I’ve been using Cursor since the 0.x versions. In the past, I was genuinely excited about every new release — I would literally click Check for Updates almost every day. Lately, however, I feel the opposite: I hesitate whenever I see a new version available.
The reason is simple: almost every update breaks something. If it’s not one issue, it’s another.
Version 1.7 felt relatively solid and stable. But from 2.0 onwards, things have gone downhill — and there are already many posts out there about the problems with 2.0. Right now, I’ve pinned myself to 2.0.69. I did try 2.0.7x, but had to uninstall and roll back because of random issues, such as being unable to connect via SSH. I don’t have the time or energy to debug or file detailed reports for every regression, and as soon as I reverted, SSH worked again. That’s just one small example of the release quality problems I keep running into.
At the beginning, I still reported bugs. But occasional bugs are very different from what I’m seeing now: frequent, repeated issues, and regressions where a bug “fixed” in one version shows up again in a later one. It really makes me question whether there is a proper QA process or any serious testing before a version is released.
As a user, I feel like I’m being turned into a free tester who has to discover bugs and report them so the team can patch them. This feeling is made worse by the lack of clear, timely changelogs at the moment of upgrade (Today - Latest Version 2.0.77 but Changelog 2.0.1-2.0.28: Bug fixes). Updating should be something I do with confidence because it improves my experience. With Cursor today, updating feels risky.
More and more, Cursor gives me the impression of a product that’s being VIBE CODING rather than carefully engineered and released with discipline. I still believe Cursor has huge potential — but without stronger quality control and a more reliable release process, it’s very hard to trust new versions.
I must agree, for past few months almost every other stable version is filled with rather obvious bugs.
I sometimes notice specific features are shipped directly to a stable channel and never hit nightly or early access.
Also nightly/early access channels sometimes miss features already in stable build.
It seems like a specific feature is created in a separate container/branch and is merged into stable/dev/nightly on ‘random’, rather than going first into nightly, then early access and then stable.
Sometimes nightly contains a specific feature, that is removed in a next nightly and then added back in next nightly.
I understand some features must be reverted and revised, but it happens so often it seems like somebody creates a feature based on stable version 2.0.x and then this feature branch is set to be a nightly branch until somebody else overrides it with their feature branch.
It is really annoying and must be a real pain in testing..
LoL, everybody agrees.
Im here for a year now, since 3 months i dont update anymore. Thats it. I treat new releases as betas and do beta testing on my own every 2 or 3 new versions. NEVER use cursor *.0!
I personally feel ashamed when a new release drops just to have 10+ patches hours after the main release. But well, thats me.
like:
12:00 1.3 came out
12:30 1.3.1
12:40 1.3.2
13:00 1.3.4
13:40 1.3.6
…
20:00 1.3.21
couldn’t you just wait the 12 hours? I take half the price from the image damage every new release and find the bugs for you in no time! promise!
That said, just switched from 1.5.11 to 2.0.77 and the main differences are slighty different colors and the “you’re using an very old version” message swapped with a “theres an update” message despite using the latest downloadable (77 instead of 28 like listed in changelogs - btw. is the changelog site nearly killing your browsers too or is it just me?).
They said they tweaked performance, but its clunkier than before for me esp. on manual code edits.
Glad it did not set the new design by default or i would’ve punched my screen by reflex for sure
Bugs are still the same: reverting some files leads to indention errors
some changes are saved immediately, some have to be saved by hand.
Yes. My agent doesnt capture any Terminal output so it cant use the terminal which ■■■■■ real bad and Cursor hasnt even reacted to many posts about it.
Many of the changes for me have been bad enough that I have concidered moving to a different IDE. I have been with Cursor for a long time so I keep holding out hope that they will improve their process soon. But I probably will leave if they release 2 or 3 more broken “fixes”.
I have been around with Cursor now for more than a year now. I have never considered leaving until the last few months because it doesn’t feel like their product direction matches the reason I joined in the first place. I still have hope, but it feels like it is fading a bit
Agree, I believe there is too much variability when it comes to windows combined with WSL. So not sure where does it breaks. Maybe I should switch to macbook lol.
Browser integration broke in the latest version. You’ve really ■■■■■■ me off with releasing ■■■■■■ unstable versions. Hire some testers already who will test your ■■■■■■ editor. Or at least write tests.
