Gary Marcus presses: is OpenAI's IMO 'basic harness' really just symbolic AI?
GaryMarcus · x · 2026-10-07
In the ongoing dispute over OpenAI's IMO-level math results, Gary Marcus presses his point: what exactly is a 'basic harness' if not symbolic? He argues that using a symbolic system for verification plus a neural network generating candidates is neurosymbolic; critics counter it's just an LLM engine with a sanity-check wrapper, not filtering-based proof search.
More from Models
- Grady Booch relays split expert takes on AI math: breakthrough or 'DDoS on mathematicians' — Grady_Booch · 2026-10-07
- OpenAI model reportedly solves long-standing GitHub problem in just 3 hours — birchlse · 2026-10-07
- WebDev Arena: claude-opus-5.5-max tops leaderboard at 1814 with 846K votes cast — arena · 2026-10-07
- This 20K-parameter, 81KB language model can still write coherent 300-word stories — x_Raincandy_x · 2026-10-07
- Mistral Large 4 Beats Opus 5.5 and GPT-6 Astra on Cybersecurity Benchmarks — Lower Refusal Rates — burny_tech · 2026-10-07
- Independent model Auro V9 ships, beating V8 head-to-head at a 10:1 clip — TheMoonMidas · 2026-10-07