The goal is to find e.g. Lagrangian, which consequences are in agreement with nature ... but calculating these consequences is quite tough - AI could generate required simulations, and it is already starting.
The question is where to search for such e.g. Lagrangian? There are already many people having own models developed for decades - AI could help to objectively verify with simulations - building kind of arena for models, to select the ones in the best agreement with nature, identify their best features.
After checking the already available physics models from humans, maybe succeeding would be also AI generated - e.g. combining what was the most successful in the tested models.
Crank science is easy to separate from non-mainstream approaches. Crank science is "everybody is selling nobody is buying" business. There are no physicists inside *or* outside mainstream who buy Consa's crank theories.
(I explained why the narrative is not valid in another comment).
One can get grants in physics nearly only if being mainstream ...
Consa brings concrete arguments regarding g-factor, I still haven't seen any concrete explanation, only saying that it is fringe science because of criticism of mainstream ... but he quotes mainstream papers.
“I must say that I am very dissatisfied with the situation because this so-called ’good theory’ does involve neglecting infinities which appear in its equations, ignoring them in an arbitrary way. This is just not sensible mathematics. Sensible mathematics involves disregarding a quantity when it is small – not neglecting it just because it is infinitely great and you do not want it!. ”
and Feynman:
"The shell game that we play is technically called ’renormalization’. But no matter how clever the word, it is still what I would call a dippy process! Having to resort to such hocus-pocus has prevented us from proving that the theory of quantum electrodynamics is mathematically self-consistent. It’s surprising that the theory still hasn’t been proved self-consistent one way or the other by now; I suspect that renormalization is not mathematically legitimate.”
This is an outdated, pre-1970s view of renomralization. Thanks to work of Wilson (1982 Nobel prize) and others on the renormalization group, we have a much better understanding.
Thanks, I will read, but generally beside e.g. the gravity problem, with increased accuracy there appear inaccuracies all over the standard model, so maybe it is worth revisiting QED? Are you saying the g-factor inaccuracies are not a problem?
The discrepancy in the article you linked is completely unrelated to how well we conceptually understand renormalization (which is broadly applicable to many quantum field theories, not just the standard model). It could be the case that the standard model is wrong, but my claim would still stand.
This patent covers rANS variant which is used for example in https://en.wikipedia.org/wiki/JPEG_XL - if granted, only Microsoft will be able to make its hardware encoders/decoders.
The question is where to search for such e.g. Lagrangian? There are already many people having own models developed for decades - AI could help to objectively verify with simulations - building kind of arena for models, to select the ones in the best agreement with nature, identify their best features.
After checking the already available physics models from humans, maybe succeeding would be also AI generated - e.g. combining what was the most successful in the tested models.