Project ideas from Hacker News discussions.

Grammarly is offering ‘expert’ AI reviews from famous dead and living writers

📝 Discussion Summary (Click to expand)

1. LLMs can imitate results but not the process that produced them

“The most interesting is the realization that if the LLM's input is only the output of a professional (human), then by definition the LLM cannot mimic the process the (human) professional applied to get from whatever input they had to produce the output.” – drbig
“An LLM can always output steps, but it doesn’t mean they are true; they are great at making up bullshit.” – latexr
“The LLM does not model text at this meta‑level. It can only use those texts as examples, it cannot apply what is written there to its generation process.” – tovej

2. “Learning” is a misnomer; LLMs are fitting parameters, not acquiring knowledge

“It all went south when we started to call it 'learning' instead of 'fitting parameters'.” – KeplerBoy
“Fitting is still too nice of a word choice… I suggest ‘randomly adjusting parameters while trying to make things better’.” – fxtentacle
“That isn’t learning, it can read things in its context, and generate materials… it doesn’t change the model weights.” – RichardLake

3. Legal, ethical, and cultural concerns over using deceased authors’ personas in AI tools

“The real issue seems more about transparency and consent around how the models are trained and how author personas are being used.” – BrtByte
“Digital necrophilia. The living ones are the ones that are going to have to make the objections here.” – jacquesm
“It feels wrong to spare nobody, not even dead writer/people.” – bayindirh

These three threads—process vs output, learning vs fitting, and the legal/ethical fallout—dominate the discussion.


🚀 Project Ideas

Process-First Writing Assistant

Summary

  • Captures a writer’s entire drafting workflow (drafts, notes, revisions) and presents it alongside AI‑generated suggestions.
  • Provides a transparent “reasoning chain” so users can see how the AI arrived at each suggestion, preserving the author’s voice and process.
  • Core value proposition: Process transparency + author control.

Details

Key Value
Target Audience Writers, students, content creators who value their own creative process
Core Feature Real‑time process logging + AI‑augmented suggestions with step‑by‑step explanations
Tech Stack Electron + React + Node.js, local SQLite, OpenAI API, WebAssembly for offline fallback
Difficulty Medium
Monetization Revenue‑ready: subscription + freemium tier

Notes

  • HN commenters lament that “LLMs only give output, not the process” (e.g., “I want to see the steps, not just output.”).
  • The tool addresses the frustration that AI “mimics” style but never reproduces the author’s internal workflow.
  • Sparks discussion on how to balance AI assistance with preserving human creative agency.

Author Persona Consent Platform

Summary

  • A web‑based registry where living authors or estates can register their persona, set usage terms, and license AI models to use their style.
  • Provides a transparent consent workflow, revenue sharing, and legal compliance for AI developers.
  • Core value proposition: Author empowerment + IP protection.

Details

Key Value
Target Audience Authors, publishers, AI developers, legal teams
Core Feature Persona registration, consent management, API for fine‑tuned models, revenue‑sharing dashboard
Tech Stack Django + PostgreSQL, Stripe for payments, OpenAI fine‑tuning API, JWT auth
Difficulty High
Monetization Revenue‑ready: licensing fees + revenue share on AI usage

Notes

  • Addresses HN concerns about “digital necrophilia” and unauthorized use of author personas (“We need to give living writers control.”).
  • Provides a legal framework that could mitigate copyright and publicity‑rights disputes.
  • Likely to generate debate on the feasibility of author‑controlled AI personas and the economics of licensing.

Local Grammar & Style Checker

Summary

  • An open‑source, offline grammar and style checker that can be integrated into IDEs and text editors, with optional LLM‑based suggestions that never leave the user’s machine.
  • Core value proposition: Privacy‑first, reliable editing without cloud dependence.

Details

Key Value
Target Audience Developers, writers, students, privacy‑conscious users
Core Feature LanguageTool‑based grammar engine, VSCode/JetBrains extension, optional local LLM inference
Tech Stack Rust for core engine, WebAssembly for browser integration, VSCode extension API
Difficulty Medium
Monetization Hobby

Notes

  • Responds to HN frustration that “Grammarly is too AI heavy” and “we need local tools.”
  • Enables users to keep their drafts on device, satisfying GDPR and data‑safety concerns.
  • Encourages community contributions and could become a standard in open‑source IDE tooling.

Read Later