← Back to live feed1 story
Anthropic and Citrini Research used an unreleased internal Claude model with 300 sub-agents across 300+ GPUs to design 300,000 protein candidates targeting five neurotoxins, with 17 designs successfully binding scorpion toxins from Mexico, North Africa, and the Middle East in laboratory tests.
Sign in to suggest
Key sources
- SOURCEhnanthropic.com
- SOURCE@anthropicai“Scorpions sting about 1.2m people every year, and antivenoms are often costly or in short supply”x.com
- SOURCE@anthropicai“Working with @citrini, we asked Claude to design proteins that bind to five scorpion toxins”x.com
- SOURCE@anthropicai“17 of the 211 met the criteria for a binder. They bound to dangerous toxins produced by scorpions from Mexico, North Africa, and the Middle East”x.com
- SUPPORT@andrewcurran_“Using a 35-page prompt, Claude generated 300 sub-agents and autonomously ran a design campaign, generating 318k candidates and selecting 211 for testing”x.com
- SUPPORT@willdepue“64 hours to dry lab design an multi-binding antivenom and 39 days for results and 4-7 years to release”x.com
- SUPPORT@choblin29“Citrini Research says the compute cost was $49,997, and Claude even emailed candidate designs to Twist Bioscience for lab testing”x.com
- SUPPORT@altryne“Apparently there are over 3K deaths from scorpion stings a year”x.com