Project ideas from Hacker News discussions.

Did OpenAI solve the wrong Navier-Stokes problem?

📝 Discussion Summary (Click to expand)

1. The forced (option C) solution is seen as a niche, less‑interesting result
- “But that specific option C isn’t what the larger math community cares about, it’s a pretty niche case” – dgellow
- “The forced version is easier since you can custom design the force function … getting the blow‑up might be regarded just as much a function of your bespoke force function as of the fluid dynamics itself” – HarHarVeryFunny
- “A solution … made by an LLM … would still not be that interesting to the math community” – etdznots

2. Debate over whether OpenAI’s work is a genuine solution or a sensationalized “loophole”
- “OpenAI did not solve the ‘wrong’ Navier‑Stokes problem … but did give a solution to the Clay Millennium Prize Problem as written and understood” – jonlong
- “SciAm writes … ‘the LLM found and exploited a loophole …’. This is pure sensationalism. Choosing option (C) … is neither a ‘loophole’ nor something ‘found by the LLM’” – jonlong
- “OpenAI did legitimately meet the Millennium Prize solution criteria, but in the most unrealistic, and therefore least interesting, way possible” – HarHarVeryFunny

3. Discussion of the problem’s formulation and who bears responsibility for option C
- “If the larger math community doesn’t care about option C, Fefferman shouldn’t have given that option in the first place” – hatthew
- “Fefferman and the larger math community are not the same actor; a divergence between their concerns is not surprising” – dragonwriter
- “But Clay chose Fefferman to represent ‘the larger math community’? … the larger math community failed to set the bounds of the problem correctly” – ChickeNES


🚀 Project Ideas

Generating project ideas…

LoopholeLint: Formal Problem Statement Analyzer

Summary

  • Detects trivializing constructions (e.g., arbitrarily smooth external forces) that could let an AI “solve” a problem via a loophole.
  • Outputs a loophole‑risk score and suggests tighter constraints to preserve problem difficulty.

Details

Key Value
Target Audience Prize problem committees, mathematicians, AI researchers
Core Feature Symbolic analysis of problem statements + LLM‑guided counterexample search to find easy exploits
Tech Stack Python, SymPy, Lean theorem prover backend, LLM API (OpenAI/Anthropic), React frontend
Difficulty Medium
Monetization Revenue-ready: SaaS subscription $20/mo per team

Notes

  • HN commenters complained that OpenAI “exploited a loophole in the framing of the question” (HarHarVeryFunny).
  • Provides a concrete way to vet future Clay problems or other open challenges, restoring community trust in problem statements.

ProofInterest: AI‑Generated Proof Impact Dashboard

Summary

  • Scores AI‑produced proofs on novelty, reliance on external constructions, and community relevance.
  • Combines automated metrics with expert feedback to highlight genuinely interesting contributions.

Details

Key Value
Target Audience Mathematicians reviewing AI work, grant agencies, journals
Core Feature Metric engine (proof complexity, use of user‑defined objects, Kolmogorov‑style length) + UI for uploading proofs and community rating
Tech Stack Node.js/Express, PostgreSQL, Python scikit‑learn for scoring, LLM for plain‑language summaries, Vue.js frontend
Difficulty High
Monetization Revenue-ready: $5 per proof evaluation or institutional annual license

Notes

  • Users noted the OpenAI result “is not interesting to the math community” and called it a “niche case” (dgellow, ChickeNES).
  • The dashboard would spark discussion on what makes a proof valuable and help steer AI research toward meaningful advances.

ProblemForge: Community Open Problem Refinement Hub

Summary

  • Wiki‑style repository where anyone can propose, edit, and vote on variants of open problems, with version history.
  • Clearly marks which formulations are considered core community interests versus niche alternatives.

Details

Key Value
Target Audience Open problem enthusiasts, research groups, prize foundations
Core Feature Collaborative editing (Git‑backed), voting, automatic diff highlighting, consistency checks via theorem provers
Tech Stack Git‑based backend (Gitea), Markdown problem files, React UI, CI running Lean/Coq checks, optional LLM to suggest clarifications
Difficulty Medium
Monetization Hobby (open source, donation‑supported)

Notes

  • HN discussants argued “the larger math community failed to set the bounds of the problem correctly” and that Fefferman shouldn’t have offered the niche option (ChickeNES, dragonwriter).
  • Enables the community to pre‑emptively refine problem statements, reducing frustration over misaligned prize challenges.

Read Later