You guys have a good product, but you really screwed up. I don’t mind paying for a plan that suits my needs. I paid for the $200 plan, which is great, I’m happy, and the service works well. But when I reached my limit, which took only a few days, I found that I couldn’t even use Claude 3.5 in task. That made me very angry. The truth is, as soon as another product matches yours, I’m leaving. That’s not how things are done. We all understand the costs, the energy you put into getting users, but if you throw it all away at the first hurdle, I don’t see a future in this. Anyway, to be clear, I love Cursor, it works great for me, but you also have to take into account that I cancel many of the requests, I accept them all but I have to fix them, and in projects with simply too many files, everything gets out of hand very quickly.
You ask Sonnet 4 for something simple and it makes 203929039 changes, which deteriorates the use of the plan.
You’re probably right. New products lure users in with cheap prices, breaking even or taking a loss to grow their userbase. Then once people are hooked, prices go up.
This was always coming eventually with Cursor. For me, it is a polished enough product that I’ll keep using it… but I’ve also never been a “heavy” user, and mostly rely on GPT 4.1 with the occasional thinking model use for tough problems.