georgiedenbro wrote:Incidentally I'd say the one weakness of Frank's conception of AI is that I don't think he correctly conceptualized what the capabilities of AI might one day be. I think he thought of machines as being linear-thinking calculators that could do so as a prodigious rate, but that certainly will cease to be an accurate description of what they do in not too long. Once we get into quantum computing and bio-neural circuitry the game will change incredibly, and we're not that far off from opening up these areas.
Do you think Ship fit that description?
Good question. I'll admit that the last time I read Destination: Void was maybe 15 years ago, and I only read one additional sequel in the series. As it happens this series is the very next thing on my reading list to both revisit and finish, so if I remember I'll come back to this thread in half a year or so and comment again. Right now I'm just finishing up the Hyperion series which coindicentally also goes into some depth about AI development in the future. From what I recall (and I feel like the 2nd book gets into this more than DV did) Ship has superior cognitive capacity than man, but there comes a point in computing capacity where it's difficult to determine whether the computer is simply processing so many steps ahead that it seems like magic, or whether it's actually using non-linear methods. Chess computers at the present time, for example, can now defeat grandmasters regularly; indeed, I'm told that even a tiny program in your phone can defeat a grandmaster using current heuristics. Contrast with 15 years ago where it took the most advanced supercomputer to basically be even with Kasparov. But the methods of the programs have advanced more or less exclusively in terms of allowing the software to ignore certain lines of inquiry, thereby saving steps. They still use the completely brute-force method of linear elimination of possibilities. But if you put a chess program like this up against a chess player from 100 years ago and told him it was a person - would he have known the difference? And if you put that program up against a mediocre player, would the chess algorithm seem like higher-level thinking? And yet the thinking is linear, just as Frank explained in the Dune series. Based on this I don't know whether Frank saw Ship as being a linear thinker but with incredible brute-force capabilities, or whether it was a different kind of processing altogether.
In Dune Frank definitely posits that there is a different kind of processing than the linear kind, and that men can do it even if most don't know it or are bad at it. The critical point of AI development is when the AI can be self-improving, and Frank alluded to this briefly with the Ixian hunter-seekers, but it was mostly expressed as a direct danger rather than as a competitive intelligence in its own right. It does seem like Ship was this self-improving type of AI where it becomes something quite other than what it was designed to do, but I don't really remember enough from the book to say more.
What do you think?