It was amazing to me how bad cursor is with using the same model I use in Claude. Even with little knowledge on how to test the llms I was able to get very minimal mvps. But I find the real trick is to have the proper tools to reign in the ai.

Thorough CLAUDE.md, that makes sure it checks the tests, lints the code, does type checks, and code coverage checks too. The more checks for code quality the better.

It’s just a bowling ball in th hands of a toddler, and needs to ramp and guide rails to knock down some pins. Fortunately we get more than 2 tries with code.

Cursor needs a paradigm shift to remain relevant, what was spectacular at first now is just banal and better done by other tools.