Project ideas from Hacker News discussions.

No Skill. No Taste

📝 Discussion Summary (Click to expand)

Three prevailing themes in the discussion

# Theme Key points & representative quotes
1 Taste as a fragile differentiator in software “Personally, it feels like taste only buys you time and taste is easy to copy.” – cjonas.
“If you build an awesome app and want to charge for it, what stops me from just pointing “Claude Epic 2.5” at it and making a pixel‑perfect replica?” – cjonas.
“Taste is not easy to copy… if that were true then there would be no bad major Hollywood movies in established genres.” – mjr00.
2 AI lowers barriers but threatens quality and authenticity “my 7 year‑old is now able to nerd out and create games using Claude even though he's just barely learned to read.” – mlapeter.
“The problem is that outsiders without taste are showing up in a space where there is a long history of dues paid by the current occupants.” – mlapeter.
“I think the problem is that people are often delusional and AI feeds these delusions.” – mlapeter.
3 Taste is subjective, hard to define, and tied to skill/experience “Kind of meaningless if you let ‘taste’ be a vaguely‑defined term.” – altmanaltman.
“Taste is not synonymous with personal preferences… it refers to one’s power of discernment as to what is good.” – wafflebot.
“Taste is intersubjective… it implicates all other humans when it is made.” – throw4847285.

These three threads—how taste can be copied, how AI changes the creative landscape, and how taste is debated as a concept—dominate the conversation.


🚀 Project Ideas

CodeTaste

Summary

  • A browser‑extension + CLI that takes LLM‑generated code, runs it through a suite of style, test, and documentation generators, then scores the result on a “taste” metric derived from readability, maintainability, and user‑feedback patterns.
  • Core value: turns raw LLM output into production‑ready, well‑documented code that feels “good” to developers and end‑users.

Details

Key Value
Target Audience Indie devs, hobbyists, and teams using LLMs for coding
Core Feature Automatic refactor, docstring generation, unit‑test scaffolding, and a taste‑score dashboard
Tech Stack TypeScript, VS Code API, OpenAI/Claude API, ESLint, Jest, Docusaurus
Difficulty Medium
Monetization Revenue‑ready: $9 / month for premium features (advanced metrics, CI integration)

Notes

  • Users lament “LLM code is 100% generated but still feels off” (cjonas). CodeTaste gives them a tangible way to polish it.
  • The taste‑score can be shared on GitHub READMEs, giving projects a quick credibility badge.
  • Encourages a culture of “tasteful” code rather than raw output, addressing the “slop” problem highlighted by many commenters.

AppLens

Summary

  • A web service that aggregates app‑store data, open‑source metrics, and community reviews to compute a composite “taste score” for any app.
  • Core value: helps users filter out low‑quality, copy‑cat apps and discover genuinely well‑crafted software.

Details

Key Value
Target Audience HN users, product managers, early‑adopter developers
Core Feature Real‑time scorecard (UX, performance, documentation, community health) + side‑by‑side comparison
Tech Stack Go, PostgreSQL, ElasticSearch, React, GraphQL
Difficulty High
Monetization Revenue‑ready: freemium API + $49 / month for enterprise dashboards

Notes

  • Commenters complain about “slop” and “copy‑cat” apps flooding Show HN (verdm). AppLens gives a data‑driven counter‑measure.
  • The score can be embedded in README badges, giving projects instant social proof.
  • Sparks discussion on what “taste” really means when quantified.

VibeShare

Summary

  • A community‑driven marketplace for personal apps, where creators can publish, license, and receive feedback while protecting their code from easy cloning.
  • Core value: empowers hobbyists to share their work without fear of instant replication, and gives users a curated list of high‑taste projects.

Details

Key Value
Target Audience Indie developers, hobbyists, early‑adopter users
Core Feature Private/public publishing, code obfuscation, license templates, community voting, anti‑copy watermarking
Tech Stack Ruby on Rails, PostgreSQL, Docker, WebAssembly for sandboxed execution
Difficulty Medium
Monetization Hobby (open source core) + optional paid tiers for advanced analytics

Notes

  • Addresses the fear that “Claude can just clone your app” (cjonas). VibeShare’s watermarking and licensing deter casual copying.
  • Community voting aligns with the desire for “taste” validation (brador, lilanthran).
  • Provides a platform for the “slop” that still has value, letting users filter by quality.

LocalAI Shield

Summary

  • A developer‑friendly toolkit for building AI agents that run entirely locally, with built‑in data encryption, audit logs, and user‑controlled data sharing.
  • Core value: lets teams harness LLMs without exposing sensitive data or giving up sovereignty, addressing concerns about data integrity and privacy (bsoles, vunderba).

Details

Key Value
Target Audience Data‑centric teams, privacy‑focused developers
Core Feature Local inference engine, encrypted data store, fine‑grained consent UI, audit trail
Tech Stack Rust, WASM, SQLite, OpenAI API wrappers, Electron for desktop
Difficulty High
Monetization Revenue‑ready: $99 / year for enterprise support + open‑source core

Notes

  • Responds to worries that “AI can’t see data constraints” (bsoles) and the need for “sovereignty” (bsoles, vunderba).
  • Enables building tools like the “AI‑powered accessibility” example from devinprater while keeping data on the device.
  • Encourages a shift from cloud‑centric AI to user‑controlled, trustworthy solutions.

Read Later