Dario Amodei is not among the most active executives on X, but his two-part reply on August 16 to investor Gavin Baker, CIO of Atreides Management, had drawn more than 14 million views as of this writing and even prompted a reaction from Elon Musk on the thread. Amodei co-founded Anthropic in 2021 with a group of former OpenAI researchers, his sister Daniela Amodei among them, and built the company — now the maker of the Claude models — with Amazon and Google among its major backers. The second part of the reply restated the prediction he is best known for: that AI will make it possible to cure most human diseases in roughly five to ten years. He acknowledged how that lands, saying the thought likely sounds crazy to ordinary people and to biologists as well.

2/2 Second, on the messaging around AI. I do not agree that my messaging has been disproportionately negative. In fact it has been about equally balanced between risks and benefits: I’ve written one major essay about each, and even in interviews where I discuss the risks, I… — DarioAmodei (@DarioAmodei), August 15, 2026

Why Amodei Was Replying at All

Baker had devoted part of an An-In podcast segment to Anthropic, saying he had been told that Amodei privately raised the possibility that his company could end up as the only private company left in the world, with governments running everything else. He followed up on X with a broader complaint, arguing that Amodei has lost the regulation argument and should act as a more positive advocate for the industry he leads.

Amodei disagreed with the framing. His two major essays, he argued, split evenly between risk and benefit, and clips pulled from interviews gain far more traction on social media when they sound bleak. Machines of Loving Grace, the 15,000-word essay he published in October 2024, was the benefit half. Most of that essay argues that AI has the capability to compress 50 to 100 years of biological progress into 5 to 10. The incredulity he acknowledged has factual anchors: a new drug typically takes a decade or more and billions of dollars to move from discovery to approval, and roughly nine in ten candidates that enter clinical trials fail. Against that baseline, the forecast amounts to a claim that the base rates themselves are what AI will change — the core of the essay’s argument.

The Admission Attached to the Forecast

The more consequential part of the post was not the prediction but what Amodei conceded around it. He called the sharpest criticism of AI companies, his own included, the fact that they have not delivered on what they promised the world, and he placed that responsibility squarely on the industry rather than on its critics. He also rejected the idea of fixing public sentiment with a marketing push, saying that promising to cure cancer now reads as cliché and mostly comes across as deceptive. Actually curing it, he argued, is the only thing that works.

He then attached a near-term timeline to the forecast. Anthropic is accelerating its work in biology and medicine, and Amodei said he expects major results in the coming years, with “some early glimmers in the coming months.”

Anthropic Has Been Building Toward the Thesis for Months

The claim is not floating free of substance, and Anthropic is not alone in betting on biology as the test of AI’s grandest promises. Google DeepMind’s AlphaFold, which predicts protein structures from amino-acid sequences, earned its creators a share of the 2024 Nobel Prize in Chemistry, and DeepMind’s 2021 spinout Isomorphic Labs has since signed AI drug-discovery partnerships with major pharmaceutical companies. Anthropic launched Claude for Life Sciences in October last year, announced research partnerships with the Allen Institute and the Howard Hughes Medical Institute on February 2, and reportedly acquired stealth biotech startup Coefficient Bio in a $400 million stock deal. Coefficient’s two founders came out of computational drug discovery at Genentech’s Prescient Design, and the roughly ten-person team is joining Anthropic’s health and life sciences group.

A Decade-Long Call Now Has a Near-Term Checkpoint

Ten-year predictions are cheap because nobody collects on them, and Amodei has now attached a shorter one to his. “Glimmers in months” is a claim that can be checked, and quickly. It also turns a diffuse industry debate — whether AI can drive genuine scientific discovery rather than simply assist it — into a dated, public test. He said that when Anthropic accomplishes something real, the world will hear about it as loudly as possible, and that until then he does not want to make empty promises. That is a firmer commitment than most AI executives put in writing, and the timing is not incidental. Anthropic filed confidentially with the SEC in June, and its backers are already modelling a $2 trillion valuation for an October listing. The early glimmers would land right around the same window.