living computation / Dave Ackley


Program, data:living computation :: machines It’s all fuzzy, dammit, relative. And ‘data’ is confusing all by itself, muddling two vastly different things: input and output.

One view sees program as actor, motivator, the source of meaning, intent, and change; data then passive, acted upon, the clay bearing the footprints of acts and intents past.

Recipes certainly sound like programs: add milk to flour… if the toothpick comes out clean, thenfor each layer…. But when we’re flipping through the cookbook, seeing what we could make without going shopping, well then recipes are just data.

The chance of rain reported on the news feels like data, but if we always bring an umbrella when it’s greater than 50%, that number is also kind of a little umbrella-carrying program, and we’re the ‘virtual machine’ it’s running on.

Maybe it’s really programs all the way down, interacting with each other, and temporarily ‘datafying’ each other—or trying to—as they do. Recipes get modified, mixed up with others, combined into fabustrous dinners, sent to friends, and so on.

Living systems are variously programmed and programmable. And people, people are way programmable; people are full of programs; people constantly create and modify them, and combine them in disastulous ways, send them to friends, and so on.

Many paths to new program acquisition for living systems, yes. Evolution rattles the hardware over generations, tweaking manufacturing, BIOS, boot sequence; the individual is shaped by pain and pleasure, conditioning and reinforcement learning, experience and discovery. ‘Higher animals’ have really killer programming technology like the mind-boggling “monkey-see monkey-do.”

Then comes us sneaky humans, of course, with our languageComputer scientists call it ‘natural language’, but it’s a programming language just the same. thing we use, that gives us all new programming storage and transmission options.

We can learn by being told; teach and be taught.

Somewhere along the way, amongst the sandlot games and the chemistry sets, my dad taught me bits of symbolic logic; my brain entrained with paths philosophic and syllogistic.

Lot to learn, at first, just to push the symbols around correctly, to make a proof that actually does get from premises to conclusion without stepping in it somewhere along the way.

Lot to learn, fun. And for the greater good, too, to reduce the total error in the universe. Reason is rational.

Of course, sometimes people aren’t interested in executing the programs they receive. How dare they.

But then there are the times you get programmed despite yourself; despite your best anti-programming defenses.

Late 1970’s: A conversation one evening, sitting across a table for two along the back wall of the Pondicherry restaurant on Boylston in Boston; there, look at mom saying to me:
“Oh, stop being so logical.”

In memory anyway, see me aghast at the very concept, slackjaw, halt, wisp of blue smoke twisting from under my shirt collar—as if, as if, as if it was possible to be too logical, as if logic could be something other than unalloyed good.

Oh, watch me reenergize, reboot, and promptly go down with the ship: So you’d prefer me to be illogical?Valiantly wielding the error I think of as The Flaw of the Excluded Middle and mom lets it go by, mild exasperation perhaps, a shrug over the water glass.

I told myself I’d handled it, pulled it out, won; but, but somewhy that day that moment, the pin had dropped. Possible to be too logical; possible to be too ‘logical’, oh yes. Logic, logician, logical—yet another thing one is never just being, another shield and club.

See tiny new pinprick hole in soft swollen water balloon mind, watch the first fat drop of thick naïve certainty welling up and oozing away, and another and…

Logic still holds me in good stead whenever the footing is firm enough, but eventually so passed Just be logical as the One Master Story, for me.

Good deal all around.

Computers have great potential range, great programmability, but so far have only poor ability to acquire programming sensibly, flexibly yet selectively.

Computers got cheated, got started at the end, at symbolic logic and reasoning, by ambitious parents in a hurry. And like child stars reaching their teens, unable to interact without a script, our computers are elaborate façades only, at a distance alluring, but at the least intimacy screaming their wrongness and incompleteness.

And hmm, the research areas I’ve pushed for computers: Genetic algorithms, neural network learning, artificial life, security and robustness, and now software evolvability. I just keep trying, I guess, to help them fill in a childhood missed.

— § —

22 Jun 2004
[comments (5)]
[permanent link]
<Previous :: Next>


This work is licensed under a
Creative Commons License.
Creative Commons License