@thor @icedquinn
In day to day stuff, sure, quality of life or happiness are good benchmarks to use. Reminds me a bit of Jordan Peterson's view of truth (basically what should be called pragmatic truth). If something works and delivers happiness, it's something worth considering as valuable, even if it's not true in the scientific sense of the word.

But when it comes to applying the scientific method, you can't really use any other value to benchmark your models than accuracy and truth.