Slashdot Mirror
Downloading The Mind
bluemug writes "The Canadian Broadcasting Corporation's popular science radio show Quirks and Quarks aired a piece this weekend about Ray Kurzweil's ideas on downloading human minds to silicon. (The interview is available in MP3 or OGG.) Kurzweil figures we'll have strong AI by 2029 and be able to copy a human mind about a decade after that. Book your appointment now!"
It wouldn't work like that. Imagine that the copy of your mind is uploaded into a new body before you die. Do you think your consciousness would suddenly transfer to the new body? All the copying could do is create a new consciousness in the new body, your old one (ie, you) would still grow old and die in the old body, never to return.
This is also the argument against Star Trek transporters. You die each time you use one but a new person is created at the other end that thinks it's you. You don't know anything about this, of course, you've just been disintegrated!
TWW
"Encyclopedia" is to "Wikipedia" what "Library" is to "Some people at a bus stop"
Copying the information would require an extremely sophisticated, as well as invasive, set of technologies. Nanotech would probably need to be used to get the proper connections throughout the mind. As far as simply linking the brain, many people have discussed 'plugs' and such that would intercept external sensory/control feeds, such as the optic nerve and spinal cord, and then allow that information to be manipulated/redirected. Thus signals to move a leg could be altered so that they would move a mechanical leg, or even something else entirely. In such a way people could transplant their brains into robotic/cyborg surrogates, not even necessarily human looking. A fighter pilot, for instance, might just transport his brain into the plane. Thus the command to 'run' or 'walk' might be mapped onto engine throttling or some such. External camera's would send a feed, acting as 'eyes', etc. However, none of this makes any attempt at all to actually access stuff in reverse, from the brain. We record memories and such in the structure of the main brain, and thus something would need to go into the brain to read those. And because the 3-D structure of the brain is so critical, preserving the meta-information of how the other memories and such were encoded is also critical. Otherwise, you might end up with a record of memories and thoughts, but no way to actually connect those to form the personality.
Heh, I seem to be ending up with a long post, but the last thing to deal with, assuming sucessful duplication (including the metainformation) is "what now?" A way would have to be found to basically create an artificial neural net that would be able to recreate the exact structure of the original brain. Who knows, it might be possible to do such a thing virtually, having different sectors connected to each other and thus having a person exist in cyberspace. That, however, is pure speculation.
I actually find a lot of the stuff going on very exciting. Brains seem to last a lot longer then the body supporting them does anyway, so being able to basically have your brain in a very strong container that could be moved from body to body would probably work pretty well, and could potentially be very doable. However, total artifical replacement seems a long ways off. In some ways, what he is talking about in this article is sort of like cryrogenics today. You can get yourself frozen, but for the time being there is no way to ever undo the process.
Probably the best way to look at it is like a fork statement in a Unix program. Suddenly, one program becomes two, with different PIDs, etc. And you could make the case that the one with the original PID is the original program, and I suppose that in many ways you'd be right, but the fact remains that the new PID begins with EXACTLY the same memory and EXACTLY the same register states. The only thing different is where that memory is located. Clearly, the programs are seperate entities, yet they'll follow exactly the same patterns (ignoring their different data input) and we might argue that, from the program's perspective, the PID is entirely artbitrary and it has absolutely no way of knowing which one was not the original program, since they start off in identical states but respond to different stimuli. I actually think that twins are a very good metaphor, only you have to imagine that the twins both have common memories for a certain number of years.
I honestly don't think words like "dual consciousness" or "death" even apply in situations like this; we need some new vocabulary.
No, I very much doubt these kinds of predictions (and it's got nothing to do with the issue of the transferrence step).
What counts as our "minds" are simply far too tied into the physical instantiation of our bodies. (Not that "mind" is too abstract, but that it's not abstract enough for separation from our bodies.) If I make a computer-based simulation of myself, will it get tired? Hungry? Thirsty? Itchy? Horny? Sick? If not, can it then get excited? Scared? Concerned? Bored? Will it have any emotional reactions at all, if all the standard physical stimuli are removed?
Even if all the "human" inputs are replaced or simulated -- you've still got an added problem of a new level of "hardware breakdowns" on whatever platform is running the simulation. Suddenly you've also got to deal with the various downtimes, pauses, glitches, etc., that will break the illusion of it being the same "mind" as in the original person.
People are simply too much a construct of their wetware to be able to remove their "minds" as a separate set of procedures.
We know where leadership by an anti-intellectual "strongman" who scapegoats minorities and likes boisterous rallies goes
The problem with depending on hardware and network advances to drive his vision is that software engineering simply cannot keep up with the pace of advances on the hardware front. Anyone who has ever read the "Mythical Man Month" understands this at a basic level. Humans simply cannot organize themselves well enough to tackle software projects of the magnitude that Ray envisions, at least not by 2029.
