If you want to really blow your mind with regards to time, look at QED. I highly recommend the Feynman lectures.
The net result of quantum interference is that light tends to move at the speed of light, particles tend to move forward through time, but only because everything else mostly cancels itself out.
(The rest of this is slightly rambling, but so was the article so it's hard to comment without matching it, so dusting off my very rusted physics degree...)
With Multiple-Worlds, I think the delayed-choice experiments more show that each universal snapshot is inherently self-consistent, rather than retro-casual. Any non-consistent snapshot is just canceled out. Meaning, there was probably a universe in which the 2-slit delayed-choice experiment didn't show interference, but it didn't survive past the delayed choice boundary. Now, the really big question (to which people keep arguing no, but I have doubts) is can you establish faster-than-light communication using a delayed-choice style effect by essentially canceling out any casualty that doesn't agree with your communication.
My personal favorite for explaining the directionality of time is looking at the boundary conditions. If you assume that at time T0 around the big bang, the universe looked like X, then it requires that particles move away from it in spacetime (otherwise it wouldn't look like X anymore). Thus, there is probably an antimatter version going the other direction in time away from us. Away from the boundary, entropy-like effects would prevent the odd positron coming back in time at just the right moment to produce a non-local causality violation.
That sounds like it might be testable by looking at the CMB for evidence of interference with the antimatter universe. It lso looks a bit like a solution to the whole "why is there more matter than antimatter" question. That's a very interesting hypothesis you've got there. :)
The net result of quantum interference is that light tends to move at the speed of light, particles tend to move forward through time, but only because everything else mostly cancels itself out.
(The rest of this is slightly rambling, but so was the article so it's hard to comment without matching it, so dusting off my very rusted physics degree...)
With Multiple-Worlds, I think the delayed-choice experiments more show that each universal snapshot is inherently self-consistent, rather than retro-casual. Any non-consistent snapshot is just canceled out. Meaning, there was probably a universe in which the 2-slit delayed-choice experiment didn't show interference, but it didn't survive past the delayed choice boundary. Now, the really big question (to which people keep arguing no, but I have doubts) is can you establish faster-than-light communication using a delayed-choice style effect by essentially canceling out any casualty that doesn't agree with your communication.
My personal favorite for explaining the directionality of time is looking at the boundary conditions. If you assume that at time T0 around the big bang, the universe looked like X, then it requires that particles move away from it in spacetime (otherwise it wouldn't look like X anymore). Thus, there is probably an antimatter version going the other direction in time away from us. Away from the boundary, entropy-like effects would prevent the odd positron coming back in time at just the right moment to produce a non-local causality violation.