[rgbrenner]

Recently canceled cursor. I think there's a shift happening right now with the improvements in the ability to process large context sizes and stay on task:

Traditional code editing -> autocomplete -> file editing -> agent mode

This is basically a gradient of AI output sizes. Initially, with the ability to generate small snippets (autocomplete), and moving up to larger and larger edits across the codebase.

Cursor represents the initial step of AI-assisted traditional coding... but agent mode is reliable now, and can be directed fairly consistently to produce decent output, even in monorepos (IME). Once the output is produced by the agent, Ive found I prefer minimal to no AI for refining it and cleaning it up.

The development techniques are different. In agent mode, there's far more focus on structuring the project, context, and prompts.. which doesn't happen as much in the ai-autocomplete development flow. Once this process shift happened in my workflow, the autocomplete became virtually unused.

So I think this shift toward larger outputs favors agent-focused tools like CC, Aider, Cline, and RooCode (my personal favorite).. over more traditional interfaces with ai-assistance.

[weego]

Yes, my initial distrust lead me to believe that I should only trust whatever ai agent it is to handle the granular issues in code. But they're fundamentally bad at it unless you spend more time prompting and re-prompting than a human with the right context would spend doing it properly.

Now I've changed my technical planning phase to write in a prompt-friendly way, so I can get AI to bootstrap, structure, boilerplate and usually also do the database setup and service layer, so I can jump right into actually writing the granular logic.

It doesn't save me planning or logic overhead, but it does give me far more momentum at the start of a project, which is a massive win.

[jjice]

If they're small corrections, I generally agree that manual changes are the easier solution. I have been recently trying to correct it and then having it generate a Cursor rule to tell itself to avoid that initial mistake (style and structure situations) in the future. Doesn't always work out, but it's handy when it does.