Yes, how scarce scientific resources are spent is an important question. I think Smolin is right to be worried that scientific talent and energy is being focused too exclusively on a particular approach to basic physical reality. And in fairness, I must grant that as a research program, string theory does appear to be developing in ways that are only very tenuously constrained by experiment.

Thanks for bringing this up

Even the preposterous Matrix scenario still needed biological brains to occupy the vats.

As you say, the history of science offers plenty of examples of correct hypotheses that were put forward before there were ways of testing them, for example atomic theory and general relativity (though Eddington did come through quite promptly in the latter case). I haven’t read Smolin’s or Woit’s critique of string theory (though I’ve seen some interviews with Smolin, who I think is an extremely gifted physicist and a very creative thinker, and have read his remarkable book The Life of the Cosmos, which is worth a post of its own). I’ve spent more time listening to string-theory evangelists like Brian Greene, and I do have to admit the theory is captivatingly beautiful. I should read Smolin’s and Woit’s arguments in depth. Certainly Smolin’s point, that if string theory is nothing more than pie in the sky we are squandering the careers of some of our brightest young scientists, is an important one.

Woit has a number of good criticisms of string theory, but he goes too far in claiming it lacks proper scientific bona fides. Testability is a virtue, but theoretical advances, especially in cosmology, often need to mature before they can testable hypotheses can be crafted. Time will tell.

Bostrom’s speculations, on the other hand, seem to me to simply skate past telling difficulties. If implementing the right program was all that it takes to be conscious, then any Turing machine, even one constructed from tinker toys (it’s been done!) would be potentially conscious. Somehow, that strikes me as bonkers.