> A bias towards "beauty" or simplicity of some form is inevitable and part of good science.

> For every theory, you can generate a more complicated theory with more moving parts that generates the exact same predictions.

I agree with the second point, but disagree with the first.

Yes, by Occam's razor, a simpler theory with fewer moving parts should be preferred over a complex theory, if both produce correct answers. But the most accurate theory might not be "beautiful".

What matters most is if the theory correctly predicts reality. If it has a weird arbitrary format, but correctly predicts everything, it's better than a theory that's beautiful but incorrect.

Most of science has been using "beauty" as a way to help guide it to the correct answers. For example, there's been a lot of work on string theory because it has a kind of mathematical beauty. But it may turn out that a search for "beauty" in this sense was misguided. Beauty is not the goal, truth is.