Here's an interesting article on the future of science fiction in the face of accelerating change.

Cory Doctorow has made more of a career tooting his own horn than writing: his novel Down and Out in the Magic Kingdom had only a single arguably-new idea (woofie) and it wasn't a particularly good one. If he's the future of Science Fiction, I'm going to have to start reading Fantasy.

The reporter has missed a few things. Frederick Pohl was writing about human consciousness uploaded to computers decades ago in his Gateway series. And some of the ideas in the stories he mentions, like instantly downloading backup copies of memories to a clone, have already been done in some lousy movies, such as Ahnold's clunker "The Sixth Day."

The reporter also should have mentioned William Gibson, who has publicly discussed the problem of writing science fiction when the rate of technological change is so great. Gibson has said that he finds it really hard to posit a future only a few decades hence because there is simply too much change to take into account. The result is that Gibson has started writing stories set in the present, but emphasizing technology.

Sure, there are a few brave souls who try to work in the few-years-ahead, but I often find those stories to be too imbued with what the writer WANTS the future to look like. Science fiction remains story driven, and any effort to try to guess the future is doomed to failure. Which is why I never notice the implausibilities of Heinlein's "Farmer in the Sky" but find some of Sterling's stuff (e.g., Islands in the Net, cool but distractingly concrete) or that British fellow who writes stuff like Cassini Gap and such to be more wish-fulfillment than extrapolation.

Of course, this isn't a new issue at all - some writers were talking about accelerating change in the '50s and '60s (and don't forget "Future Shock"). In the early '80s, I came to the sad realization that my monthly Byte magazine described more interesting technology than most science fiction stories at the time. Vernor Vinge's "Marooned In Real Time" was a breath of fresh air, and Drexler's "Engines of Creation" suggested how some of it might happen.

While there is great change, I've become a bit more skeptical of a true "singularity" coming any time soon. AI research is still working on rather low-level stuff. We still don't know how our own brains work in anything like the detail needed to duplicate or expand on it. I do think we will see self replicating machines sometime in this century (whether nano, micro or macro) and that will open huge new possibilities. But I suspect regular humans will be running things for some time yet.

Is it just me or does the idea of uploading you consciousness seem very stupid. So you duplicate your memory, personality, everything. Now there are two versions of your consciousness but you are still stuck with version one. Its great for everone else, friends, family, consciousness 2, but for you living in consciousness 1, well you're still dead even if a second consciousness lives on. That was the main problem with the Arnold movie the 6th Day.

It doesn't seem much of an improvement than having a child.

yank,

I had never really thought about that until now, and I think you're right. If our consciousness is simply an emergent quality of brain activity, and not some actual thing composed of matter or energy that can be "caught" and transferred into a computer (I have no idea how), then we cannot extend our lives by "uploading" our consciousnesses that way. We can at best create a duplicate of ourselves to live on after we die.

Dang. It's a little depressing to realize that. But it's true.

Hans Moravec (robotics researcher, this is from "Mind Children") has an interesting take on that. Assuming you have the technology, take a person and scan part of the brain, say 1/1000 at a time. Duplicate the function in a computer and wire it up using nanowires to the rest of the person's brain. Give him a switch that allows him to flip back and forth between meatware and software. When he is satisfied there is no difference, remove that portion of the brain. Repeat until complete. At what point does he become a copy?

Moravec believes, and I think he is right, that identity is based on data and structure. Most of the chemicals in your brain that create what we think of as a "mind" are in a constant state of flux. Therefore, it could be said that the "I" of now is only an inexact copy of "I" of yesterday.

A note on "consciousness": I am always amazed at how hung up people are on this nebulous concept. Look: We just aren't concious of much of anything going on in our head. That is the key reason we don't have good AI. What we call "consciousness" is more of an operating system that coordinates other "unconscious" processes. It is a one process along with many, many other important mental processes.

Look at research on how brain damage can affect behavior. A person can have very specific problems - no longer able to recognize people's faces, recognizing people but no longer able to realize they are family, not believing that an arm is his own, etc. All due to point failures in various subsystems. We are a sum of those subsystems.

Yes, Minsky, Dennett and others say that consciousness is an illusion, which is probably true in some sense, but it begs the question, just who do we think we're fooling?

Rather than saying it is an illusion, I'd be more inclined to say that there are different interpretations of the phenomenon. Our current interpretation is deeply ingrained into our culture. In a similar way, we still speak of the heart as the seat of emotion, and that was widely accepted at one time.

One interpretation of consciousness is as a conversation or interaction between parts of the brain, much as we converse and learn from other people. If you think about how you do things, this makes sense: Some parts are handling vision input, others are interpreting what that input, others are "discussing" that interpretation based on their specialties, still others are directed to react, etc. The "illusion," such as it is, is because we are only aware of the results, not the process of how the brain works.

I think the "singularity-style" technology with the most near-term potential is probably neural interfaces. It's amazing how much progress has been made in the field in just the past few years. One or two years ago we had the first monkey controlling a robot arm with its brain waves; this year we're actually decoding planned behaviors.

Couple this sort of technology with massive data storage and wireless networking, and you've got some -very- interesting possibilities.

The nature of an uploaded personality is an interesting one. It will be a copy which will st
art diverging at the point it is uploaded. There certainly won't be continuity, but that's possible not necessary.

A similar problem could be considered with Teleportation - looking at most postulated mechanisms where an exact duplicate is created, rather than the Star Trek approach of sending all the atoms, you've got the same problem. One of you dies in the process.

I'm not going to let it worry me, something which thinks like me will, one day, I hope, be uploaded.

Dave, that's the "body centric" point of view. If you use the Moravec transfer scheme, there is no discrete copy, and there would be continuity. And by the data/structure point of view, using the "Star Trek" like transporter you aren't killed and copied but disassembled and reassembled elsewhere. I suggest that the "body centric" viewpoint is an artifact of our culture and current level of technology.

Moravec does get into some of the "more than human" possibilities after uploading, in which identity becomes a more complex issue. Multiple upload clones could go off to experience different things, then merge experiences later. "You" could remember doing several different things at the same time. You'd probably want to enhance the memory process over merely human ones so "you" could follow your various experience threads.