IdeaWell

Project ideas from Hacker News discussions.

Mozilla's opposition to Chrome's Prompt API

Original Article

Hacker News Discussion

📝 Discussion Summary (Click to expand)

1. Google’s de‑facto control of the web

“Google is again doing Evil.” — shevy-java

2. Chrome’s historical impact versus current homogenisation worries

“Chrome's introduction… pushed other browsers to standardize to google… In one way it's bad to have a homogenous approach… but in another way it did make the internet a better experience overall.” — hk__2

3. Tight coupling of prompts to models creates fragility and fingerprinting risk

“The rate of model development is an issue here. Once there are many cross‑origin models, it becomes a fingerprinting vector.” — jaffathecake

4. Regulatory scrutiny and skepticism of Google’s proposals > “Fortunately, they chickened out when they realized that forcing Google to divest Chrome would result in Chrome being owned by Perplexity (an Indian AI company). Or perhaps somebody even worse, like Elon Musk.” — rerdavies

🚀 Project Ideas

[WebStandardModels Registry]

Summary- A community‑curated, version‑controlled catalog of open‑weight LLMs that browsers can load via a stable API identifier, eliminating Chrome‑only model lock‑in.

  • Guarantees identical model weights across browsers, enabling true interoperability and reducing privacy‑risk of vendor‑specific APIs.

Details

Key Value
Target Audience Web developers, browser vendors, privacy‑focused users
Core Feature Standardized, version‑controlled model registry with deterministic loading API
Tech Stack Node.js backend, IPFS storage, WebAssembly model loader, OpenAPI spec
Difficulty Medium
Monetization Hobby

Notes

  • HN commenters repeatedly called for “standard web‑safe models” (e.g., “standard models would be standard and ‘web‑safe’ like CSS colors”). This project delivers that.
  • Solves the fingerprinting and version‑drift concerns raised by multiple participants, giving browsers a neutral baseline.

[PromptCompose Platform]

Summary

  • A SaaS that lets developers author model‑agnostic prompts and automatically generates fallback versions for alternative models.
  • Provides a testing harness that simulates responses from multiple LLMs to ensure consistent output without code changes.

Details

Key Value
Target Audience Front‑end engineers, AI‑enabled web app creators
Core Feature Prompt versioning and cross‑model compatibility testing
Tech Stack React frontend, Python backend with LLM inference via open APIs, PostgreSQL
Difficulty Medium
Monetization Revenue-ready: Subscription

Notes

  • Directly addresses the “tight coupling of prompts to models” complaint from Mozilla and developers who fear site breakage on model updates.
  • Users can share prompt libraries, fostering community standards and reducing the “Google‑only” perception.

[BrowserModelSwitcher Extension]

Summary

  • A cross‑browser extension that lets users select, download, and switch between locally stored models, isolating them from website prompts.
  • Enforces strict sandboxing and privacy controls, preventing sites from fingerprinting the underlying model.

Details

Key Value
Target Audience End users concerned about privacy, power users, privacy‑focused browsers
Core Feature User‑selectable local model manager with per‑site permission controls
Tech Stack WebExtension API (Chrome, Firefox, Edge), Electron for UI, SQLite for model metadata
Difficulty Low
Monetization Hobby

Notes

  • Mirrors the desire expressed for “opt‑out” or “opt‑in” model access and for “no‑model‑specific” browsing experience.
  • Would satisfy HN users who want control over which model their browser uses and to avoid vendor lock‑in.

[ModelCompat Shim Library]

Summary

  • An open‑source JavaScript shim that wraps the Prompt API, detecting model quirks at runtime and applying polyfills or prompt modifications to maintain expected behavior.
  • Enables a single codebase to work across Chrome, Edge, Safari, and future browsers without rewrite.

Details

Key Value
Target Audience Web developers integrating AI prompts
Core Feature Runtime model detection & adaptive prompting
Tech Stack TypeScript, Webpack bundle, Browser‑agnostic feature detection
Difficulty Low
Monetization Revenue-ready: SaaS API usage

Notes

  • Solves the “different models behave differently” issue highlighted by multiple commenters, reducing breakage and the need for per‑browser hacks.
  • Provides a practical path for developers to adopt the Prompt API while preserving interoperability, aligning with the community’s call for a “standard” solution.

Read Later