Please reach out if you have an open role or any connections!

Some projects I've lead:
- Complete normalization of all tables and fields in a MySQL database
- Migrated a production database from MySQL to PostgreSQL (leading to an ETL pipeline that ran 50% faster)
- Designed database schemas for multiple data collection systems supporting federal reporting

For some reason, I find myself switching between Zed's built in agent tooling and opencode.ai in the terminal.

The big reason I like Zed's AI tooling is being able to individually Accept and Reject code blocks

True 😂

What are we actually expecting from this open model at this point? A SOTA model you could host locally? A GPT4 level model?

The thing I'd be most interested in would be them giving us some insight in their training data.

I wonder if this will force out a lot of older farmers, and move us to more corporately owned farms that have larger acreages and less employees

How are you configuring that for Claude Code? I'd be interested in seeing if I can get a PR in OpenCode to allow that

Have you thought of also integrating with OpenCode? opencode.ai. I personally have switched over to using it over ClaudeCode

This reminds me I need to recalculate the layout for my side project 😂

Any chances of getting a RSS feed?

I use @lmstudio-ai.bsky.social since it can use MLX as the backend without any tweaking

Was there a reason you chose Qwen2.5-Coder, or was it just the fastest/simplest to get working?

I'd be interested in having it use multiple models and comparing their outputs.

I've used Ollama and then the command line to `cat` a bunch of markdown files into a prompt and found some relative success with it.

It was not as nice as I would've wanted it to be, though, and felt a little clunky