Saturday, March 26, 2005

Darkness at the Break of Noon, Goddamned Turtle Tank, etc.

"Darkness at the break of noon / Shadows even the silver spoon / The handmade blade, the child's balloon / Eclipses both the sun and moon / To understand you know too soon / There is no sense in trying."

Dylan usually gets it right....

Arrrghh.... I'm in an oddly dark mood this Sunday at 5PM, probably not a good frame of mind to be blogging, but it's a good way to delay cleaning out my son's turtle tank (my son doesn't live in it; his turtle Rick does) -- I don't really want to clean the tank but I know I have to do it, and if I start working on something intense or start playing the keyboard, the turtle will probably end up swimming in its own excrement for yet another day....

Hmmm ... the fact that I'm blogging about turtle shit probably indicates that I'm in a bad mood....

Not a bad day overall -- I got lots of interesting AI thinking & writing done, and took a long walk in the woods with the dogs. Sorta miss my kids as usual when they're at their mom's for the weekend. And, an interesting guest from New Zealand is arriving in a couple hours. Oops, better mop the dogs' mud off the floor, too.... (Sometimes I wish I lived in a country like Brazil where you don't need to be rich to have a maid! Is cleaning up really the best use of any fraction of my potentially far too limited lifespan? Well, if you saw the general state of my house you'd realize I don't think so!)

Maybe Buddha was right: all existence is suffering. Well of course he was right, but he left out the other half that Nietzsche said so well: "Have you ever said Yes to one joy? O my friends, then you have said Yes to all woe as well! All things are enchained, all things are entwined, all things are in love." Or something like that. In German, which I can't read, except for a few words and phrases. Everything is all mixed up, that's the nature of humanity. Almost every experience has some suffering in it -- only the most glorious peak of joy breaks this rule. And semi-symmetrically, almost every experience has some joy. Semi-symmetrically, because the mix of joy and pain seems rather differently biased for different people, based on variations in neurochemistry and situation. Most of the time I have an asymmetrically large amount of joy, I think -- as Dylan was well aware, it's not always easy to tell -- ....

Blah blah blah.

In moods like this I seriously consider giving up on the whole AI business and doing something easier and more amusing. I could become a professor again, write philosophy books and math papers, record CD's of weird music and write novels about alien civilizations living inside magic mushrooms.... I'm really more of a philosopher/artist type, software engineering isn't my thing ... nor is business. Not that I'm bad at these things -- but they don't really grab me, grip me, whatever metaphor you want today ....

Getting rich would be nice but I don't care too much about it -- I could live quite comfortably according to my standards without being rich, especially if I left accursed Washington DC for somewhere with cheaper land. Wow, I miss New Zealand, Western Australia and New Mexico ... great places I lived back in the day ... but it seems I'm stuck here in the DC metro for another 10 years due to a shared-child-custody situation.... Well, there are worse fates. And it's a good place for business....

Well, OK, time to clean the damn turtle tank! I could try to portray the stupid turtle (yeah, they really are stupid, though my son claims he can communicate with them psychically, and they're not as dumb as snakes) swimming in its own crap as a metaphor for something, but I don't feel perverted enough right now. Or at least, I don't feel perverted in the right sort of way.

Years I used to delude myself that I was just, say, 6 or 12 or 18 months away from having a completed thinking machine. That was a fun attitude, but it turned out I wasn't quite self-delusional enough to keep it up forever. I've now gained a lot more respect for how idiotic we humans are, and how much time it takes us to work through the details of turning even quite clear and correct abstract ideas into concrete realities. I've tried hard to become more of a realist, even though it makes me significantly less happy, I suppose because getting to the end goal is more important to me than being maximally happy.

I still think that if I managed to turn Biomind into a load of cash, or some rich philanthropist or government body decided to fund Novamente R&D, I could lead a small team of AI geniuses to the creation of an AI toddler within a few years. But realistically, unless a miraculous patron shows up or DARPA undergoes a random brain tremor and suddenly decides to fund one of my proposals, it's likely to take several years before I manage to drum up the needed funding to make a serious attack on the "Novamente AI toddler problem." (Yeah, I know, good things can sometimes pop up out of nowhere. I could get an email out of the blue tomorrow from the mystery investor. That would be great -- but I'm not counting on it.) Honestly, I just barely have it in me to keep doing software business for 3-5 more years. Not that it isn't fun sometimes, not that it isn't challenging, not that I don't learn a lot -- but it's an affront to my "soul" somehow (no I don't believe in any religious crap...). And no, it's not that I'm a self-contradictory being who would feel that way about any situation -- there are lots of things I love doing unreservedly, software business just isn't one of them. The difficulty is that the things I love doing don't seem to have decent odds of putting me in a position to create a thinking machine. I love music but I'm not good enough to become a star; and I'm about maximally good at fiction writing IMO, but my style and taste is weird enough that it's not likely to ever make me rich..... Urrgghh!!

