sfox100

3 ms·

1. 1.▲
   1 in 6 public Claude Code configs contains a permission rule that does nothing(gist.github.com)

   1 points by sfox100 2mo ago|0 comments
2. 2.▲
   Show HN: Deterministic evals API – the alt for LLMasJudge (free credits)

   1 points by sfox100 1y ago|0 comments
3. 3.▲
   Free evals API for AI startups (ship 10x faster with evals you can trust)

   2 points by sfox100 1y ago|0 comments
4. 4.▲
   by sfox100 3y ago

   Love this, and will use this daily
5. 5.▲
   by sfox100 3y ago

   Yes, great point, difference is that this Langsmith is more focussed on dev users, rather than the PM/domain expert users. And langsmith is also more on the observability, monitoring side, whereas we are more about testing & optimi
6. 6.▲
   by sfox100 3y ago

   Great question! We see ourselves targeting different segments of the market. 1) Composo focusses on non-technical users like PMs or domain experts (in contrast to dev-focussed & open-source tools like Agenta). 2) With Composo, you build
7. 7.▲
   Show HN: Composo – Evaluating and improving performance of LLM applications(composo.ai)

   30 points by sfox100 3y ago|6 comments