What Can't Fable Do?
A Hacker News user asks if anyone has seen the AI tool Fable genuinely fail at tasks, beyond its guardrails being triggered. They observe online discussions polarize between praise for Fable's capabilities and complaints about restrictions, with little reported neutral failure.
Background
- Fable is a coding AI agent (similar to tools like Cursor, Copilot, or Devin) that can autonomously write, debug, and fix code across a full project, not just suggest snippets. It was developed by Sentry, the well-known error-monitoring company.
- The "restrictions" and "guardrails" the post mentions refer to safety filters that block certain prompts (e.g., generating malicious code or copyrighted material), which some users see as overly aggressive.
- This thread is on Hacker News, a tech-community forum where developers discuss tools and share experiences. The post is asking for honest failure stories, since most public discussion is split between extreme praise ("Fable is magic") and complaints about censorship, leaving little middle ground about actual technical limitations.