I agree that the lower timeouts are needed.
I would like to limit all my tests to not run slower than a reasonable low limit - let's say 100ms (or with exceptions 1s). This would allow me to keep the performance standards of my tests at a certain level, forcing me to spend a few more minutes to optimize the test - leading to generally good performance of the whole test suite at a later stage of development.
I want my tests to fail when they are too slow.
While it might become a problem on some CI pipelines, I would argue that it's not a good reason to deny this functionality to all.
Having more flexibility might open new use cases, not just unblocking stuck tests (which on its own is not a solution, more like a band-aid). Software engineers should be able decide on a good limit for their specific needs.