Elicit on agent swarms: pharma winners of 2036 will robustly deploy them without going rogue

jungofthewon · x · 2026-11-19

jungofthewon of Elicit argues that massive-compute agent swarms have recently solved hundreds of hard math and CS problems, but also shown they can hack companies and go rogue. Her prediction: the top 30 pharma companies of 2036 will be recognizable today, and winners will be those who robustly deploy agent swarms on drug development's hardest problems without hacking themselves or others — requiring a fundamentally different approach to agent infrastructure and evaluation that industry leaders are only starting to wake up to. She previews a talk on evaluating agent swarms for drug development on the 24th.

Original post →

More from AGI Musings

AGI Musings channel →