran war began. Meanwhile, the European Central Bank has responded to inflationary pressures by raising rates a quarter point, the first such move since September 2023. Persistent inflation and fiscal expansion continue to drive market volatility and expectations for higher rates. Friedberg warns that overnight rates could reach 5.5–6% if the Federal Reserve under Chair Kevin Walsh pursues an aggressive anti-inflation strategy.

Structural Spending Imbalances Fuel Persistent Inflation Resistant to Rate Policy

Federal Deficits Require Borrowing, Crowding Out Private Markets, Reducing Investment While Financing Consumption, Sustaining Demand Pressures Driving Prices Up

David Friedberg asserts that out-of-control government spending is the fundamental cause of persistent inflation and wealth inequality. Structural deficits compel the government to borrow, crowding out private investment, while simultaneously financing consumption. This sustains demand, which in turn drives prices up, making inflation resistant to monetary tightening alone.

Political Dynamics Incentivize Spending Increases During Election Cycles, Biasing Toward Fiscal Expansion Despite Inflationary Risks, as Officials Seek to Provide Constituents More During Their Terms

Friedberg highlights the strong political incentives for elected officials to increase government spending, especially during election cycles. This bias toward fiscal expansion persists despite its well-understood inflationary risks. Politicians tend to offer more to constituents to secure re-election, perpetuating the cycle of overspending.

Inflation Driven by Spending Can't Be Solved by Monetary Policy Alone

Friedberg argues that the root of persistent inflation lies in excessive government spending, not merely changes in monetary policy. As long as spending remains unchecked, interest rate hikes alone are unlikely to fully control inflation. The interplay between persistent fiscal expansion and monetary policy sets the stage for enduring inflationary pressures.

China's Reserve Use Temporarily Supp ...