I updated to the latest version and it appears that the response space has been nuked. I was earlier able to get full file codes in the responses but now it’s barely finishing one.
Worse, continuing from where it left off breaks the formatting of the rest of the response.
Version: 0.31.3
VSCode Version: 1.86.2
Commit: fcd6e551d6f014d13071eac9bd4904e7e0346720
Date: 2024-04-02T20:20:11.019Z
Electron: 27.2.3
ElectronBuildId: undefined
Chromium: 118.0.5993.159
Node.js: 18.17.1
V8: 11.8.172.18-electron.0
OS: Windows_NT x64 10.0.22000
Thanks for reporting this - which model are you seeing the broken markdown with? Or is this happening reliably with all models?
The screenshot I attached happened with claude-3-opus. But I have seen it happen with other models as well.
I’m also getting the same error…
model: GPT-4o.
Lots and lots of bugs on your pimped vs-code IDE, makes it 98% unusable at a production level. It would be great if cursor simply kept it simple and just gave us your own unlimited slow API keys for $20 a month rather than forcing us to use a broken IDE. I’ll pay $40/month if that makes cursor happy, just give us the unlimited slow API. I don’t need the bells and whistles for code completion on an IDE which crashes or have tons of bugs. Cursor needs to make their own scratch made (written in c/c++) IDE.
1 Like
