Given how quickly things evolve, it’s easy to get lost in the numerous offerings and hard to get the best deal. So, what do you use? Both clients/harnesses and LLM providers or local setups would be interesting.
Personally, I’ve been using opencode with Github copilot for work. I’m currently looking for cost-effective provider for personal work. Maybe openrouter with one of the cheap models?
I haven’t found anything that isn’t a shell backdoor that you let an llm run commands in, so strictly via chat interfaces.
Codex has a web interface that hosts little virtual machine environments, you give it access to a git repo and then you can approve it’s pull requests and test them locally, and roll back crappy code…
It’s a little weird to setup at first but pretty nice.
If they weren’t helping the military autonomously kill people I’d keep using it, but I don’t want to make it stronger.
You could also run your own virtual machines to let other ai fuck around in so it doesn’t escape whatever root you give it.