https://xcancel.com/mitchellh/status/2036866220449030168

I was expecting something more pragmatic than 'lean into AI even harder' to be honest.

GitHub needs to slow down with the AI shit and spend manpower fixing what's broken. Actions is a complete fucking disaster.

Also I have no idea what Pierre is but their website is horrible.

Being pragmatic means admitting AI is unstoppable whether you like it or not. Stopping copilot BS doesn't conflict with that

AI should not even be part of this discussion, it’s frankly irrelevant. None of the complaints Mitchell or anyone else has would be fixed by deleting Copilot or adding new agentic features. Deliver a stable and secure development platform first, then do AI stuff.

The big idea is all agentic interactions should critically rely on GitHub APIs. Code review should be agentic but the labs should be building that into GH (not bolted in through GHA like today, real first class platform primitives). GH should absolutely launch an agent chat primitive, agent mailboxes are obviously good. Etc. GH should be a platform and not an agent itself.

Why do I want that running on somebody else's computer? It's bad enough that most developers already rely on Anthropic or OpenAI. What value does a remote working repo add?

This sounds like massive centralization on GitHub and super ugly product coupling, instead of rolling open standards. I'm now glad Mitchell doesn't run GitHub.