Anthropic CEO Fires Back at Silicon Valley Critics Over AI Governance

Anthropic CEO Dario Amodei broke from his usual social media silence on August 15 to rebut a prominent investor's claim that his AI safety messaging has damaged the industry — dismissing the framing as "an overly simplified picture of the world."

The dispute began when investor Gavin Baker, on the All-In Podcast, claimed multiple trusted sources had heard Amodei privately suggest Anthropic could become "the only company left in the world." Anthropic researcher Sholto Douglas publicly called the claim "completely false." Baker then broadened his attack: Amodei's AI risk warnings, he argued, had failed to produce meaningful regulation and were now being weaponized by anti-datacenter advocates.

Amodei rejected Baker's core premise — that regulation inevitably concentrates AI power in the hands of a few companies and politicians. He argued that fair institutions vest power in rules rather than people, drawing an analogy to the formal court system, which he said defends individual rights better than mob justice.

He cited Anthropic's policy record as evidence. California's SB 53, which the company backed, imposes its heaviest obligations only on firms above $500 million in annual gross revenues — a tiered structure that, Amodei argued, advantages smaller competitors over frontier labs like Anthropic itself. The same logic, he said, applies to the testing framework Anthropic has pushed at CAISI (the Center for AI Standards and Innovation) and the White House, which applies stricter scrutiny to the most powerful models.

He also rejected the claim that his regulatory push had failed. The Trump administration's reported approach — pre-deployment testing for frontier models, with additional oversight for advanced open-weights releases — is "one that I am very supportive of," he said, contrasting it with six months ago, when most of the industry was pushing for federal preemption of all state AI regulation.

Amodei Says Public Distrust Runs Deeper Than AI Messaging

Baker urged Amodei to adopt a more optimistic public posture now that strict regulation appears off the table. Amodei declined. Public skepticism toward AI, he argued, is a decades-old crisis of institutional trust — not a messaging problem fixable with "a glitzy marketing campaign with a positive spin." "Saying that AI will cure cancer is more a cliché than it is inspiring, and most people think it is deceptive," he wrote.

He also disputed the charge that his messaging has been disproportionately negative, pointing to his essay "Machines of Loving Grace," which argues AI could help cure most human diseases within five to ten years. Pessimistic social media clips, he noted, reflect curation rather than his overall message: "Short clips from my interviews that end up on social media tend to be disproportionately negative, as that gets clicks."

Amodei's sharpest concession cut in a different direction: "By far the most accurate criticism of AI companies including Anthropic is that we haven't yet delivered on our big promises to benefit the world. That is totally on us."

He gave a personal account of what is at stake. His father died of Hepatitis C just before sofosbuvir — a direct-acting antiviral that now cures 95% of patients — reached the market. "I feel the urgency here," he wrote. Anthropic, he said, is rapidly expanding its biology and medicine work and expects to show "early glimmers" in the coming months. When results are ready, "the whole world will hear about it, as loudly as possible," he pledged.

The exchange marks one of the most substantive public debates over AI governance philosophy to emerge from Silicon Valley in recent months.

Sources: 

X posts by @_sholtodouglas, @GavinSBaker, @DarioAmodei ; California's SB 53: The First Frontier AI Law, Explained – Future of Privacy Forum 

（Related： Opinion | Kimi K3 and the AI Cold War's Costlier New Battlefield ｜ Latest ）