You asked, I delivered! ⚡️ Introducing Automode for Claude Engineer. When in Automode, Engineer will just go through all the tasks it sets to solve without asking for confirmation. Create multiple files at once, etc. You can enter/exit this mode all within the same chat.

You can change the amount of max iterations you want Claude to go through, but of course, use it at your own risk! API costs can add up quickly and it's still a non-deterministic system.

Repo here:

I am so impressed with the reasoning capabilities of Claude 3.5 Sonnet. It's wild to see how Claude first thinks, then automatically decides when to search, when to read files, when to edit. It's really wild. No other model can do this so well, trust me, I've tested them all lol. @AnthropicAI really cooked with this one! 🔥

oh and of course it made this beautiful readme for itself lol

Was just thinking about this the other night. Having the AI engineer work overnight to solve a big problem without constant prompting would be incredible.

This was my first real "feel the AGI" moment in a few years. You can obviously tell this is the future and that everything is going to change forever

will be firing off PRs 🚀

thanks for sharing this, i learned a lot 😋

Great! Code up a stack for the AI to hack its own weights and let it self optimize. See if you can hit ASI before Ilya. 😉

This is absolutely insane, well done. Made a game in one command and it built everything