There was some publicity recently about John Cramer not being able to get funding for an experiment that he claims could prove "retrocausation" or backward time travel. I'm confused as to what the experiment is supposed to show. Cramer seems to be saying the outcome will depend on whether or not his transactional interpretation of QM is true, but that can't be the case, since all interpretations are supposed to make the same empirical predictions. Elsewhere, he seems to suggest the "sending information into the past" part works only if the world deviates from pure QM with small nonlinear effects. Does anyone understand what's going on? I certainly don't think QM lets you send information into the past, but the experiment still bothers me. Let's say that Cramer is right, and this sort of experiment really does let you set up a causal loop. If this is something Nature hasn't safely tried out yet (and I don't know if that's true -- anyone?), couldn't there be all sorts of risks we don't know about? For one thing, it seems to me that once you have a time machine, it's consistent with the laws of physics for absolutely anything to come out, so long as it goes back in later. People have made some predictions of risks (like black holes and strangelets eating the Earth) associated with particle accelerators, but at least we have strong evidence that those risks aren't real, since they would already have happened naturally due to cosmic rays. Do we have some similar argument why we know successful time travel experiments wouldn't blow up the world? If the answer is, "Oh, don't worry, there's only a 1 in 1000 probability that this experiment will destroy the universe", I don't find that very reassuring. Is doing the experiment worth one thousandth of a universe? On the other hand, if you can destroy the world with a 20,000 dollar experiment, I guess that in the long run we're screwed anyway.