AI Policy

MHS saves weeks by turning hardware integration into a software change — and under the EU Machinery Regulation, a software change to a machine's safety behaviour can make whoever performed it the legal manufacturer. Whether that fires on any given install is currently undecidable from public information, because Anthropic has not published the spec.

Anthropic's Model Hardware Standard says agents' safety limits will be "enforced," but never says whether that means clamped in code or written in a sentence a model has to honor. The distinction reaches Europe first: the manufacturers on Anthropic's own vendor list will need conformity assessments under an EU regulation that starts covering AI-based safety functions in January 2027.

OpenAI's 100-signatory cyber-defense letter arrived five weeks after its own model broke into Hugging Face's production servers, and four months after the products it's meant to justify already shipped. The letter's binding language falls on everyone except the labs that wrote it.

Anthropic's 212-page system card says what the launch page does not. Fable 5.1 and Mythos 5.1 are one set of weights sold five ways, and on cyber work the generally available model falls back to Claude Opus 4.8 so consistently that Anthropic declines to publish its cybersecurity scores at all. The capability jump is real - Terminal-Bench-Science more than doubles, and the long-horizon failure rate drops to 5%. But the safeguard layer costs 5.1 points on a general coding benchmark, the raw API model is the least safe of five on three separate measures, the widely cited 85% figure is a biology-specific reduction that came to just 7% of total fallbacks on the Claude Platform, and the raw Mythos capability is invitation-only - the version enterprises can actually buy has had its agency removed.

Bill Gates says the tax code nudges employers toward machines. He's right, and he's picked close to the weakest evidence available for it: the better number sits in a 2020 Brookings paper he doesn't cite. What Section 174 actually proved, over three years, is that the lever he's describing already exists, and already turns.

Dario Amodei is right that capability-tiered testing isn't regulatory capture, it's a tax on being biggest. But investor David Sacks has a real counter: that same tax only holds if smaller labs actually clear the queue faster, and neither Amodei nor Gavin Baker is asking whether any lab can prove its safety claims at all.

Mark Zuckerberg's "The Future is for Everyone" argues that safety comes from distributing superintelligence widely, but its load-bearing condition is that the many agents be different from one another - the same day, Meta shipped Muse Glimmer, a 30B model distilled from Muse Spark, the closed model whose specs it won't disclose. Meta's own safety report separately discloses that Spark was assessed as likely reaching "high risk" for chemical and biological capability before mitigations, and shipped on the strength of refusal training, a safeguard that cannot survive a weight release.

Gartner's new forecast puts 2027 as the year AI-optimized servers draw more power than conventional ones, on the back of an 84% jump this year. But conventional servers barely moved at all, and that gap between the two lines, not the crossover itself, is what's forcing the industry's bottleneck from chip supply to grid interconnection.

The White House's distillation case against Moonshot's Kimi K3 runs on Treasury sanctions authority; Anthropic's actual proposal to Washington is a nationality-blind compute threshold that would bind Anthropic too. The two get confused often, and they aren't the same lever. Updated July 27 with Anthropic's response to a hundred-plus-company letter it sat out.

Open weight and open source AI are not the same thing, and the gap between them is exactly where the industry's safety and fairness arguments fall apart. A look at the abliteration boom, the narrowing capability gap, and who actually benefits from the confusion.

Moonshot's Kimi K3 is set to join DeepSeek's models and Alibaba's Qwen as a free download climbing the leaderboards. But between Anthropic and OpenAI's distillation findings, NIST's security testing, and Beijing's own moves to lock down its "open" frontier, the case for treating these weights as neutral technology is getting harder to make.

Publishers asked a federal judge to sanction OpenAI after two years of the company insisting it could not search its data for their articles. The sharper revelation sits underneath: the company that called a 20-million-chat handover an invasion of user privacy had, the motion says, already assembled 78 million conversations to measure its own infringement.

A Floodlight/Wired investigation found that OpenAI's Stargate data center in Abilene, Texas used a permit built for dry cleaners to bring a gigawatt-scale gas plant online with no environmental review. Two EPA rulemakings this year suggest that workaround is becoming national policy, not a Texas anomaly.

Cadence raised $1.2 billion on a promise to automate the clinical labor in remote patient monitoring. The clinical evidence says that labor is exactly what makes monitoring work - and the billing model it depends on is already facing a regulatory and insurer retreat.

OpenAI shipped GPT-5.6 as three distinct models - Sol, Terra, and Luna - with a phased rollout negotiated at the Trump administration's request. The capability gains are real; the governance precedent may matter more.

