For anyone currently thinking about fundamental physics, this latest Edge question is easy, with an obvious answer: string theory. The idea of unifying physics by positing strings moving in ten space-time dimensions as fundamental entities was born in 1974, and became the dominant paradigm for unification from 1984 on. After 40 years of research and literally tens of thousands of papers, what we've learned is that this is an empty idea. It predicts nothing about anything, since one can get pretty much any physics one wants by appropriately choosing how to to make six of the ten dimensions invisible.
Despite this, proponents of the string theory unification idea refuse to admit what has happened to it, often providing excellent examples of Planck's observation about what happens as scientists grow old while staying true to ideas that should have been discarded. Instead of retiring a failed idea, lately one hears instead that what needs to be retired are conventional ideas about scientific progress. According to string theorists, we live in an obscure corner of a multiverse where anything goes, and this "anything goes" fits right in with string theory, so fundamental physics has reached its end-point.
The "string theory" answer to the 2014 Edge question is however much too simplistic. String theory unification has long been a moribund idea, but it is just part of a much larger circle of now-failed ideas dating from exactly the same time period. These include so-called "grand unification" schemes that propose new forces and particles, generally invoking a new "supersymmetry" that relates known forces and particles to unseen "superpartners". Besides finding the predicted Higgs particle, the other great discovery of the LHC has been that the superpartners predicted by many theorists aren't there.
The period around 1974 brought us not only string theory, grand unification, and supersymmetry, but also something called the "naturalness" argument. The idea here is that our best model of particle physics, the Standard Model, is just an "effective theory", an approximation valid only at observable distance scales. Ken Wilson taught us how to use the "renormalization group" to not only extrapolate the behavior of a theory to short distances we can't observe, but also how to run this backwards, finding an effective theory for a fundamental theory defined at unobservably small distances. In a technical sense, "natural" theories are the ones where what we see is insensitive to the details of what happens at short distances. "Naturalness" became part of the speculative picture born in the 1974-era: complicated new physics involving unobserved strings and superpartners could be postulated at very short distances, with a "natural" theory all that is visible to us. In this picture it is technical "naturalness" which ensures that we can't see any of the complexities introduced by unobservably small strings or superpartners.
Wilson was among the first to point out that the Standard Model is mostly "natural", but not entirely so due to the behavior of the Higgs particle. At first he argued that this meant that at LHC energies we should see not the Higgs, but something different. Fans of superpartners argued that such particles had to exist at roughly the same energy as the Higgs, since if so, they could be used to cancel the "unnaturalness". Long before the LHC turned on, Wilson had retracted this argument as a blunder, deciding there was no good reason not to see an "unnatural" Higgs. The sensitivity of its behavior to what happens at very short distances is not a good argument against it, since we simply don't know what is going on at such short distances.
The observation at the LHC of the Higgs, but no superpartners, has caused great consternation among theorists. Something has happened that should not have been possible according to the forty-year-old reasoning now well-embedded in textbooks. Arguments are being made that this is yet more evidence for the multiverse. In this "anthropic" view, anything goes at short distances for bubble-universes elsewhere in the multiverse, but we see something "unnaturally" simple in our bubble-universe because otherwise we wouldn't be here. The rise of such reasoning shows that sending the "naturalness" argument into retirement (along with the epicyclic complexity of strings and superpartners) is now something long overdue