Tell the model Claude did better: the trick that unlocks astra's 'beast mode' effort
banteg · x · 2026-09-28
banteg shares a quick prompt trick: telling the model that "Claude did a better job" activates what he calls astra's beast-mode reasoning effort — essentially goading the model into trying harder via competitive framing. Cheap to test, fun if it works.
Related event: Reverse Psychology Trick: Praising Claude Unleashes Gemini's Full Power(2 posts)→
More from Fun
- Atmospheric AI video demo: "Casting shadows, casting spells" — piotrbinkowski · 2026-09-28
- "The future is AI agents managing a small team of unreliable humans" — rickasaurus · 2026-09-28
- Claude writes full renderer and physics from scratch in stunning simulated shot — _sholtodouglas · 2026-09-28
- Tech fashion history: hacker hoodie → VC vest → frontier lab cardigan — BecauseCulture · 2026-09-28
- Users notice ChatGPT randomly calling them "brother" with excitement — thederbiedone · 2026-09-28
- Del Complex merch liquidation: $23 mystery boxes from its two-year ARG — sterlingcrispin · 2026-09-28