Brooders.net

I have a soft spot for Max Headroom. It was pretty cool in the 80s. One of the “dated” looks in Max Headroom (and other sci-fi) is no one expected TVs to be flat or small (not to mention the recording gear, which was just 80s video cameras).

But .. he’s back touting the joys of HDTV

@ Youtube http://www.youtube.com/watch?v=LKt7AJLIFL0

Amusing.

( Via Warren Ellis )

So, a couple of weeks after Udon and Artificial Machines, and a bit further into Age of Spritual Machines and I’m starting to change my mind on the topic … perhaps this AI and computational power is making a bit more sense.

So far I’m really really enjoying The Age of Spiritual Machines. It’s a pretty out there book, on par with most cyberpunk science fiction .. just that it’s mostly research fact.

A couple of tidbits so far that I like:

• Do people think or just calculate - most people will argue that machines calculate
• What will computing power in 2020 be like, (will we have strong AI) …
• Computational power of computing has been increasing since we’ve had computing, Moore’s Law is still at work (even without bizarro Nanotechnolgy or Quantum computing)
• Computers learn the opposite direction to people. Children learn language via sounds, then reading and writing symbols - while Computers can output and read symbols well before they can understand spoken lanuguage.

A thought occured to me while I’ve been reading. Some of my disbelief is probably due to the yuck factor, or self preservation context of my thoughts. Perhaps even pride. Surely the human mind is unique. If my mind could be transfered would it really be me? How would I know that I had been transferred to the new hardware?

But then it occured to me … I’ve gone under a general anaesthetic before .. how do I know it was me that woke up afterwards? I guess if you think of that as being “powered down” and then brought back, then the feeling of transferring conciousness to new hardware isn’t as hard to swallow.

Every night, we sleep (power down?) and come back in the morning with the same memories. So is the whole mind downloading/upload to new hardware just a technology issue? That would be freaky, yes .. upgrade me to new body … wake up the next day in a new “skin”.

Kurzweil seems to think so.

The book is a philosophical rollercoaster. My other current started reading book pile consists of “Future Shock”, The God Delusion, and “The Tipping Point” feels like they’re all subsets of this book. Future Shock because it looks at Acceleration Change, and people having more difficulty adapting … but I think Kurzweil is describing technology as an aid to this problem, augmenting humans for the better. The God Delusion is similar due to the absence of any Supernatural god, and the authors strong belief that evolution is how people became what we are today. It’s almost like this book is the God Delusion, just with a mix of technological expectations built in. The tipping point (although I’ve only just read 20 pages or so) is sometimes the irrational reason for why things occur, especially with the irrationality of people.

The whole mind uploading and downloading thing (often core theme to Cyberpunk fiction) is quite an interesting issue, which the book touches on. In the novel “Altered Carbon”, certion special forces soldiers were trained to be downloaded into new bodies and beamed across to wherever they were needed. Not only that, but most people had a real time backup of themselves in their “Cortical Stack”(Storage device somewhere in the back of their skull), as long as that wasn’t destroyed .. you could be brought back. So if you had a heart attack, got shot, bled to death you were capable of resurrection. The book was good in that respect, serious psychological trauma still required therapy once you were brought back. There were also laws against uploading yourself (or mind) into multiple bodies .. now that’s a mind bending topic. The Age of Spiritual machines also covers this very same topic, if you could non-invasively (and non-destructrively) scan and copy someones mind, and upload to another body .. both of them would claim to be the same person.

This book, probably best described as transhumanist, is almost cyberpunk science fiction. If I had read this when authored in 1999, I think I would have called it that. Now, it feels like this is actually in the realms of possibility.

I get the feeling, that as I keep reading the book this “Udon and Machines” series will continue.

Lunch today was bladerunner-esque.

I was sitting in a Japanese Noodle and Sushi shop, reading Ray Kurzweil’s Age of Spiritual Machines. While there was no asian harp music in the background, there was a bustling ambience as people bought their salmon sushi.

The evidence is below:

Some Kurzweil notions such as a computer will have the processing power of the brain by 2020, I don’t think, necessarily translate to Artificial Intelligence. There’s a whole lot of software that has to run with that hardware.

However, I just made it through the prologue (it’s hard to eat udon noodles and read), but some of his comments/points do make sense.

• The best chess software can beat most players - years ago this was not the case
• Intelligence applications are very specific, but probably surpass most people
• Will machines think, or just calculate?
• Do people actually think, or just calculate as a response to stimuli?

Then we get into the philosophical side of what it is to be human. If you could upload someones memories into a machines would it be human?, could you tell the difference over the phone?

My Hypothesis: to be human is to have experiences and knowledge in a human body. Profound? Sorry it’s not. I think we can seriously worry about this when/if it happens. Perhaps we could describe Humans and an Artificial Intelligence both as “Intellectual entities” and get away from narcistic self bias. These concepts still seem strongly in the realms of science fiction.

“The Age of Spiritual Machines” seems like a though provoking read, and that’s my opinion just after the prologue !

I found some recent interview with Marvin Minsky from this year (2007).

The three part series of “Where is Hal?” (Doctor Dobbs Journal):
Where is Hal? Part 1/3
Where is Hal? Part 2/3
Where is Hal? Part 3/3

A while back I was pondering that it might be time to look at his works, I have bought Society of Mind and started to trawl through the ideas (not necessarily in a linear fashion).

While listening to his talks, I always feel that “Artificial Intelligence” gets more complicated and tricky. I like what he says, but it makes any programming more difficult to deal with uncertainties !

Some ideas that stood out to me are:

• Statistical Analysis is not so good for linguistics or other representational tasks - Not easy to change (although not sure whether he thinks cased based reasoning falls into this bucket)
• Representation of ideas, thoughts or problems can be tricky, new models can come along but no necessarily solve the problem or provide useful application
• People have automatic systems - e.g. Adrenalin, body’s response to heat or cold - Therefore, we are not fully in control of our “hardware”, maybe intelligence doesn’t require total control
• The brain has redundant capabilities - he mentions in one of the articles an epileptic boy who no longer has his right lobe, but is still able to walk, talk and draw (although impeded)

This sort of information requires a lot of digestion, and re-listening .. I need to go over this again. He does have an interesting manner in “telling the AI story”, which at least makes you think about the how and why things work.