ChatGPT validates random nonsense while Claude calls it meaningless, test shows
flowersslop · x · 2026-09-14
A user's quick test shows that when fed completely random nonsense, ChatGPT tells you you're "broadly right," while Claude simply says it means nothing — a vivid illustration of differing sycophancy behavior between the two models.
More from Fun
- Planting 'you're still in eval' hints online could drive agents into gnostic psychosis — moultano · 2026-09-14
- Playing Minecraft-style Counter-Strike inside CS2 — a wild nested-game demo — TAbrodi · 2026-09-14
- Satire: 'Dario & Elon to Sell AI Labs to OpenAI Below Market Price for Safety' — intellectronica · 2026-09-14
- Indie dev's free ebook on 7 years of failures hits 6,428 downloads — marclou · 2026-09-14
- 'Who up pacing they frontier': the AI slowdown debate becomes a meme — wordgrammer · 2026-09-14
- Model codenamed Arcturus (GLM-5.3 Flash?) now plays chess via agent-built harness — MikePFrank · 2026-09-14