@tejasmanohar We used claude code as our agentic tool calling loop here. Great to prove out ideas before going further. We didn't really look at framworks as a counterpoint though, claude code was a good fit for our experiments. https://t.co/r4uYsNmp84
@hwchase17 Our team put out a lib to evaluate llm responses yesterday. You can use evals or create your own with it. Maybe we should write a blog post… 🤔
https://t.co/FXnZw1NxOQ
@Stretchedwiener I am in WA. Was down in ATX visiting an old friend not too long ago, hadn’t been there in 20ish years. Been here long enough to be incapable of handling that humidity any more.
@jeffwilcox@stevesi So your HOA is a bit less than the mortgage we had on this place when we got it 20 years ago. That might be coloring my attitude on this.
@jeffwilcox@stevesi HOA fees alone are enough to keep me from ever considering a condo in Seattle or anywhere else in the US. In Europe I would buy a condo, charges de copropriété seem reasonable.
🚀 v1.101 of @code is here! Check out what’s new:
- Supporting all MCP spec capabilities
- Chat tool sets
- Copilot coding agent integration
- Source Control Graph improvements
…and so much more: https://t.co/A9p63tiyY2
🧶 Here are some of the highlights…