IdeaWell

Project ideas from Hacker News discussions.

Amateur armed with ChatGPT solves an Erdős problem

Original Article

Hacker News Discussion

📝 Discussion Summary (Click to expand)

1. Tier‑gated access to advanced models > "Yes, but don’t we expect GPT 5.5 Pro will eventually be a free tier?" – jasonfarnon

2. Skepticism about “real” intelligence; LLMs are statistical generators

"They are obviously statistical token predictors… there are probably some very niche circumstances under which statements … are useful but by and large they are not." – baxtr

3. Cost‑effectiveness of spending on AI for math breakthroughs

"I think we should at least ask the latter, if it turned out it cost $100,000 to generate this solution, I would question the value of it." – Eufrat

4. Human verification is essential; AI output must be vetted

"That’s great if it works. But it’s way harder to produce a formal proof. So my expectation is that this will fail for most difficult problems, even when the non‑formal proof is correct." – dbdr

🚀 Project Ideas

Generating project ideas…

[ProofFormalizer]

Summary

  • Automatic conversion of AI‑generated mathematical proofs into verified Lean 4 code. - Eliminates the manual validation bottleneck that HN users repeatedly complained about (e.g., “Formalize it in Lean”).

Details

Key Value
Target Audience Mathematicians, graduate students, research engineers
Core Feature Parse natural‑language proof output from LLMs and generate complete Lean 4 scripts
Tech Stack Python backend, Lean 4 API, GPT‑4‑Turbo for parsing, FastAPI, React UI
Difficulty Medium
Monetization Revenue-ready: SaaS subscription with free tier ($15/mo)

Notes

  • Users lamented the difficulty of verifying AI proofs and the need for formal checking.
  • By automating Lean formalization, the tool turns a tedious step into a single click, increasing trust in AI‑generated results.

[PromptCraft Studio]

Summary

  • Curated library of high‑performing prompts for solving math competition and Erdős‑style problems.
  • Directly addresses the “prompt lag” and lack of effective templates that free‑tier users experience.

Details

Key Value
Target Audience Hobbyist researchers, educators, Pro‑users seeking reproducible success
Core Feature Searchable prompt database, A/B testing optimizer, auto‑generated meta‑prompts for different models
Tech Stack PostgreSQL, Node.js, ElasticSearch, Next.js front‑end
Difficulty Low
Monetization Revenue-ready: Freemium with premium packs ($9/mo)

Notes

  • Commenters repeatedly asked whether the free plan even gets access to “thinking models” and how to craft prompts that actually work.
  • Providing vetted prompts reduces token waste, speeds experimentation, and saves users money.

[TokenWatch Dashboard]

Summary

  • Real‑time monitoring of token consumption and cost across multiple AI APIs.
  • Answers the “how much did those tokens cost?” question that surfaced in the discussion.

Details

Key Value
Target Audience Power users, researchers, investors tracking AI spend
Core Feature API‑agnostic usage tracking, cost alerts, ROI analytics for R&D projects
Tech Stack Serverless (AWS Lambda), DynamoDB, Grafana, Stripe billing integration
Difficulty Medium
Monetization Revenue-ready: Subscription $20/mo per user

Notes

  • Users queried about the lag between Pro capabilities and free tier access and the cost of token‑heavy experiments.
  • Visualizing exact spend helps teams decide when to allocate resources or stop “burning tokens.”

[ErdosProblem Solver Playground]

Summary

  • Community platform aggregating unsolved Erdős‑style problems and automatically running them through multiple frontier LLMs.
  • Provides a systematic harness for evaluating AI performance across many problems at once.

Details

Key Value
Target Audience Math enthusiasts, academic collaborators, AI developers
Core Feature Bulk prompt execution, result aggregation, novelty scoring, export to markdown/Lean
Tech Stack Python, Celery workers, Docker, Angular UI
Difficulty High
Monetization Revenue-ready: Enterprise tier $50/mo for private problem sets, free community access

Notes

  • The thread referenced a repo of Erdős problems but lacked a centralized tool to test them systematically.
  • Building such a playground would let users identify effective prompts, democratize access to Pro‑level results, and avoid repeatedly paying for individual token burns.

Read Later