« ITF #165 | Main | Batman Begins »

Blue Brain

One of the favorite tropes of both science fiction and extropian speculations about the future is the idea of uploading human consciousness into a computer. Uploading will require two things:

1. An appropriate storage medium for holding not only the data that a brain contains, but the metadata that defines relationships between the data, as well as the "application logic" that knows what to do with this data and the "operating system" on which the whole thing runs.

2. Sufficiently robust processing power to emulate the hardware functions of the brain.

Of the two requirements, the second seems the more daunting. Surely we have enough storage that we could back a brain up (should we figure out a way of doing that.) But reading a brain and playing it back...? That's going to take some doing.

Of course, these requirements overlay the "computer" paradigm onto brain function, which defines relationships between hardware, software, operating system, and database that are very different from what you would find somewhere behind the screen you're now looking at. But ultimately, that's what we have to do, unless we're prepared to create a machine that operates less like a computer and more like a brain...

...which leads to this highly interesting development:

Yorktown Heights, NY and Lausanne, Switzerland, June 6, 2005 – IBM and The Ecole Polytechnique Fédérale de Lausanne (EPFL) are today announcing a major joint research initiative – nicknamed the Blue Brain Project – to take brain research to a new level.

Over the next two years scientists from both organizations will work together using the huge computational capacity of IBM’s eServer Blue Gene supercomputer to create a detailed model of the circuitry in the neocortex – the largest and most complex part of the human brain. By expanding the project to model other areas of the brain, scientists hope to eventually build an accurate, computer-based model of the entire brain.

By the way, for those of you unfamiliar with the term, the neocortex is that special core brain part that only mammals have. That's how you can tell a mammal brain from a reptile brain. Look for the neocortex. From this site, I learned that the neocortical column (NCC) is a handy building block for higher brain function, and that a human brain is really nothing more than a robust collection of specilaized NCCs working together in harmony.

So in other words, once IBM gets a single NCC emulation running, they are well on their way to emulating an entire neocortex and, eventually, an entire brain. The question, then, is this: will a computer emulation of a brain produce a computer emulated mind? (For an interesting discussion on that point, go here, via Kurzweil AI.)

If so, Blue Brain will be the first step towards electronic immortality via mind uploading. And even if the "playback" piece doesn't work yet, the "backup" piece will be very encouraging to those who are looking to upload eventually. After all, if IBM can create a working electronic replica of a brain, they should, before long, be able to make a working replica of any brain, including yours or mine. And if the whole brain is backed up, it's just a matter of time before we have computers sophisticated enough to tease out that elsuive "mind" thing. (Unless it turns out that there really is something more to a mind than brain function, but I doubt it.)

But uploading isn't the only possible application. Personally, I'm kind of suspicious of uploading. I can't get past the idea that I get lost somewhere in the transaction, and that what survives bears no more relationship to me than, say, a different version of me living in a parallel universe. Sure, that other guy -- parallel universe or uploaded version -- is exactly like me, thinks the way I do, remembers the things I do, etc. But that's still not me.

I am some kind of process, some sense of continuity.

Here's another possible application for Blue Brain technology (somewhere down the road).

Years ago, I read a short story by (I think) Greg Egan in which everyone gets a tiny crystal storage device implanted at the base of his or her brain. Via nanotechnology, the crystal is "wired" to the entire brain and is essentially just backing the brain up as it goes, neuron by neuron. In a sense, the crystal is a sort of "warm standby" backup of the brain, the ultimate disaster recovery system.

The crsytal is diamond, and thus highly resistant to shock, changes in temperature, etc. So no matter what happens to you, no matter how traumatic, chances are the crystal comes out just fine. So whatever happens to me, I am a restorable entity. Via stem cell technology, they can grow back anything that's missing from my body, up to and including major chunks of brain. Then once my body and brain are whole, they just do a restore from the crystal, and voila!

It's me. Just like before the accident.

But then, what if my house burns down while I'm asleep and there's nothing left of me but some ash and the crystal? Well, from stem cells they could grow my body back one part at a time (to avoid cloning a separate human being, which would be wrong) and then do the restore. But at that point, how different is it from just loading the contents of the crystal onto a computer and running an emulation of me?

Why does it seem that one version would be me, and the other wouldn't? Am I just an organic bigot, or is there something else at work?



"I am some kind of process, some sense of continuity."

At the risk of getting excessively trekkie, isn't that the 'transporter' problem. That creature produced at the remote location isn't you however much he may think he is. He's just a point-in-time copy. Existence or the semblence thereof seem to rely exclusively on the perception of continuity. For all I know, every time I go to sleep 'I' cease to exist. When 'I' wake, the semblence is restored. Sort of like a computer program paged out. How would I know the difference.


"Ash and the crystal" - didn't they have a top-ten hit a couple of years back? ;-)

You would want the chrystal to contain a digital copy of your complete genome - just in case there was nothing left of your physical body to clone. A homing beacon would be a good idea too.

"how different is it from just loading the contents of the crystal onto a computer and running an emulation of me?"

Continuity. Well...only if the chrystal was doing more than data storage. If, after installation, the chrystal became actively involved in your thought processes or slowly compensated for those areas of your brain that wear out, this would become a part of the mind rather than just a backup. That continuity could allow you to make the trip to a new body or to virtual reality.

Any other backup scheme, including a destructive scan backup, leaves me wondering whether its live or Memorex.

"If, after installation, the chrystal became actively involved in your thought processes ... this would become a part of the mind rather than just a backup. That continuity could allow you to make the trip to a new body or to virtual reality."

Excellent point. If memory implants were available now, I would get them immediately. When eventually, all my memory is mecha, would I still be me? Will souls stick to crystal?


I'm for the full-uploading option. My thoughts here.

The application software may well be available, and it's open source too! Have you read "On Intelligence" by Jeff Hawkins (the guy who founded Palm Computing and Handspring)? He's a life long brain science geek and believes he's come up with a working theory of how the neocortex functions. He used his substantial fortune to create the Redwood Neuroscience Institute to further the research and just recently announced the formation of a new company called Numenta, to market software based on Hawkins' "Common Cortical Algorithm". There's a developer sign up to get access to the software as soon as it becomes available (later this year). If Hawkins is right, AI research has been going in the wrong direction and intelligent machines based on his theory may be a reality this decade.

You are right: the big problem will be recreating the "I" experience.

While Francis Crick & Company have been doing a great job with delving into specific brain function models and showing how much of an automaton each of us is when it comes to "free will" (how much of our "free will" is biophysical reaction), I think they are quite wrong in asserting that the observed phenomena are all there is to the problem of consciousness. Crick & Co. fail to account for "experience." I may not have free will, but "I" certainly experience the effects of my will, whether or not my "will" is simply a complex biophysical response to changes in environmental states.

A playback mechanism may be created for our high-fidelity brain backup system, and it may certainly appear that the playback mechanism is faithfully doing its job to the point where any observer cannot tell the difference between the "will" of the mechanism and the "will" of the original biological entity associated with that backup, but there is currently no way for us to know whether or not the doppleganger is actually "experiencing" anything at all.

Yep; the other insuperable objection to uploading is the duplicates problem. Any construct of hard- and soft-ware and data can be duplicated, as can, therefore, any "upload". Since both can't be the the original, neither is the original. Q.E.D.

The 'wake-up as a rebooted self' problem is also insuperable, IMO. Short of staying awake forever, there is no way to be certain that "my" existence didn't begin this morning and won't end when I next fall asleep. Maybe an explanation for my frequent insomnia? :D

Post a comment