I’m writing this post about a rival editor which name is prohibited to mention here and gets censored by the machine. It starts with W, and honestly, it wasn’t good. It’s riddled with bugs and doesn’t feel nearly as polished as Cursor. I swear this is my honest take as of March 13, 2025. After watching the W presentation (https://youtu.be/bVNNvWq6dKo?si=dvYCQp9ERyixvVGr), I was totally sold. I thought to myself, “I’ve got to try this!”
You can give it a shot too—it’s got unlimited DeepSeek V3, so there’s plenty to test out.
I also got the pro subscription and paid 18,5 USD for it to make sure that it is not a LLM issue.
But right from the start, it was a mess. It couldn’t find my Dev Containers properly, and then the ghost text (those autocomplete suggestions) was from time to time practically unreadable—definitely a bug there. Worse, it had no clue about the context of my task. It was generating random files all over the place, like a kid you have to babysit. The presenter in the video made it sound like W was steps ahead of the dev, anticipating your every move, but in reality, it just created more problems.
I ran the Linux version, and I am not aware if it behaves better on other OS.
There are a couple of features in W I actually like, but as it stands right now, it’s pretty much a waste of time and money to bother with.
W does excellent marketing while cursor seem to have not spent any on it. I got curious with W after a seeing a ton of posts on X coming up on my feed talking about how great it does great with context and what not. Complete disappointment as it was not able to fix simple issues despite me saying “feel free to completely refactor this component”. I fed this same prompt to claude code and in 1 shot it fixed the issue. Cant say anything good about cursor today as its been a non started. refuses to edit code and refuses to throughly search through the codebase to find the relevant code. maybe its my mistake to turn on “early access” in settings. on 0.47.3 now!
Aug… is probably the best competitor in my oppionon - but I’m not a good enough coder to use it. 0.46.11 here - also pretty useless today - have nothing good to say about it. I don’t have any experience with Claude Code but I’m probably going to give it a bit of a shot.
I tried W also and it was terrible(Had 2 accounts to get things done but still so bad) I moved back to Cursor+Aug and its be wonderful, yes we sometimes have fights with claude but I love it 100x better
I think this is the difference between a company like Cursor with amateur leadership versus the alternative. I for one would have a section in this forum to encourage discussion about rival services. It would provide Cursor with information about rivals that you just can’t buy. But to think it a good idea to code out the ability to even mention rivals is an amateur move!
The truth is that when you can get Claude-3.7 Thinking to work, it is exceptional! The biggest issue at the moment is the consistency and stability of it actually working.
Make Claude 3.7-Thinking reliable and consistently available and I wouldn’t need to consider rival services. It really is a testament to the coding team at Cursor, when it is working. Its the stability and testing free update policy that really needs some leadership driving it!
My experience today has been a prime example… It is 2pm here at the time of writing.
From 7am till 12. Cursor worked fantastically! Very astute! limited randomness. Had to go out until 1:30pm and now the last 30 mins, with the exception of one response, has been constant High Load Errors!!
Yes, I also think that removing the names of the rivals in the first place and not letting us to mentioning them, is kind of immature, and I have no clue what kind of threat W imposes to C.
They need to listen to users and customers and the community in general to adapt fast. Like very fast.
The C developers must know that UI comes is the last priority. changing names has 0 priority. stability and functionality and integration with external platforms is the key for this editor to hook into the mindset of devs.
Agreed!
Information is difficult to obtain. To actively go about preventing free information is exactly that. A childish approach that comes with lack of experience and professional leadership. Talented Coders should code, and be given the tools and boundaries to be creative with it by the leadership who can consider the bigger picture. Much much better if the leadership team only have a rudimentary coding knowledge, so they focus on the setting the right boundaries and strategy to deliver growth.
Coders running the show will only ever focus on trying to outshine competitor coders! Average coders with strong leadership will ALWAYS eventually outperform!
I had the same experience. Saw the video, had to try it. The “flow” was good, but the results were wrong. The worst part is that it thought it was done and that the task had been completed. Whiff!