It really blew my socks off. This time, last year, I was using Windsurf and getting the agent to stop acting retarded was like pulling teeth. At times, I thought goddamnit I could have just done this ...
Yes. A little over a year ago I didn't trust LLMs to write code more than half a screen long. And sometimes it even produced nonsense there.
I still don't trust it for entire projects like the one you posted. I doubt it can write production quality code without anything strange or unnecessary hidden inside. But it can do largeish merges and refactors rather nicely these days.
I still don't trust it for entire projects like the one you posted. I doubt it can write production quality code without anything strange or unnecessary hidden inside. But it can do largeish merges and refactors rather nicely these days.
1