Dario Amodei made a big claim at a convenient time.
In a 42-minute IPO roadshow video released on November 10, Anthropic’s CEO predicted that “cancer, Alzheimer’s, heart disease, mental illness — will be relics of the past in less than a decade.”
Scientists have been trying to beat these diseases for centuries, yet Amodei is giving AI less than 10 years. Naturally, pharmaceutical executives, physicians, and researchers are pushing back, citing complicated biology and long clinical trials.
The Pitch
Amodei thinks AI can greatly speed up biological research, drug discovery, and medical breakthroughs.
“I think if we do this right, we can create an age of miracles where humanity is able to solve problems that have eluded us for millennia,” he said in the video.
Other tech leaders have said similar things. According to NDTV, Google DeepMind’s Demis Hassabis has suggested AI could compress a decade of drug discovery into weeks. Elon Musk has claimed AI will give everyone the same quality of medical care the U.S. president gets.
Anthropic has put money behind the idea. It has bought several biotech startups, including a one-person company working on muscular dystrophy treatments.
Per Business Insider, a person familiar with the matter said it also opened a “wet lab” in the Bay Area.
In an October blog post, the company said it had “begun drug discovery efforts of our own.” The push also helps soften Amodei’s reputation as an AI doomer, and it gives investors something to get excited about.
The Reality Check
The drug industry is polite about the idea but doesn’t think it will happen.
“I think that would be amazing and I think it’s not likely,” Eli Lilly CEO Dave Ricks wrote on X. “The map of biology is still largely a blank page and randomized human proof is the long pole in the tent,” he added.
Jeff Leiden, executive chairman of Vertex Pharmaceuticals, was far blunter. He called these timelines “complete BS.” He noted that Vertex tested roughly 50 AI companies on about 10,000 compounds, and not one of them produced a working drug on its own.
The problem is the slow middle of the process. AI may help find promising molecules faster, but those molecules still have to go through animal testing, regulatory review, and years of human trials.
One X user pointed out that the mean time from first human test to FDA approval was 14 years for the 20 most recently approved drugs.
Data is another obstacle. “The data exists for <1% of biology,” computational chemist Dr.Rich Law replied on X.
The Cliché
NYU professor emeritus Gary Marcus said the video contained overpromises “that would have made Elizabeth Holmes blush.” Meanwhile, biotech executive Rami Elghandour noted that “these are insane things to say in an IPO pitch.”
Amodei himself said something similar not long ago. In an August X post, he wrote that “saying that AI will cure cancer is more a cliché than it is inspiring, and most people think it is deceptive.”
This time it didn’t sound like a cliché, but when the company starts trading on Thursday, November 19, the market will get to decide whether it believes him.
Image: Shutterstock
Recent Comments