[Kitsune]

Phil Masters wrote:

Well, I have, and I agree with Jürgen on this.

And so did I. That's what I said. But whatever. To be misunderstood is often the fate of great minds. :)

I'm sure that they will. I've read Permutation City. But in order to make these changes - especially if you want to make them with some kind of solid guarantee that the subject's personality will be preserved - you've got to have an extremely detailed understanding of human brain function, and especially that of the brain in question, right down to the level of the individual cells. And then you've got to tweak the structure, as modelled, with some kind of certainty that you'll get the desired result.

Well, that sounds not very convincing, doesn't it? You need a detailed understanding of brain functions to make the Ghost in the first place. Take (again) the example of the guy with the bad memory. If your understanding of his brain is detailed enough to make a computer imitate this dudes forgetfulness it should not be that difficult to simply stop doing that.

Can it be done in the TS setting? Reliably enough that people would volunteer for it? Possibly. It'd be a resource-hungry and expensive sort of task, though, I suspect. And for that matter, nanomods can do similar things to the meatbags - so you should be worrying about the same problem for flesh characters, too...

Hmmm? Do you really think that the problem could be to find Ghost guinea pigs for the tweaking? I mean, when most SAI's and Ghosts do not even fear death because they always have a backup of themselves made? And as far as the wellknown "is the copy really me?" problem is concerned...since most Ghosts had to undergo the destructive uploading process to become what they are, it's hard to believe that they should be very troubled by this. Plus, they most probably did the transformation to ghost status to transgress their limitations in the first place. And now they should stick to those silly limits artificially imposed by a simulation programme forever? I don't think so. I suspect that many of them would start to tweak and to change as hell asap...and if something does not turn out right, there is always the backup.

And by the way, personally, while I agree that AIs will be not human in fairly significant ways, I don't think that they'll be too alien. Remember, they're created and largely trained by humans, to work with and for humans, If they came out too unlike humans in their behaviour, that'd be inconvenient, and the next software generation would be designed to be more user-friendly.

I agree, a typical AI will seem quite human. But I think it is one thing to, let's say, imitate emotions (to be "user-friendly") and another to truly have feelings. So, I suspect that a typical AI will have an "interface" part that imitates human behaviour or even cracks a joke now and then...but deep down its workings are quite differently from a human brain. (For example telling a joke to "ease the tension" - after concluding that this is now a good idea - but not really finding the joke funny itself. Which does not keep the AI from making its avatar or its android cybershell or its voice partake in the laughing after the joke has been told. Otherwise the easing of the tension would be less effective...)
And let's not forget something very important about AI's: their minds content, their memories, their values, their beliefs...can all be modified. Changed instantly. And back again if need be. That may be their most "alien" feature of all.