Jetbrains AI Assistant

Hi,

I tried out the JetBrains AI Assistant and I was initially confused as to why it only had 2 stars. I initially thought that people just had unrealistic expectations for the awesome company JetBrains, at times.

To be clear, the product isn't bad. However, I dug deeper to try to understand the rating. I gave Cursor AI a shot, and now I see what everybody's talking about. I'd say the plugin, as it stands right now, is 2 or 3 stars.

But… looking at Junie, that one competes well with Cursor. Basically, what I would like to see:

1.) A great diff tool. This should be supported for both inline and chat prompts. Right now, it's just inline from my understanding.

2.) A nice smooth workflow. As I accept the commands, as I get errors, it would be nice to quickly relay that back to the AI model.

3.) The ability to upload images. Heck, why not the ability to upload business rules (feature files, etc.)?

4.) Of course, an easy way to revert back changes.

5.) A composer.

I think Junie has most of these. I signed up for beta testing. For JetBrains AI Assistant, I feel like I'm using technology from a couple of years back. But I can see it getting really good by 2025.3, since JetBrains has always been good with meaningful updates.

But… I don't know how Junie and JetBrains AI Assistant would coexist. Would JetBrains AI Assistant maybe stay mostly the same and be a competitor of Tabnine, and then for those developers that really want to take advantage of AI, they can subscribe to Junie, which will cost roughly the same as Cursor AI ($20/month) and have all of those nice features that I spoke about in my bullet point list (numeric).

Also, I would have said to JetBrains, be careful embracing AI too much. If developers get TOO productive because of the AI support, then companies would get rid of some and not need as many JetBrains licenses. But… JetBrains isn't really in a position to dictate that. If they don't invest in AI, then some of their users would just go to Cursor AI and those same developers might get so productive that we still have companies' layoffs.

Probably the best thing for JetBrains is that their customers don't embrace it too much. Just enough to give JetBrains a reasonable number of great profits. For the ones that do embrace it, they would be happy and stay here, instead of going to Cursor. But to reiterate, it's in JetBrains' best interest for the AI revolution to not explode too much.