17 October 2007

Living Longer in Our Bodies vs. Uploading Our Minds

My natural preference between mind uploading and SENS/mental augment is to go with Aubrey de Grey. Anders' approach appears far too naive, neglecting many necessary details of brain to mind emergence. But--if I did want to convert a "wetware mind" to a "hardware mind," how would I do it?

There is only one approach that I consider even slightly plausible--the gradual replacing of wetware modules with hardware modules. Rudimentary mindware chips are able to act as "neural language relays" to bypass damaged/infarcted brain in animal studies. Such chips appear to facilitate neural recovery after infarct. Other chips may act to translate brain wave activity to electronic commands for robots or computer software--as in "Second Life" avatar or video game control.

These are only very crude beginnings in the path to eventually replacing wetware with hardware. Clearly, it will be necessary to utilise advanced nanotechnology methods that have not been invented yet. Hardware will have to be able to accurately emulate wetware over both time and space. The brain is an asynchronous device that at the same time is exquisitely sensitive to the timing and sequence of signals received.

The highly significant difference between an electronic computing device and an electrochemical brain anchored in a lifetime's experience of gain/loss, emotions, physical drives, etc. appears to be lost on many commentators on mind uploading.

Rodney Brooks and Jeff Hawkins are two of the more realistic researchers in the field of AI. I would be interested to know either of their positions on the possibility of mind uploading within the next century. My prognostication is significantly less hopeful than that of Anders Sandberg, Ray Kurzweil, or others who place this achievement well in the middle of the 21st centure--or sooner.

Contrast my pessimism on that front, with my optimism regarding Aubrey de Grey's SENS project--which I expect to yield many dividends before 2020. We are learning to manipulate the machinery of life to our own benefit. Through biological means, we will be able to live longer and think better. Longer life and clearer, more powerful thinking will allow us to better emulate human intelligence and consciousness in machines.

Hat tip Kurzweilai.net

Labels: ,

Bookmark and Share


Blogger IConrad said...

Ted Berger plans to have his alzheimer bridges completed & commercial by 2025. That's worth thinking about in this.

Of course, so too is the fact that no electronic medium has yet been built that could endure continuous operation for more than 5-10 years. (To my knowledge). And this problem is getting worse. Now that we're removing lead from our electronics, the tin-based soder is beginning to develop "whiskers" which create static electricity and kill the machinery.

View This entry on my blog for a more detailed explanation of the tin-whisker thing.

Wednesday, 17 October, 2007  
Blogger al fin said...

Interesting point, IC. Berger and Donoghue are doing fascinating work with nerve-machine interfacing. The newer cochlear implants and artificial retinas are also impressive.

Compensating for neurosensory and motor deficits with signal processors is a wide open and promising field. Going from there to a mind upload is like comparing a trip to the corner grocer with a voyage to the Andromeda Galaxy.

I don't want to take away from the incredible research that's being done. At the same time, the magnitude of the challenge should not be understated.

Wednesday, 17 October, 2007  
Blogger Michael Anissimov said...

"All" that is required for mind uploading is a scanning method that can image the brain at molecular or slightly-below molecular resolution, throw away as much knowably unnecessary data as possible (like water molecules), and instantiate that data in a virtual machine. Based on past trends in hardware and scanning devices, Kurzweil thinks this will happen in 2029.

It should be emphasized that no philosophical insights are necessary to do this. No one will need to develop the special "soul chip" that makes it possible to transfer over the mind into a computer. We just need a really high resolution scanner, a really high bandwidth data pipe, and a really high capacity supercomputer (probably around 100-10,000 times better than today's best, actually).

Jeff Hawkins doesn't even believe in human-level AI, and I don't think he has considered mind uploading. Rodney Brooks -- I don't know. There is little inclination to think about these things seriously. Nick Bostrom and Anders Sandeberg, who both have degrees in neuroscience, are somewhat optimistic (not sure how much).

SENS will provide benefits by 2020, but not immortality of course. By the time SENS really progresses to the point where it can give us a youthful body practically forever, I think mind uploading will be ready and many people will choose it, because it would make us practically immune to accidents.

Wednesday, 17 October, 2007  
Blogger Michael Anissimov said...

PS. It sounds like they told Aubrey to slow down the usual rate at which he talks, in this video. :)

Wednesday, 17 October, 2007  

Post a Comment

“During times of universal deceit, telling the truth becomes a revolutionary act” _George Orwell

<< Home

Newer Posts Older Posts