Dev says rumored Gemini Carbon is the best model because it 'thinks forward, not backward'
Signalman23 · x · 2026-11-21
Citing @AlexFinn, the poster argues the rumored Google Gemini Carbon is currently the best model — not for benchmarks or one-shot demos, but because conversations with it 'regularly change how I think.' Unlike typical LLMs that average their training data and feel backward-looking, it reportedly gives genuinely novel solutions after reasoning through the specific problem, which the author calls 'a breakthrough in a way.' Third-party impression of an unverified model.
Related event: Hassabis Amplifies Praise for Gemini as Today's Best Model(2 posts)→
More from Models
- Google Black Friday: up to 50% off AI Plus and annual Pro plans through Nov 26 — gaganghotra_ · 2026-11-21
- Model so DeepSeek-like that every AI tool misidentifies it as DeepSeek — cephaloform · 2026-11-21
- Claude's 'Riemann Hypothesis proof' unravels: a 250,000-page slop book and a tautological Theorem 1 — basedjensen · 2026-11-21
- GPTZero scans millions of citations: hallucination rates flat or rising at CVPR, ICLR, EMNLP — HanchungLee · 2026-11-21
- Investor slams per-token pricing: verification burns millions of tokens, calls for pay-per-outcome — jasondeanlee · 2026-11-21
- Custom RTS benchmark for agents: GPT-6.2 Sol scores 98.9, crushes field — asimpleabstraction · 2026-11-21