Dario Amodei's Blunt Message to Congress: Three Things to Do Now

Anthropic’s CEO sits down with Axios co-founders Jim VandeHei and Mike Allen for a follow-up to his essay “The Adolescence of Technology,” distilling his 72-hour manifesto into a direct, no-hedging message for lawmakers. The conversation is short, pointed, and surprisingly specific about what Congress should actually do.

The Three-Point Agenda

Amodei doesn’t hedge. When asked what Congress should do right now, he offers three concrete actions:

First: Transparency legislation. Require AI companies to disclose what tests they run, what risks they find, and label their products accordingly. The key argument isn’t just consumer protection; it’s that if each company studies risks in isolation, afraid to share findings because of competition, the scientific community can’t learn collectively. Mandatory disclosure turns competitive secrecy into shared knowledge.

Second: Cut off the chip supply chain to authoritarian adversaries. Amodei frames this as a prerequisite for everything else. The US is years ahead in chips, and that lead creates a buffer to handle AI dangers properly. Without it, any domestic regulation faces the impossible dilemma of slowing down while a rival keeps accelerating.

“It’s hard enough between the companies in the US to handle this crazy commercial race. But it’s almost impossible to do that if we have an authoritarian adversary who’s out there building the technology almost as fast as we are.”

Third: Rethink wealth distribution for a post-AI economy. This is where Amodei gets uncomfortable, even for himself. AI will create enormous economic growth (curing cancer, cheaper energy, new materials), but precisely because it replaces white-collar work, wealth will concentrate from labor to capital. The result: a world of unprecedented abundance with a severe distribution problem.

The Trillionaire Problem

The most striking part of the conversation is how directly Amodei names the problem that most tech leaders avoid. We’re heading toward a world of trillionaires. He acknowledges he’s one of the people benefiting from this concentration, which gives the warning more weight.

His argument to fellow future trillionaires is both moral and pragmatic:

“If your answer is just ‘screw you, there’s nothing we can or should do about this,’ then that’s going to create a lot of discontent. It already has. We’re already starting to see the beginnings of it, and it’s just going to get worse.”

The pragmatic version: “You’re going to get a mob coming for you if you don’t do this in the right way.” If well-designed solutions don’t come first, poorly designed ones will follow. He specifically calls for “economically literate” tax policies, not the old ideological debates, but new frameworks for a world that’s never existed before.

Why Three Years Is an Eternity

When pushed on the risk of Congress doing nothing for three years (the most likely scenario under the current administration’s hands-off approach), Amodei makes the exponential case vivid:

• In 2023, models were “maybe as smart as a smart high school student”
• Now (early 2026), Anthropic engineers have models writing all their code, approaching “mid to high professional level”
• In another three years: “a country of geniuses in a data center,” maybe sooner

His framing of uncertainty is notably honest. On bioterror risks, he says: “We just don’t know.” But he argues for a paranoid engineering stance, the same as building a rocket. You don’t assume parts will survive tensile forces; you do scenario analysis for everything that can go wrong.

The Memo Behind the Message

Amodei reveals that “The Adolescence of Technology” was written in a 72-hour sprint at the end of winter break, after a week of “zoning out and playing video games.” Claude didn’t write any of it, but served as editor and research assistant.

A Few Observations

• The self-implication is strategic. When a future trillionaire tells Congress to tax future trillionaires, it short-circuits the usual “easy for you to say” response. Amodei explicitly says he’s one of the beneficiaries arguing against his own financial interest.
• The bipartisan claim deserves attention. Amodei insists this isn’t partisan, and that people on both extremes of the political spectrum say “remarkably similar things” in private. If true, the gap between private acknowledgment and public action is the real bottleneck.
• “Poorly targeted concern” is a useful frame. Public anxiety about AI water usage or power bills is real but misdirected. The actual stakes, according to Amodei, are wealth concentration, national security, and the pace of capability jumps that will reshape professional labor within years, not decades.
• The 14-minute format works. No time for platitudes. Every exchange cuts to a specific policy recommendation or a named risk. More tech leaders should be interviewed this way.