[<< | Prev | Index | Next | >>]

Tuesday, August 26, 2003

Robot Economics

I've been saying for a few years that AI is going to be the next big thing, and I'm finally starting to feel vindicated and encouraged to see the word popping up with ever increasing frequency, along with "robot" and "humans losing jobs". (I'd like to see a chart of the prevalence of words like "AI" and "robot" on the net over time--seems like a feature Google or the Wayback Machine should have, but I haven't been able to turn one up. Anybody?)

I do find it a bit discouraging though that there seems to be more mind-share given to doomsday concerns than to the technology itself. It's pretty rare I hear anyone ask "how do I code up an AI?", but I frequently get "how will all the out-of-work humans survive?" and "how do we keep the robots from taking over?"

My own relatively safe prediction is that AI is going to happen a lot quicker than most people expect, but a lot slower than most people fear. And the consistency in that is in recognizing that intelligence, and even consciousness, just isn't all that big of a deal--neither so terribly hard to implement, nor so overwhelmingly powerful. (I think it would be a lot less confusing if we would just stop anthropomorphizing humans.)

Well, eventually the robots will take over, but if you consider that the natural human population is projected to level off and start declining by the end of the century anyway, it will probably happen as a gradual replacement rather than a war. In the long run, humans will live on as novelties in a robot world, and many will probably upload and graduate to immortality at the end of their biological lives anyway (or sooner, after enough "oh, I remember when I was a human--boy did that suck" conversations).

In the meantime, it is an interesting question to ask how the economic landscape will evolve as robots start displacing unskilled human labor. I suppose the most convincing analogy that this is really a problem is to look at the productive value of horses over time. There was a period when a horse was a valuable worker--on the farm, for transportation, and elsewhere. But now they're pretty much relegated to entertainment, having been displaced in the workplace by more efficient (generally more specialized) machines. I don't know the stats, but I assume the number of (captive) horses to humans has declined sharply in the last couple of centuries.

Unfortunately, the number of humans of average intelligence isn't going to decline so easily, and we can't just turn them into glue, so what will they do? Joel* and others posed this view to me recently and I was more optimistic that the market would continue to find uses for the humans--but clearly I'm becoming less convinced of that. Perhaps the critical distinction from the historical evolution of technology and economics is that we are no longer talking about replacing someone with a machine that can do some job better, but rather we are replacing them with a machine that can do everything better. I.e., the machines are reaching a point where they are no longer just competing for jobs, but for the very role of being human.

But the flip side of that, I continue to contend, is that efficiency correspondingly rises, and the cost of surviving becomes ever more trivial, and so ever more trivial pursuits become viable jobs. An acquaintance of mine once dressed as a homeless person and sat with his hand out on the streets of Santa Cruz, just to see what it was like. He came home with $300--after one day. There are still lots of horses in the US even though most of them don't do much more to earn their keep than carry someone around for half an hour once or twice a month. Why would anyone hire a human when a robot could do the same thing for less? Why does anyone still ride a horse, or keep a dog or a cat?

As efficiency increases and menial labor migrates to the machines, it may be simply the human touch becomes ever more valuable--to see a live band, to be served by a human waiter, even to have some guy walk up and greet you as the robots fill your gas tank, tell you the local news, ask you how your day is going. For these extra touches, you, the rare skilled human, may pay a little extra, and that little extra will go a long way--perhaps so long that that waiter serving your table only works a few weeks a year, kind of like that horse does now.

But my final argument in favor of the unskilled-laborers continuing to prosper is simply this: Humans from all but the lowest tail of the bell curve have already long proven they are individually capable of producing more than they consume. This alone is sufficient for survival (at current and increasing standards of living) baring a battle over basic resources. Robotco could give a poo about rain and fertile farm lands (any more than it needs to feed the few remaining human employees of Robotco) so it's hard to see how Robotco and its army of cheap labor could do anything but add efficiency to the existing chain of human needs and production. I can see there being an ever widening disparity in wealth between the owners of Robotco and the unskilled, uninvested masses, but that's been the case with every vital technology over the ages from oil to Pepsi--everyone is becoming wealthier with time, just some much faster than others.


[<< | Prev | Index | Next | >>]

Simon Funk / simonfunk@gmail.com