Y'know, if I didn't have kids and obscenely excessive alimony payments (which were determined at a time when my businesses were more successful, but now my income is a lot lower and the alimony payment remains the same!! ... but hey, they only go on another couple years ;-p), I might just retreat to an electrified hut in some Third World country and program for three years and see if I could make the Novamente toddler myself. No more business and management and writing -- just do it. Very appealing idea. But Zarathustra (oldest son) starts college in a couple years. The bottom line is I'm not singlemindedly devoted to creating AI even though I think it's the most important thing for me to do -- I'm wrapped up with human attachments -- family attachments, which mean an awful lot to me.

Funny, just this morning I was reflecting on how great it was to be alone for a change -- the kids are with their mom, my wife is overseas visiting her family, the dogs and cats and turtle and gerbil don't quite count (OK, the gerbil almost does...) -- how peaceful and empty it felt and how easy it was to think clearly and work uninterruptedly. But now I see the downside: if a dark mood hits me there's no one to lift me out of it by showing me a South Park rerun or giving me a hug.... Human, all-too-human indeed!

And now this most boring and silly of my blog entries comes to an end. Unlike the previous ones I don't think I'll publicize this one on any mailing lists! But I guess I will click "Publish Post" in spite of some momentary reservations. Maybe someone will be amused to observe that egomaniacal self-styled AI superheroes have the same erratic human emotions as everyone else....

How important is it for this kind of human chao-emotionality to survive the Singularity? I'm not saying it shouldn't -- but isn't there some way to extract the joyous essence of humanity without eliminating what it means to be human? Perhaps there is. After all, some humans are probably "very happy" 5-10 times more often than others. What percentage of happiness can you achieve before you lose your humanity? All human existence has some suffering wending through it, but how much can it be minimized without creating "Humanoids"-style euphoridic idiot-bliss? I don't know, but even though I'm a pretty happy person overall, I'm pretty sure my unhappiness level hasn't yet pushed up against the minimum euphoridiotic boundary ;-p

And in classically humanly-perverse style, I find that writing about a stupidly unpleasant mood has largely made it go away. Turtle tank, here I come! Suddenly it doesn't seem so bad to do software business for a few more years, or spend a year going around giving speeches about AI until some funding source appears. Why the hell not? (No, I haven't taken any drugs during the last 10 minutes while typing this!). There's plenty of joy in life -- I had a great time doing AI theory this morning, and next week I'll be canoeing in the Everglades with my wife and kids. Maybe we should bring the turtle and let it swim behind the canoe on a leash?

Ahh.... Turtle tank, turtle tank, turtle tank. (That thing has really gotten disgusting, the filter broke and I need to drain it entirely and install a new filter.) Yum.


Mentifex said...

Ben, did you know that you have been Slashdotted by Mentifex? IMHO you've got a great chance of being _the_ AI leader of the next several decades, and so I make sure to link from Mentifex AI to Goertzel AI at every best opportunity. I sincerely hope that links from me turn into funding for you -- somehow. -Arthur (Mentifex)



Weird post Ben. Interesting though. Keep the interesting posts coming. I've linked to your blog from my web-site:

Regarding AGI I'm sure it's equally frustrating not only for the few heavy-weights like Wilson, Yudkowsky and You, but also for the thousands of interested ameuters like myself looking into it as a hobby, but seriously hobbled by lack of time, funds etc.

In terms of a time-frame to real AGI I used to be a big skeptic, but the longer I looked into it, the more I found msyelf revising my projections towards the near-term.

There are quite a few 'big money' players getting into the AGI game now. Millionaires like Jeff Hawkins for instance:

I continue to be blown away by the snow-balling pace of events.. news from the labs continues to pour in at a ferocious, ever accelerating rate.

I've tried to be skeptical, I've tried to be conservative... but now it seems very hard to believe that the Singularity could be more than 15 years away.

Even assuming the need for further major theoretical break-throughs (for instance in probability theory, as I speculated on SL4), I can't see this Pre-Singularity existence lasting more than another 15-30 years.

I think if we hustle a bit a 15 year time-frame is well within the bounds of possibility. I still think it could take twice that long. But certainly not more than 30 years.

Anyone who wants to play the AGI game should probably make up a 15-year plan, and save money, study etc... it should be possible for someone to save and study for the first 10 years enough for a no-holds barred full-time assault on AGI in the final 5 years.

Anonymous said...


Amy Watson said...
This comment has been removed by the author.
Amy Watson said...

Parkinson's disease usually appears gradually and progresses slowly. At first the victim may be troubled by “resting tremors” of the hands and feet that diminish or disappear when the patient begins to move. Parkinson's treatment