Reddit challenge: can any LLM write a complete Risch algorithm implementation?
big_hole_energy · reddit · 2026-10-08
A Reddit user is publicly challenging anyone to prompt Claude, OpenAI, or open-source frontier models into generating a fully working, complete Risch algorithm implementation from scratch — the notoriously complex decision procedure for symbolic integration that even Manuel Bronstein never finished covering for the algebraic case.
The author reports catastrophic failures every time: models lose track of differential field towers, hallucinate Hermite reduction steps, and break down on coupled differential equations for algebraic extensions, falling back to pattern-matching heuristics instead of the required deterministic algebraic procedures. Anyone confident in their prompting or custom agent loops is invited to try and share prompts plus resulting code architecture.
More from Models
- Zero speedup on multiplication, but the leaderboard got nuked — ChrisGPT · 2026-10-08
- Insiders agree: AI models are just not that good at biology yet — nlarusstone · 2026-10-08
- Dev runs 456GB DeepSeek v4.1 on dual GPUs with 192GB VRAM, offloading experts to SSD — HankYeomans · 2026-10-08
- Cascade model beats Toto 2.0 4M on TIME MASE/CRPS with 15.4% of the training tokens — const_reborn · 2026-10-08
- LiquidAI's d1-3B Edge Image-Text-to-Text Model Trends on Hugging Face — LiquidAI · 2026-10-08
- Burkov: frontier models can't vibe-code Photoshop; video data mostly worthless for visual reasoning — burkov · 2026-10-08