> but in any case market efficiency is a very general claim, it doesn't rule out some "special" people with innate talent to predict future prices.

The "efficient market hypothesis" is one of those hypotheses that has Strong and Weak versions, with the weakest versions being obviously true, and the strongest versions being equally obviously false, and practically pseudo-religious. The strongest version of the Strong Efficient Market Hypothesis does indeed rule out the existence of such exceptional people (or at least their ability to profit), because it says that the market will always instantaneously reflect the knowledge of all participants. Such a prognosticator would find their ability to beat the market asymptotically approach zero, as his attempts to take advantage of opportunities cause those opportunities to evaporate in his hands at the instant of investing.

The survivorship bias argument to explain people like Buffett is ridiculous. Before you continue to tell myself and others to read up on CAPM, the EMH, portfolio theory, and so on (which I have, extensively), I suggest you read up on Buffett and his approach to investing. He vocally rejected the EMH decades ago, and has continued to produce outsized returns. The proof of the pudding is in the eating and Buffett has managed to be right much more often than the academics and their insane asset and derivative pricing models.