I’ve spent more than 15 years reading financial statements and making decisions with real money on the line, and one lesson keeps repeating itself: when a dominant player starts asking regulators to please raise the cost of doing business, it’s almost never because they care about the greater good. They’re worried someone smaller is about to eat their lunch. So when Dario Amodei — CEO of Anthropic, one of the two or three companies with the most to gain if the AI race ends in an oligopoly — published a 3,800-word essay calling to “pace” AI development, co-signed by Sam Altman, Demis Hassabis, and even Elon Musk, my first question wasn’t “are they right?” It was: “who benefits from me believing this?”
What’s Actually Real
I’m not going to pretend the fear comes from nowhere. Days before Amodei’s essay, Jacob Coxon — a researcher who passed through both OpenAI and Anthropic — publicly resigned, forfeiting equity that hadn’t even vested yet, saying neither company is acting responsibly and that both are racing straight toward self-improving superintelligence, “gambling with our lives.” What stands out is that Evan Hubinger, Anthropic’s own head of alignment science, came out to back the substance of that warning, putting the probability of AI-caused human extinction at above 10% within the next decade. That’s not an influencer chasing clicks — that’s someone inside the building. The technical risk of increasingly autonomous, less supervisable systems is real, and it would be irresponsible to dismiss it just because the messenger has skin in the game.
What Nobody’s Telling You
But here’s the part almost nobody connects in the press coverage: Amodei’s own policy proposal includes permanent on-site third-party audits for frontier models. And that kind of requirement doesn’t cost the same for everyone. For Anthropic, with billions in capital and an army of compliance staff, it’s another line item in the budget. For a three-person open-source lab in a garage — or in Bogotá, for that matter — it’s an insurmountable barrier to entry. David Sacks, the White House’s AI czar, called Amodei’s strategy exactly what it looks like: “regulatory capture” — a DMV for AI — and Chamath Palihapitiya went further: “Dario is making the case to shut down open source and concentrate enormous technological and economic power at Anthropic.” Meanwhile, more than 270 companies — Nvidia, Microsoft and Meta among them — signed a letter weeks earlier calling for exactly the opposite: more open models, not fewer.
And the market, which puts real money on the table, isn’t behaving like it believes in an imminent apocalypse. Today the Fear & Greed index sits at 50 — neutral — and on Reddit, where retail investors live, mentions of Nvidia jumped 104% in a single day and Palantir’s spiked 328%, sentiment mostly neutral. If there were real consensus that this could “kill us all” within the decade, as Coxon put it, we wouldn’t see that appetite. (To be clear: this is social-sentiment signal, not a verdict — but it cuts against the tone of the letter.)
The Play Behind the Fear — and What to Do About It
My read, filtered through the Método CRITERIO: technical risk and regulatory risk are not the same problem, even though Amodei presents them together. One calls for genuine caution in how models are designed. The other calls for skepticism about who’s writing the rules, and for whom. The real opportunity for anyone reading this isn’t picking a side in Silicon Valley’s ego fight — it’s understanding that the regulation being negotiated right now in Washington will decide whether, five years from now, the AI ecosystem looks like banking (three or four giants, sky-high regulatory barriers) or like the early internet (thousands of players competing). That matters for your portfolio, for your business if it’s built on open models, and for a country like Colombia, which today depends almost entirely on foreign technology and has far more to gain from an open ecosystem than from one captured by two or three names in San Francisco.
So the question I keep asking myself — and the one I’ll leave you with — isn’t whether AI is scary. It’s: who benefits from you being scared, and are you willing to let that decide for you?
This is exactly the kind of intersection between technology, power and money we break down every morning in the Daily Briefing. If this was useful, subscribe — it’s free and takes two minutes.