I think there's a rationalist antipattern, where an area is very data-poor, and one then decides that one can come up with the correct theory for that area using reason, e.g. Occam's razor/Bayesian updates/minimum description length/etc..
My experience is that this works surprisingly poorly in practice. Probably the reason is that people massively underestimate just how many models can be made which fit through the data one has observed.
In particular, one thing that has an especially high tendency to go badly is this:
So I'm asking you, LessWrongers to figure out how to weight credences between these theories to at least identify the most probable theory for our universe.
To illustrate how that can go bad, consider a biased coin with 60% chance of ending up tails, and imagine you flipped it 10 times. What's the most likely sequence you could observe from this? HHHHHHHHHH. But is this a typical sequence, whose properties you can use to probably predict things right? No, you'd often (not always, depends on various obvious things) be better off with something like HTHHHTHTTH as your prototypical sequence, even though it is less likely, because its aggregate properties are more accurate.
Though you'd be even better off by integrating over all the possible sequences weighted by the probability, which in this metaphor corresponds to keeping an open mind to different possibilities while also being very skeptical about any 1 of them.
This is a good point, but also kind of an oversimplification of the situation in physics. Imagine Alice is trying to fit some (x, y) data points on a chart. She doesn't know much about any kinds of function other than linear functions, but she can still fit half of the points at a time pretty well. Half of the points have a large x coordinate, and can be fit well by a line of positive slope. Alice calls this line "The Theory of General Relativity". Half of the points have a small x coordinate, and can be fit well by a line of negative slope. Alice calls th...
Near the end of his life, Feynman said, we do have plenty of data, it's all the unexplained parameters in the standard model.
Another hint: Witten once said, there has never been a theory which contained as many of the ingredients of reality as string theory (fermions, gauge fields, gravity), that turned out to be unrelated to reality.
Loop quantum gravity, on the other hand, is extremely overrated. The nonperturbative form of the theory has been incapable of recovering space-time.
Penrose's twistors are great, they have revealed deep new structures in field theory and string theory, and even have some relationship to the Ashtekar variables that are employed in loop quantum gravity (it's just LQG's method of quantization which is probably wrong).
The old paradigms of supersymmetry appear to be wrong. The old paradigms of dark matter are probably wrong too, given the number of correct predictions that come from "MOND". However, there are many known variations on these ideas, and we are also simply very far from knowing all possible forms of field theory and string theory.
The odds are that we are living in some kind of string theory possible world, that is slightly or markedly different from the ones that have received most of the attention in particle physics. However, non-string-theory hypotheses deserve to have people investigating them too, they may turn up overlooked pieces of the puzzle.
To the extent that these opinions of mine provide appropriate guidance, I can't say it's due to the use of any rationalist heuristics. It just comes from lots of study and thinking about the subject, and having Internet guides, for the best orthodox wisdom (Lubos Motl) and for promising ideas on the fringes (Marni Sheppeard, RIP).
I think you are asking the wrong question. What do you want to use these credences for? Are you planning to work on fundamental physics? Then the relevant question is something like "In what area am I likely to make progress?" or "Which of these fields offers the best opportunities for advancement?" If one does not expect to ever know what model, if any, will successfully unite QM and GR, then asking for credences is a fun but pointless speculation.
(Now, to express my personal speculative view for fun, I suspect that none of the existing models come close to anything like a "Theory of Everything", we are short a couple of paradigm shifts and about 10 orders of magnitude off in experiments, but there will be progress once experiment catches up by e.g. measuring gravcats.)
Now my personal credences are 55% String Theory with broken Supersymmetry, about 1-5% to Loop Quantum Gravity, and about 40-44% credence in other theories.
That means String Theory still is the most likely theory that applies to reality, but definitely not overwhelmingly confidently likely.
One other problem for String Theory is the black hole information paradox has been solved solely by normal physics, that is there is no need to posit strings to solve the paradox. Conjunctions are never likelier than their conjuncts, so solely solving the black hole information paradox using normal physics is strictly more likely than positing String Theory and the AdS/CFT correspondence.
Here's a link:
https://www.quantamagazine.org/the-most-famous-paradox-in-physics-nears-its-end-20201029/
the black hole information paradox has been solved solely by normal physics, that is there is no need to posit strings to solve the paradox
Well: this "solution" to the paradox was discovered by working within string theory. Then they found a way to carry out the argument at a level of approximation that doesn't need to know what the fundamental theory is. That is the only sense in which they were able to dispense with string theory. They don't actually have an alternative fundamental theory in which the argument can also be carried out.
Meanwhile, I p...
I would expect asking lesswrongers about this is approximately the same as walking into a meetup between CS, AI, and philosophy departments at a university and asking the same thing - there might be answers but they're gonna involve a lot of wishing there was a real physicist to ask, or perhaps getting lucky if one happens to be attending the AI meetup.
personally I think string theory is a very promising overcomplete representation space and we should scale it up and train it on large hadron collider data to create a superintelligent string theory. we'll know we succeeded when the new physical theory can tell jokes.
It's not an exaggeration to say that the flood of data in the 20th century for theoretical physics has now effectively vanished, and that leaves theoretical physics in a limbo as they figure out which TOE is true.
String theory, Loop Quantum Gravity and more theories try to duke it out, but almost no data can discern between these theories.
So I'm asking you, LessWrongers to figure out how to weight credences between these theories to at least identify the most probable theory for our universe.
(One hint: Supersymmetry, at least in it's non-broken forms, is unlikely to exist. If Supersymmetry is correct, then it has to be broken in a way that doesn't solve the Hierarchy Problem.)