Dario Amodei's essay on how AI could transform biology, neuroscience, poverty, governance, and work - if we get the safety part right.
“The CEO of Anthropic giving his most optimistic take yet. Unusually specific about mechanisms - not just 'AI will be great' but how each domain could actually change. The biology section alone is worth the read.”
3 comments
The biology section is the most compelling part. If AI can compress 50 years of drug discovery into 5-10, the downstream effects on healthcare access are staggering. The question is whether regulatory systems can adapt that fast.
I appreciate that Dario acknowledges the 'if we get safety right' caveat isn't just a footnote - it's the whole ballgame. The optimism is conditional, and he's honest about the conditions being hard.
Interesting contrast with Eliezer's piece posted yesterday. Same technology, two very different priors about what happens by default. I think the truth is closer to Dario's view but Eliezer's concerns deserve serious engagement.