Google Search AI leaks its chain-of-thought and internal formatting tags to users
ConsistentMap5775 · reddit · 2026-11-21
A Reddit user's screenshot shows Google Search's AI answer exposing its internal reasoning — self-talk about ambiguity, deciding whether to search for memes, formatting constraints and literal <FollowUp>/<Carousel> tags — before appending the user-facing response. A classic pipeline leak where the model's internal scratchpad got concatenated into the public answer.
More from Fun
- Claimed Claude-assisted proof of Riemann Hypothesis slammed as reading 'like a suicide note' — zetalyrae · 2026-11-21
- 'Math is solved, what's next?' — AI circle's punchline: Physics — JensHonack · 2026-11-21
- AI art community rallies to bring back prompt sharing and creative challenges — azed_ai · 2026-11-21
- GTA6 dev on rebuilding the weather system: "my code is staring right back" in your weather app — flngr · 2026-11-21
- OpenAI researcher re-ups 5-year-old bet: AI to crack 6 of 7 math problems by 2040 — littmath · 2026-11-21
- Mystery solved: Astra, Fable and Neon are Jason Dean Lee's PhD students, not models — jasondeanlee · 2026-11-21