I think the idea is to disassociate newness from improvement and only associate it with improvement when it's actually improvement. It's not wise to pick a diff

I think the idea is to disassociate newness from improvement and only associate it with improvement when it's actually improvement. It's not wise to pick a different tool for every job, nor is it wise to only stick to one tool all the time. The fatigue imo comes from producing approximately the same output in ever more complex but not necessarily more interesting or better ways. If it does arguably provide a marked improvement, then definitely experiment and see how it might help