On whether string theory is how reality actually works...
In our current state of half-knowledge half-ignorance about reality and about string theory, it is at least possible that this is so. String theory on a specific background can contain all the necessary ingredients to explain what we see. The immediate problems are
(1) we have no specific string model which predicts the exact values of the standard model parameters, and in fact it's not presently possible to calculate those properties of a "string standard model" to that degree of precision
(2) string models also contain other stuff that isn't seen so far.
Even when they don't contain "exotics", completely new particles with no relation to the known ones, they do contain superpartners of the known ones, and these have not been seen directly. (In various places there continue to be discrepancies between SM theoretical prediction, and experiment, so it's conceivable that we are already seeing the indirect effects of beyond-SM particles, but these discrepancies may also just go away as calculation and measurement improve.) The scalar "moduli" fields arising from the size and shape parameters of extra dimensions are another potential source of trouble, which also could be behind osberved new physics (e.g. aspects of the dark sector).
We can acknowledge these problems while also acknowledging that they are nothing like a falsification.
The following fact is also often presented as a problem for string theory:
(3) there are zillions of possible string models (possible shapes for the extra dimensions, etc).
The argument is that string theory can't be falsified, because there is always another model, so it isn't science.
I feel that the clearest way to interpret the significance of fact 3, is to think of two levels of ambition, in the application of string theory to fundamental physics. First is the quest to find a string model, among the zillions, that just matches reality - which gives the right ratio of electron mass to muon mass, and so on. Then comes the quest to find a reason why the world is like that, rather than some other way.
We are still very far from meeting even the first level of ambition. See my characterization of string phenomenology in #28, and also the second half of "fact 1" in this comment. There is definitely progress in string phenomenology, today's best string models look much more like the real world than those of the mid-1980s, and the ability to calculate has made great progress, but we are still very far from calculating quantities like that electron/muon ratio to several places.
Regarding the second level of ambition, the anthropic principle gets a lot of attention for the past ten years. This is the part that critics of string theory find really threatening: what if, instead of theories that make sharp falsifiable predictions, we settle into a new dogma in which there are handwaving arguments that various qualitative features of the standard model are generic in the landscape of string vacua? But before we debate the merits of anthropic thinking, I would like to say that it is very far from being shown that this is how string theory must work.
The issue of "vacuum selection" is basically an issue of cosmology - what is the large-scale structure of the world in string theory. And we just don't know whether the branching world of eternal inflation is how it works (and even if that is how it works, we don't know that it then goes on to ergodically explore the landscape). There are issues about de Sitter space, there is the question of where inflation's initial conditions come from. There may be a wavefunction of the universe which shows a sharp preference for particular vacua. None of this is worked out properly, it is all discussed very heuristically, and we probably won't really know which heuristics were the right ones - and thus, which picture is the right one - until there is further progress with the mathematical and conceptual fundamentals of string theory.
So at the second level of ambition, we may eventually get sharp predictions produced by cosmological initial conditions. But OK, suppose eternal inflation is how it works, and instead we get a prediction that the zillions of different vacua are being realized in different places, and all the properties of the standard model have a high degree of contingency.
A historical precedent would be the attempts to understand the detailed arrangement of the solar system as expressing some divine intention or natural principle. Today we can see that stars and planets come in numerous arrangements; they all obey the same celestial mechanics, but the particular arrangement we have in this solar system is not the only way it is. There are principles visible in it, e.g. Bode's law, and perhaps it represents a common "type", but still, ultimately there's no deep reason why e.g. the outer planets have the tilts and eccentricities that they do.
The same may apply to the standard model. There may be patterns there arising from deep relationships, and there may be other patterns which are just contingent randomness, which are only constrained by the anthropic necessity that physics be consistent with the existence of observers and/or atoms and/or a long-lived universe. The random patterns would be the junk DNA of particle physics.
Again returning to our current state of half-ignorance half-knowledge: in that state, this must be regarded as possible. We don't know it's true, we don't know it's false, we don't even know whether it is what string theory actually predicts. Anthropic string theory can't be excluded apriori - though any predictions it generates are liable to be suspect - and it leaves the first level of ambition untouched and still valid; one can still hope to identify a specific string vacuum which predicts measurable quantities to arbitrary precision.
Returning to that first level of ambition, where we just try to match experiment and make predictions, and don't especially care if our model is one among zillions of models - we try to falsify the model, and not all of them at once - I started by saying that it's possible that the existing program of string phenomenology may yet pay off. Maybe we'll see superpartners, and so on.
But, it is also possible that some of the guiding phenomenological assumptions are wrong. This is related to my comments #25 and #28, where I tried to explain that string phenomenology derives many of its ideas from a consensus about what to look for and what needs explaining, that originates outside string theory, in the culture of particle physics. There has been a prevailing assumption that the bare standard model is an unnaturally tuned theory, that there needs to be something else in order to make it "natural", and that superpartners with masses within an order of magnitude of the Higgs fit the bill. All this is in question now. The anthropic alternative is currently dominating discussion - the idea that the Higgs is tuned to a thousandth or a millionth part, in order to make atoms and a long-lived universe possible - and so anthropic string phenomenology is also getting a boost. But there are actually many other ideas on the margins, and it may be expected that they will also get their turn to be investigated, and that many surprises will show up.
Just at the level of field theory, there are many other paths to explore, and it's also true of string theory. People may begin to look at new parts of the landscape, and they may even look at neglected versions of string theory itself. There are many unexplored options, at that first level of ambition for string theory, and it may be that the experience of repeatedly being disappointed by the failure of superparticles, large extra dimensions... to show up, will have been necessary to drive people off the beaten track and look at the other possibilities.
