Trending PR Campaigns
Claude’s Scorpion Antivenom Experiment
Citrini Research teamed up with Anthropic to see whether an unreleased version of Claude could help create scorpion antivenom. The experiment produced protein candidates that bound to some venom toxins, but none has been shown to neutralise venom or work as a medicine. The hook was a high-stakes, open-ended test of AI drug discovery, with a clear gap between the bold ambition and the modest scientific result.
Why it worked: It hooked journalists with a bold, time-boxed, dollar-tagged experiment that publicly stress-tested AI claims and documented the anticlimax with receipts.
The lesson: Consider staging a constrained, measurable challenge against a field’s hype, publish the full protocol, costs, and independent expert verdicts.
Coverage over time
How this story spread across the media
| Publication | Headline | Date |
|---|---|---|
| anthropic.com | How Claude designed protein binders against scorpion neurotoxins \ Anthropic | Tuesday 20th October, 2026 |
| benzinga.com | Anthropic’s Claude Targets Scorpion Venom - Benzinga | Tuesday 20th October, 2026 |
| interestingengineering.com | Claude's 64-hour experiment finds 17 binders for scorpion toxins | Tuesday 20th October, 2026 |
| aiweekly.co | Claude Autonomously Designs 17 Scorpion-Toxin Binders in 64 Hours | AI Weekly | Tuesday 20th October, 2026 |
| theverge.com | Claude did not invent scorpion antivenom | Thursday 22nd October, 2026 |