Hi.
I cannot understand the Cloud Agents feature of this Cursor. What problems or pain points is this feature designed to solve? Is it for writing code in the cloud? I can fully use Cursor locally, so why would I need to write code in the cloud instead?
Do programs written in the cloud then need to be merged into the local environment? Isn’t this an unnecessary extra step?
The Cursor 1.7 version feels quite good; at the very least, it can gradually get a task done. However, after upgrading to Cursor 2.0, it seems like Cursor can’t do anything anymore. Cursor’s Auto feature has become extremely clunky. With the previous Cursor 1.7, I could still see the Auto feature display some thinking processes, but Cursor 2.0’s Auto has none of that at all. Now, the 2.0 version can only handle some very simple tasks.
Cursor also keeps updating automatically. I uninstalled version 2.0 and reinstalled version 1.7, but after closing Cursor, it would update to the latest version again. I can’t understand why it’s designed to behave this way.
When I use Cursor’s Agent to write code, I have this experience: once the plan is finalized, Cursor Agent can quickly write the entire program. However, the program is completely unreliable at this point—basically, every function has issues. Then, as you continue to use Cursor Agent to debug each function, the program gradually becomes polished. But one day, when you close the original conversation and create a new one, Cursor Agent loses its previous memory and reverts to being that extremely clunky AI. Why am I saying all this? Because I feel like the Cursor team is just like this Cursor Agent: they always destroy the Cursor IDE right when it’s on the verge of being perfect.
–Best wishes from China