Hacker News

New stories

Live mirror
30 storiesupdated just nowView source snapshot
  1. No errors, no warnings, no gods, no masters – HTML Purity is a Fetish (shkspr.mobi)
    —discuss
  2. Oracle Extends Fusion Agentic Applications with Introduction of Fusion Claw (oracle.com)
    —discuss
  3. Apple's New CEO Moves to Overhaul Company to Run Faster and Leaner (bloomberg.com)
    —discuss
  4. Show HN: ProgressCove, calm and smart to-do app with Home Assistant integration (progresscove.com)
    2comments
  5. San Diego overbills sanitation customers due to file copy error (voiceofsandiego.org)
    —discuss
  6. Another World Ported to ZX Spectrum (github.com/antirez)
    —discuss
  7. A list of publications that accept cartoons and their associated payment (colemantoons.com)
    —discuss
  8. Turso 0.8: Concurrent writes without SQLite's single-writer bottleneck (turso.tech)
    —discuss
  9. The Genius Trick Behind Exile's Impossible Map (youtube.com)
    —discuss
  10. Every model (incl. Jev) we tested inflates security finding severity (casco.com)
    1comments
  11. Wikifunctions (wikifunctions.org)
    —discuss
  12. Slow running benefits: Boosts in mood and brain function at very light intensity (direct.mit.edu)
    —discuss
  13. Singularity – design a data model, get the REST API and the MCP server (github.com/dantesabatier)
    —discuss
  14. Show HN: Squidbrake – self-hosted approval gateway for AI agent tool calls (github.com/batrapulkit)
    —discuss
  15. Enforce positive security with Cloudflare Application Profiles (cloudflare.com)
    —discuss
  16. Bootstrapping an Infrastructure [pdf] (usenix.org)
    —discuss
  17. Author Dropped from Literary Prize over AI Allegations (plagiarismtoday.com)
    —discuss
  18. Nvidia Donates $12,000 to the Perl and Raku Foundation (perl.com)
    —discuss
  19. TLA+ helped us fix and10 issues in our OSS project (github.com/desplega-ai)
    1comments
  20. People building AI think it might kill everyone. Hear from them directly. (frominside.ai)
    —discuss
  21. Ciao, Control coding agents from your phone (apps.apple.com)
    1comments
  22. SlopOne: Winning the code quality fight with Jev (qlty.sh)
    —discuss
  23. Have we reached peak Markdown? (strata.space)
    —discuss
  24. Does software performance still matter? (lemire.me)
    1comments
  25. AI: Igniting the Spark to End Stagnation (lemire.me)
    —discuss
  26. Antifragile Programming and Why AI Won't Steal Your Job (lemire.me)
    —discuss
  27. Fallout: New York (fallout.nyc)
    —discuss
  28. Weapons of Mass Decentralization (worksinprogress.co)
    —discuss
  29. LeRebot So-101 Crash Course (medium.com/rob.bercik)
    —discuss
  30. Show HN: Postgres MCP server with a context file the agent can add to (github.com/contextflo)
    1comments

TLA+ helped us fix and10 issues in our OSS project

1 pointsby 13m agogithub.com
1 comments
13m agoHN ↗

Hi there!

A few weeks ago, Boris posted how formal verification helped fix issues in their codebase (https://x.com/bcherny/status/2102543349102338309) and I definitely wanted to try it out in the swarm codebase.

The idea was simple:

1. Go over the codebase and fine most critical parts that could be modelled 2. Pick the first few and try to model them based on main code 3. Find counterexamples 4. Generate failing tests for them (single tests file diff pr) 5. Go over each of them and fix them + add the counterexamples in the repo for reference

I think this pattern is super nice tbh, and it helps a lot on reasoning around the code and implementation, specially when AI was involved.

The main issue: we had to add a daily wf to check for any updates to the spec, so that we keep it in sync (price you pay when the language can not be expressive enough, e.g. js).

Another win was that for the heartbeat system, which I wanted to refactor for a while, it managed to propose a simplification from 1.3M states to only ~4k (lol).

I believe it can be used for complex systems refactoring, while keeping the same functionalities.

Blog post about it (beware AI generated, but it contains some of the prompts we used during the session, just point your agent to the link and let it cook) -> https://www.agent-swarm.dev/blog/tla-plus-races-agent-swarm

I've been wanting to use tla+ for years, but always struggled w the syntax when working on something non-trivial. AI generating the specs and explaining the reasoning behind them helped a lot!

If you curious about the specs -> https://github.com/desplega-ai/agent-swarm/blob/main/specs/t...

Cheers,