@FectaTr1@theo Absolutely! I literally can't use the official releases at work without multi-repo ๐
And thanks to you for getting orchestrator v2 into Logan's fork!
@RanuareG@theo I've been using it, but just through a fork that adds multi-repo. I literally can't use the official release at work until multi-repo exists ๐ at least not very smoothly
@jullerino I'm not sure how much work would need to go into dynamically setting up T3 Code though, would need to setup the agent auth automatically somehow as well, so I completely understand if this isn't planned!
@jullerino I'm currently thinking: PR webhook trigger -> create env (or reuse if it's a follow-up event) -> run everything in the new env. Without dynamic envs, all the threads will have to run in the same machine and use an MCP to access anything in the VM
@kentcdodds Yeah that makes sense. How does useOptimistic handle the original state changing while the optimistic update is still active? Does it rerun the update function with the new data?
@mattpocockuk I just still have moral issues holding me back from going all in. The training data is morally grey, the energy usage is really high, and the generative stuff when it comes to artwork and articles/story is straight up stealing. I do still use AI, but only as a last resort
@mattpocockuk Do you have an article or some examples of these "type-only namespaces"? Sounds interesting, I think I recall something about them being useful for specifying types as part of other things? Not sure if I'm remembering correctly ๐