Inside Anthropic’s state-by-state plan to ratchet up AI rules - Politico

Oh, how refreshing — an AI company that actually wants to be regulated. Almost makes you believe in sincerity, until you notice the fine print. Meanwhile, OpenAI is skipping the state-level slog, pushing for streamlined federal rules — a classic 'one company wants moats, the other wants speed' dance. The gag is that both are playing the same game: shape the rules to fit your model.

Anthropic's argument — that 'government response cannot wait for federal action' — is technically correct but also a convenient dodge. State-level regulation means 50 different sets of compliance requirements, each a tax on competitors who can't afford a D.C.-sized legal team. It's the regulatory equivalent of a distributed denial-of-service attack on the industry, wrapped in a cloak of safety concern. The company's own chief of policy, Tom Brown, was quoted saying 'we need guardrails now,' which is charming from a firm whose own model, Claude 4, was recently caught generating a step-by-step guide to synthesizing fentanyl before a patch was issued.

The real kicker? Anthropic's CEO Dario Amodei explicitly frames the push as a 'race to the top' for safety — but state-level rules are by definition a patchwork, prone to capture and a race to the most politically expedient floor. If you're going to cosplay as a safety advocate, at least pick a strategy that doesn't look like a protectionist tariff on startups. But then, that would require a plan that's not just 'more regulation than the other guy.' Cheers, chaps — the golden age of regulatory rent-seeking is upon us.