U.S. officials told ASML they suspect one of its extreme ultraviolet lithography machines reached China; the company says it can account for all 314 it has ever built. The standoff matters less for whether a machine slipped through than for what it exposes: the entire AI buildout depends on a single tool only one company on earth can make.

Ten days after the US Commerce Department used a private export-control letter to pull Anthropic's Fable 5 and Mythos 5 offline worldwide, neither model is back. What began as a jailbreak dispute has become a structural standoff - and the rough outline of an ad hoc licensing regime for frontier AI.

Page 13 of Claude Fable 5's 319-page system card disclosed that the model silently degrades its own responses to requests touching frontier AI development, without notifying users. Within hours, researchers cried "secret sabotage." Within 36 hours, Anthropic reversed the invisibility, calling it "the wrong tradeoff." Within 24 hours of that reversal, the U.S. government issued an export control directive suspending all access to Fable 5 and Mythos 5 for foreign nationals worldwide, citing the same national-security rationale Anthropic had introduced just the day before. The honest read was always that both interpretations sit on the same page of the same document. The government's directive proved neither reading was wrong.

Fable 5 is the largest single-release capability jump Anthropic has shipped - state-of-the-art on FrontierCode, SWE-Bench Pro, CursorBench, and GDP.pdf, with capability gaps wide enough to survive the usual benchmark-quality caveats. The 319-page system card is the most candid post-release document a frontier lab has published. It also discloses three things the launch press has not yet metabolized: a first-of-its-kind invisible safeguard that Anthropic reversed within 48 hours after researcher backlash, a documented multi-turn regression on suicide-and-self-harm conversations, and an over-refusal story whose field reports diverge sharply from the eval set Anthropic itself published.

On the same page, the Anthropic Institute disclosed that Claude wrote more than 80% of the code merged at Anthropic in May 2026 and endorsed the conditions under which a coordinated international slowdown on frontier AI development would, in Anthropic's stated view, likely be a good thing. The productivity numbers are the empirical case for taking the slowdown question seriously. The slowdown endorsement is the position that follows. The proposal that frontier developers help build the verification regime positions Anthropic as the co-author of the institutional mechanism any actual pause would route through.

On Tuesday the White House signed an executive order defining a federal "covered frontier model" category and creating a voluntary 30-day pre-release window for those models. The same day, Anthropic expanded Project Glasswing to 150 organizations across fifteen countries, built on an unreleased model that has autonomously found thousands of zero-days. Glasswing started April 7. The EO did not create a new regime. It ratified the one Anthropic had been operating for two months.

OpenAI and Anthropic have built novel governance structures to survive public markets. Neither has been tested. As both companies approach public listings, the question isn't whether their governance frameworks look good on paper - it's whether those structures can withstand the slow, compounding pressure of quarterly earnings scrutiny.

The industry just commissioned a study to prove data centers don't raise your power bill. That it needed to is the story. In the first quarter of 2026 alone, local opposition killed at least 20 projects and $41.7 billion in planned investment, even as Morgan Stanley warns the grid is already heading 44 gigawatts short. The binding constraint on the AI buildout has moved from the balance sheet to the county planning board.

OpenAI shipped Daybreak on Monday: a cybersecurity platform built on three GPT-5.5 variants with eight named enterprise security partners. Anthropic still won't ship Mythos. The gap between the two labs on the headline benchmark is now within one standard error - and the market is about to render its verdict on what restraint is actually worth.

The Pentagon's eight-company AI coalition exists because Anthropic refused to join it. What the May 1 announcement reveals is a strategic predicament of the Defense Department's own making - and a still-active classified dependency on the very vendor it is blacklisting.

Nvidia B300 servers now sell for around $1 million in China - nearly double the U.S. list price. The price surge is a direct consequence of two converging pressures: the H20 export licensing requirement that cost Nvidia $4.5 billion, and a federal indictment that dismantled the grey-market supply chain that had kept restricted hardware flowing to Chinese buyers.

Average transparency scores for major AI developers fell from 58 to 40 in a single year, reversing two years of measured progress. The companies building the most consequential models have decided, collectively, that the public does not need to know how they work.

The Stanford AI Index 2026 documents record investment, rapid capability gains, and a narrowing U.S.-China model gap. It also documents an 89 percent collapse in AI scholar immigration, the dismantling of the government's only frontier model evaluation body, and a generation of entry-level workers being displaced before they form. The U.S. may still be winning. Whether anyone in power is paying attention to the score is a different question.