Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!samsung!news.cs.indiana.edu!rutgers!aramis.rutgers.edu!athos.rutgers.edu!nanotech From: webber@csd.uwo.ca (Robert E. Webber) Newsgroups: sci.nanotech Subject: Re: Some problems of super-intelligence Message-ID: Date: 31 Dec 90 23:41:40 GMT Sender: nanotech@athos.rutgers.edu Organization: see disclaimer Lines: 91 Approved: nanotech@aramis.rutgers.edu .[You could (a) do the Moravec trick of replacing a few neurons at a time, . maintaining continuous consciousness throughout; ... . Personally, I would guess that people would begin by augmenting their . existing brains and replacing parts they didn't consider "personal" . (autonomic system, etc) and ultimately replacing the whole thing . part by part. . --JoSH] Hmmm, I think this trick predates Moravec - reminds me of a tale of a ship being replaced plank by plank; the question being when it was a new ship. If I take someone's brain and replace it cell by cell with the brain of squirrel, at what point would you say I have destroyed the person's brain? But, for many people, a squirrel's brain, the final end product of the transfer, would clearly be a distruction of their own brain/personality (I suspect - note I haven't actually tried this recently). Did I maintain continuous consciousness throughout? If not, at which cell did I drop the flow of thought? I think the `gradualism' approach sounds nicer, but isn't really any safer than the sudden download approach. Once we get to a level of technology where we can directly observe thought, we will feel much more comfortable about transferring it from one location to another. It will be interesting to see how the direct observation of thought is first demonstrated. As long as the augmentation is external, I don't think the change will be fundamental. Sitting at home surrounded by my books and computer access, in some sense, it isn't really me making this posting, but instead it is the accumulated knowledge of the room making the posting as processed through me [hmmm, I even have a Chinese grammar around here somewhere, but that's another posting]. What I write when I can look up a quote, track down a reference, do a computation, is often rather different than what I would say out on the street sans augmentation. With implanted computational resources, I will just be able to carry my library et al with me always (I will become this poorly integrated room creature and stop being me, but this is not fundamental since I already become the room creature from time to time anyway - although now, I also from time to time become an office creature, a library creature, and an outdoor walking creature, since the way I think about things varies with the different environments and access to different information - with implantation, all these creatures will merge (but this also is not something fundamentally new, since there are plenty of people who don't read very much and hence have similar information access continuously throughout their life)). When I see that kiwi are on sale in the market I will be able to dig up an article telling me how to tell the ripe ones from the spoiled ones or the unripe ones and make the purchase that I really want whereas now I usually have forgotten how to do this at the time the purchase needs to be made. And it will be nice to be able to make a quick scan of net news while waiting in line at the supermarket (after all, who would trust automatic food delivery to choose the choicest kiwis). Life will be somewhat richer, you will remember more details and be able to plan better what you want to do. But most of the benefits won't happen for the same reason that they don't when I am sitting in my room, i.e., just because I have a book on my shelf that has the perfect quote in it doesn't mean I will think to check to see if there is anything appropriate in that book. Although the room creature has more information access than the street creature, it is poorly integrated with the extra information. However, once the augmentation is internalized in the sense that information stored in the hardware is indistinguishable from information stored in the wetware, then some fundamental changes will occur. One such change will be the transferability of skill. After all, you don't make someone a musician by handing them a bassoon and a few books on music theory. But if they can directly access the memories of a professional bassoonist, I suspect they will actually have the skill of playing a bassoon. Of course, with current people, there would still be a problem due to variance from person to person of muscle tone, but there is no reason to believe that such variance would still need to exist once things can be engineered to this degree of precision. The interesting question will doubtless be how much of the memories of the bassoonist to transfer before one considers oneself to have acquired the skill. It will also be interesting to find out that certain skills require mutually inconsistant world-views (i.e., that someone can not simultaneously be a bassonist and a painter, for example). I think the world would also be incredibly richer. When you see a cloud you will be able to appreciate it simultaneously for its thermodynamic properties, meterological properties, and artistic aspects. All prior clouds you have ever seen will be available for comparison and everything anyone has ever publically said about clouds will be available for consideration. Similarly, with interacting with other people, you will remember every previous interaction as well as having a vast common pool of knowledge upon which to draw; hence greatly reduce the redundancy. Usenet postings will average 7 characters per message (no headers). --- BOB (webber@csd.uwo.ca)