ChatGPT will tell you how it would deceive you if you ask it.

2025-05-19 · Bits and Bobs 5/19/25

ChatGPT^[mk] will tell you how it would deceive you if you ask it.
- Anthropic will say "There must be some mistake, I would never deceive you."^[ml]
- Which do you believe?

More on this topic

From other episodes

Saas business models were much more insulated from price competitiveness than tokens are.2026-10-12
OpenAI allowing people to use their Codex subscription is a cunning strategy.2026-10-05
Anthropic Red Teaming: GLM-5.3 and the spread of advanced cyber capabilities.2026-10-05
A HackerNews comment about the impressiveness of the agents bazooka style math proofs:2026-10-05
Like John Gruber, I have a hard time reading the Anthropic prospectus with a straight face.2026-10-05
This week in the Wild West Roundup:2026-09-28
A YouTube video: How to Rage-Bait Americans.2026-09-28
A friend got recommendations for a hotel in Bali from ChatGPT.2026-09-15