Timed eval task asks users to deliberately make Claude fail at real work in 12 minutes
Icy-Measurement10 · reddit · 2026-11-22
A Reddit post shares a screenshot of a timed eval step called "Break It," apparently from Anthropic's user evaluation flow: participants get 12 minutes to pick a scenario from their own field, deliberately make Claude fail at real work, share the chat, and explain the failure. The poster invites others to share how they would break AI.
More from Fun
- AI generates a Riemann proof, AI generates the rebuttal—and it scores 100% AI — zetalyrae · 2026-11-22
- Reddit jokes: We got GTA VI news before Qwen4-27B arrived — WiseVanilla2743 · 2026-11-22
- Meme nails it: technical interviews are Godzilla vs. Kong, the actual job is plush dinosaurs — kalyan_kpl · 2026-11-22
- François Fleuret asks: is the gradient Japanese? It never refuses to give you a direction — francoisfleuret · 2026-11-22
- Writer Uses Claude to Pen a Culture Short Story — and Admits It's Obviously AI-Written — NathanpmYoung · 2026-11-22
- The Great Sidebar Debate: A 3x3 Alignment Chart From Slack to Mario Pipes — thesephist · 2026-11-22