Ray dismisses this argument by saying we'll have software that writes the software. Well, there's a tautology for you. If you can't write the software you need because it's too complicated, how can you possibly be expected to write the software to write that software? Genetic algorithms are useful for some very specific trial and error sorts of problems. But using them to random walk our way to a billion lines of debugged, functional AI code seems a bit of a stretch.
My money sure isn't on Ray's pony...
Shut up and eat your vegetables!!!
Thought experiments like that are fun, and yours is convincing, but consider some of the implications...
During the transfer from inside to outside, suppose you use a machine that has redundant circuits. Each nanobot is replace by a trio of simulated external neurons, so that they can check each other for errors. (If the presumably-binary output of the three disagree, the majority wins and the disagreeing unit syncs to the final result.)
Ok, up to now it's exactly the same situation that you describe, but with additional reliability.
But after the transfer is complete... The trio-links are broken, resulting in 3 perfectly synchronized systems.
Which one is you?
I'm not sure that "continuity" proves anything. Maybe your original consciousness would die slowly, neuron by neuron, as the new consciousness comes to life. If it even does come to life.
Honestly, I don't think the human race yet has the terms to describe the problem, much less speculate about the answer. It's fun to talk about, but so was "how many angels can dance on the head of a pin?"
Understand anything?
.sid Saturday morning.
My Windows 2000 SP2 Athlon doesn't understand HFS+. My OS 10.2 iMac needed alot of help understanding a
And niether of them know anything about the cordless phone, nor have they met my washer and dryer or the sexy new fridge just 3 meters away.
As for all the awareness coming to the desktop near you, I'll state here and for the record, each new CPU from Intel and AMD will be brought to thier knees by the new versions of Gnome, Windows, and Quake, leaving the desktop user with the same number of free cycles as before.
Now for this mythical AI thats coming surely by 2040, I say poppycock! My stance towards such "advanced thinking", and "futurists" is the same stance I have towards people telling me the World is ending right after the World Series.
I didn't buy this whole Strong AI nonsense a while back, but I did read one of his books. I was left with a sense of wasting my time then.
I popped over to Ray Kurzweil's site and poked around. This bit got my attention.
"If we can combine strong AI, nanotechnology and other exponential trends, technology will appear to tear the fabric of human understanding by around the mid 2040s by my estimation."
Those are alot of "Ifs". Strong AI, which is a buzz word. Nanotechnology, which is something that is built 1 or 3 or 10 at a time and photographed, but does nothing at this point. And my favorite "other exponential trends" In other words this whole idea of a Bowie-esque Savior Machine depends on crap he doesn't understand or can't put in words, but he is sure is coming.
Poppycock.
The capacity for thought we have is an intensely complex combination of the neural processes of survival and reproduction, with all those billions of years behind it, plus the geologically recent development of a whole lot of extra cognitive juice in the frontal lobe department, plus a couple of million years of tweaking this wetware system in the context of social, tool-using behavior, plus several tens of thousands years of social behavior combined with the meta-social instruction of language, art, text and such...
We have some bare inklings and theories of how we acquire language, intelligence, social functioning. The barest inklings. We're working on it. There is still a helluva lot of controversy on what exactly intelligence is, and no end in sight.
At the dint of enormous effort we have computers that can take a stab at interpreting meaning of isolated phrases based on context and a whole lotta cultural and semantic training by humans. The most powerful computers built for the job can hit-and-miss beat the finest human chess players... a game with a fully mathematically limited scope which is almost entirely susceptible to a brute force approach. "Seeing" computers can be trained to make some decent interpretations based on heavily patterned information. Voice recognition still has to be tuned to every individual, and it's pretty damn iffy for all that. Nowhere near to a computer that can hold anything resembling a conversation.
So where in hell do we get an estimate like "Strong AI by...?" As far as I'm concerned science has barely framed the question of what that would mean... and only in qualitative terms at that. So I'll tell you where these pointless predictions come from: ballpark some meaningless figure about biology - numbers of neural gaps, firing rates, impulses per second, whatever. Connect it in some arbitrary manner to some measurable function in a computer, extrapolate based on some law of technological development with far less than a century of statistical evidence and no basis mechanism whatever behind it (statistical evidence without an explanation is ALWAYS suspect in interpretation) and - viola! - you're a futurist. Or, as I like to say, a worthless dumbass.
ANd how do you get from there to the process of downloading consciousness, despite the fact that there is not even an inkling of a glimmer of the slightest valid theory about how an active and continuously shifting neourochemical proccess of personality and intellectual template, stored memory and present cognition (not to even touch the primal, the emotional, the glandular, the spiritual) gets translated to something that can be interpreted by a machine or stored in a meaningful sense or caused to be active outside of a biological framework? Well you just pull that one right out of your ass because it doesn't have even the flimsiest basis in the "reality" of doodling with a few facts and figures on your scientific calculator.
I'm not sure exactly why but the idea of people making carreers based on this bullshit makes me so mad I could kick puppies.
It Is the Nature of Information to Transgress Artificial Boundaries