why would you use an IDE if you are not the one coding? You can review the code in github/gitlab using prs, but not necessarily code.. I feel IDEs are not the best environment for coding agents UI because the prioritize code editing, and the interaction with agents has a different priority.
i still code, but i'm searching an IDE that preserve the human in loop, for example showing better the code added/changed by the agent, but yeah i would like to have a better interaction with the agent while coding, infact i tend to use small models with high inference so i can procede with him without losing the flow.
I mean if you are using version control like git you can always see diffs. I would say it might be easier to go that way into checking small changes into commits and dev branches, then you can try a very good diff UI that will refresh each change. Thats one way.
Thats very interesting view. I actually had some experience with that kind of dense language for specifying when I start testing using two llms to code: one to instruct the coding agent and the agent itself. The instructor start creating quite dense english that sometimes I could not easily understand, but seems to make the coding agent pretty efficient in understanding the requirements.
Do you mean Claude Code? If so, that's what I use(d) primarily for development, and Claude Desktop for general chats. My issue with Opus was that, every time I start a new task in Plan mode, it'd use 50k - 100k tokens and that'd by about 20% of the session limit. A bit of back and forth and its done for most of the work day. Just not feasible at all. The tasks I wanted it to perform were fairly small and contained, "Look at these three files @@@ and add xxx to @file. DON'T read any other files. If you need more context, ask me.". That worked sometimes but not always, still burned a lot of tokens.
reply