Models refuse CAPTCHA solvers but happily build semantic-segmentation click tools
cgarciae88 · x · 2026-09-06
An X post highlights a safety-guardrail paradox: models will refuse to implement a CAPTCHA solver outright, yet will happily implement a click tool based on semantic segmentation — functionally the same capability with different framing. The observation underscores how current safety refusals key on surface intent labels rather than actual capability or consequences.
More from Fun
- Mechanical koi unfolding into a floating garden, built with GPT-6 Astra + Blender — taherdhanera · 2026-09-06
- Ex-OpenAI VP of research mocks startup trend: the label 'lab' doesn't make it so — docmilanfar · 2026-09-06
- Claude Navier-Stokes proof rumor officially debunked by Elliot Glazer — basedjensen · 2026-09-06
- Dev predicts Chinese open-source clones of Astra within months — cephaloform · 2026-09-06
- Bored waiting on coding agents, dev builds Agentopoly — a shared Monopoly world via MCP — Bubbly-Importance-90 · 2026-09-06
- GPT6 Asked to Draw Michael Jackson in Google Calendar Fails Hilariously — gabrielchua · 2026-09-06