Accelerating Future

SIAI Interview Series: Ben Goertzel Monday, Jul 30 2007 

AI Michael Anissimov 3:43 pm

Trackback URI

6 Responses to “SIAI Interview Series: Ben Goertzel”

1. IanC says:

   July 30, 2007 at 4:34 pm

   Something — wildly fictional at this point — popped into my head whilst reading this, that rebuts the “Computronium Friendliness failure” and the “Formoulli’s Paradox” as well — it is based on M-theory branes being eventually accessible by some radically advanced society, which itself isn’t really all that fantastic a statement.

   IF these two things are the case, a sufficiently advanced AI would quite possibly prefer to manufacture or connect to a brane with physical laws far more conducive to its operation; this kind of expansion would relegate turning the planet into computronium so obsolescent by the time it would reach a point where such an attempt, quite feasibly, would be useful to the machine (given the logarithmic complexity / time lag issues of expanded networks), it would have access to such a technique (trans-’braneous’ travel). The reason why this applies somewhat equally to the Formoulli Paradox is because it begins to explain why, exactly, radically advanced societies do not expand greatly beyond their point of origin, even if exploratory in nature — it’s a matter of additional sums and increasing the sphere of travel necessary to accomplish ubiquitous presence a la the Paradox.

   Call it the Membranous Transcendance hypothesis. :)

   Reply
2. Nick Tarleton says:

   July 30, 2007 at 7:23 pm

   But it only takes one civilization that decides not to do that (or never figures out how) to start expanding through the universe. If civilizations are anything close to common in our past light cone, it probably would have happened.

   Also, I don’t think your logic applies to a utility-maximizing AI (as opposed to a satisficer – which is probably much safer but harder to build). If utility(convert Universe + other brane) > utility(convert other brane) – as seems likely – bye-bye universe.

   Reply
3. Tom McCabe says:

   July 30, 2007 at 10:40 pm

   “But it only takes one civilization that decides not to do that”

   You don’t even need an entire civilization- just a couple of functioning Von Neumann probes.

   Reply
4. steven says:

   July 31, 2007 at 6:10 am

   But can you build Von Neumann probes smart enough to colonize the universe and stupid enough not to colonize another brane?

   I find the utility maximizing argument convincing too, *but* I don’t think you can 100% rule out the possibility that even the relatively tiny amount of resources involved in building a Von Neumann probe could be better used toward colonizing the other brane.

   On the other hand, trans-membrane travel, if it’s even possible and other membranes even exist (I have no idea how all this would work), would probably require lots and lots of energy. AIs would be needed to gather it, and who knows what they’d do if left behind.

   I think transcension scenarios like this are very unlikely, but I wouldn’t know how to disprove them 100%.

   Reply
5. IanC says:

   July 31, 2007 at 8:26 am

   On the other hand, trans-membrane travel, if it’s even possible and other membranes even exist (I have no idea how all this would work), would probably require lots and lots of energy. AIs would be needed to gather it, and who knows what they’d do if left behind.

   Oddly, it’ll either require a great deal of energy or a very small amount of energy.

   Also, I don’t think your logic applies to a utility-maximizing AI (as opposed to a satisficer – which is probably much safer but harder to build). If utility(convert Universe + other brane) > utility(convert other brane) – as seems likely – bye-bye universe.

   I suspect you missed this part of the idea: “[...]to a brane with physical laws far more conducive to its operation[...]“(Emphasis this post). If the other brane were more conducive to computation as designed by the optimizing AI, then game theory dictates that it abandon this universe, essentially, for that other one; for while both are effectively infinitely expansive as resources, one will result in vastly or significantly greater return on investment; thus, all investment would likely be directed to this other brane (or series of branes, for that matter).

   And where von Neumann probes are concerned, what evidence do we have that there aren’t any in this stretch of the cosmos? Or that they haven’t already been in our Oort Cloud? We’ve mapped, what, (.000000001)^2% of it? That’s the teapot orbiting the sun argument, I know — but it’s all hypothetical, right?

   But can you build Von Neumann probes smart enough to colonize the universe and stupid enough not to colonize another brane?

   That strikes me as a very interesting question. I tend torwards the “yes, you can” side — at least, initially. I don’t know.

   Reply
6. Can ed drugs cause impotence. says:

   July 18, 2008 at 7:43 am

   Can you buy generic ed drugs without….

   Can ed drugs cause impotence. Is there free samples of ed drugs. Drugs used for treatment of ed. Ed drugs and bleeding. Is there free samples od ed drugs….

   Reply