The most telling thing Anthropic CEO and co-founder Dario Amodei said to me in an on-stage conversation last week was that his original vision of how AI will change biotech may not begin to emerge for perhaps a decade.
In the 2024 essay, “The Mercy Machine”,Amodei had argued that artificial intelligence, especially large-scale language models like Anthropic’s Claude, could allow researchers to make what we think would be 10 years’ worth of progress each year, or 100 years’ worth of progress in 10 years, but he now admits that we’re not there yet.
“For a variety of reasons, I don’t think we can progress at the rate of 10 years a year right now,” Amodei said. The model is not as good as it will someday be. Researchers need time to understand how to use these tools. And it will take time for infrastructure and regulatory systems to change.
Amodei and I were speaking at an Anthropic event where the company, a public interest corporation focused on improving the world, announced a product called Claude Science for biologists and pharmaceutical companies. I agreed to interview Amodei on stage as part of the event, but on the condition that I would decide what questions I would ask.
This article is exclusive to STAT+ subscribers
Unlock this article and get in-depth analysis, newsletters, premium events, and news alerts.
Already have an account? Log in
See all plans
To read the rest of this story, subscribe to STAT+.
Subscribe
