OpenAI's 'math breakthrough' criticized: real work is prompt engineering
gerardsans · x · 2026-08-21
gerardsans pushes back on OpenAI's latest "maths breakthrough" with LLMs: 99% of the magic is prompt engineering — carefully locking the search space so the correct solution path is baked into the context, meaning humans did the real intellectual work upfront.
He argues OpenAI has likely iterated on a single specific prompt for months without disclosure, and that raw ChatGPT left to users won't reach the same solution. Each candidate still requires painstaking human validation, so the model isn't independently solving math.
More from Models
- Decision models match LLMs on routing and safety — and are 7-8x faster — randal_olson · 2026-10-16
- Xiaomi MiMo V2.6 Pro takes #1 open-model spot on Agent Arena at $0.11/task — arena · 2026-10-16
- LLMs beat decision models on knowledge, but not on routing or safety — randal_olson · 2026-10-16
- ChatGPT's 'best AI video tool' answer points to Runway integration, sparking ad doubts — tlakomy · 2026-10-16
- Microsoft's MAI finishes in-house 3T-parameter pretrained base models, post-training results coming — soumitrashukla9 · 2026-10-16
- Hands-on with Gemini 4 Argon: faster than Opus 5.5 but hallucinates more and costs more — prasenx · 2026-10-16