Hacker News Reader

Comment by gojomo | Hacker News Reader

While it can't be proven either way, something I find interesting about Bostrum's formulation (and before it, Hans Moravec's 'Pigs in Cyberspace') is that it does seem possible to infer some probabilities.

As we get closer to a world where our own ability to create world-simulations approaches verisimilitude to our 'real world', it becomes more likely that we are ourselves in such a simulation.

Some people find this idea existentially-threatening, but if it's any consolation, those running the simulation can't be sure they're not in a simulation either.

For example, let's say we're inside a simulation run by a race of hyper-intelligent turtles, whose universe simulation, after a few tredecillion ticks, has given rise to our current perceived reality.

Well, it's turtles all the way up.

View on HN

"Probabilities" are deceiving here. It's pure speculation and so any talk about probabilities is only adding a pseudoscientific veneer to what is really a high class entertainment.

Science fiction minus fiction does not equal science.

What if the turtle's universe was a dream thought up by the great flying spaghetti monster?

My probability math is rusty, but it's my understanding that probability is only relevant in a finite system. The article (in a very sneaky way) assumes, as it's premise, an infinite set of 'simulations'. But if this is the case, wouldn't that negate the usefulness of their 'probability'? In other words, if your conclusion is that we could be living in a simulation running inside a simulation running insi.....(keep going until blue in the face) doesn't that undercut your ability to say that one course of action is more likely than another?

Or, even better, isn't this the same as the proof for god that goes: "Our world is so complex and the probability of it coming about by by accident is so small that there must be an intelligent being controlling it'

(I was a CS/Philosophy double major in college, so my theoretical probability skills may be off, are there any math majors out there that can clear this up?)

The real numbers are infinite, and probabilities involving continuous functions on the reals make sense; it's just that P(X=c) = 0 for all constants c, and you have to integrate the probability density function to obtain probabilities. See http://en.wikipedia.org/wiki/Continuous_probability_distribu...

The trick is to come up with a probabilistic model for how they could have implemented this simulation that we're in, assuming that we are simulated.

Is there a way to exploit the simulation?

If I were to create a simulated universe, the program would have assertions that test for an interesting condition. (Say, if the simulated humans can achieve X, then break, and print a stack trace so we can study what lead to X)

When the universe simulator breaks, does that cause a perceivable difference for us, the simulated? Not necessarily. We would never experience such a break because by definition there is nothing to experience.

We can exploit the program if it is likely that the simulation program includes glitches. There are likely to be glitches if certain computations are innately difficult to do (e.g. NP hard?) so our simulation architects had to "hack" an estimation. For example, the strangeness of quantum mechanics may be just a poor (albiet optimized) implementation of newtonian mechanics.

Robin Hanson's suggestions for what behavior would best exploit living in a simulation are worthy of a debate all in themselves:

http://www.jetpress.org/volume7/simulation.htm

Abstract:

If you might be living in a simulation then all else equal you should care less about others, live more for today, make your world look more likely to become rich, expect to and try more to participate in pivotal events, be more entertaining and praiseworthy, and keep the famous people around you happier and more interested in you.

Reply on news.ycombinator.com