4 ms·

I suspect the simulation hypothesis is false. It relies on the premise that computers will get faster and faster until eventually we are simulating universes,

I suspect the simulation hypothesis is false. It relies on the premise that computers will get faster and faster until eventually we are simulating universes, but I suspect this will not be the case; we have been running into physical boundaries in making CPUs faster for quite some time. It's true that simulating a universe would lend itself well to parallelization in some respects, and it's also true that the universe would not have to be simulated in realtime in order for it's occupants to perceive it in realtime, but still it would be a monumentally expensive undertaking with current technology. Quantum computing may give us some big shortcuts here. I also read a clever paper some years ago that assumed a minimum amount of energy required to perform a computation, and then asked, if our CPUs continue to double in power every year, how long will it be until we need to convert all the matter in the observable universe into its rest energy to power our computer? As with all numbers that double on a regular basis, the answer is shorter than you'd think. But, there's a far more worrisome subset of the simulation hypothesis; when we train an image classifier, we show it a whole bunch of pictures of cats and dogs and "train" it. It stands to reason that when we want to build a general purpose AI, we will need to train it with general experiences. You could easily do that by recording a person's entire life and then feeding it to an AI. Hard to start recording a person right out of the womb, but you could make some minor modifications to the recording so that everyone in this "recorded universe" thought it was perfectly normal for everyone to not remember the first four of five years of their life. If this hypothesis is true, then I have good news; when you "die", there will indeed be an afterlife. There will be several in fact, as you are instanced across many servers and put into service doing natural language processing for users of the iPhone MMLX.

orwin 8y ago
I remember my philosophy teacher saying to the class that we can make 3 hypothesis: - Technology can advance until we can make simulation undicernable from reality: in this case, we are likely to live in this simulation - Any civilisation comming close to the previous point is likely to already have destroyed itself. In this case, living in a similation is unlikely, - Technology can't advance to this point in our universe. Living in a simulation is impossible. On of these proposition is true, but we are unable to know which one. So we have to put an equal weight on those, and the probability that we live in a simulation is 1/3. Sadly (or hopefully), Church-Turing conjecture is somehow backing the 3rd theory too, and i'll by anything coming from Turing.

orwin 8y ago

I remember my philosophy teacher saying to the class that we can make 3 hypothesis: - Technology can advance until we can make simulation undicernable from reality: in this case, we are likely to live in this simulation - Any civilisation comming close to the previous point is likely to already have destroyed itself. In this case, living in a similation is unlikely, - Technology can't advance to this point in our universe. Living in a simulation is impossible. On of these proposition is true, but we are unable to know which one. So we have to put an equal weight on those, and the probability that we live in a simulation is 1/3. Sadly (or hopefully), Church-Turing conjecture is somehow backing the 3rd theory too, and i'll by anything coming from Turing.