Monday, April 30, 2007

An Exercise In Futility? Not Anymore!

Have you watched the movie Pi? Do you remember the personality of the main character? The odd scientist effect and all that? Even if you don't it doesn't matter. This post isn't about him, except maybe tangentially.

Humans have come pretty far down the evolution chain. That is, if you believe in evolution. I don't know what there is to doubt about it, but there are doubters. How can we have emerged from monkeys, and that sort of thing.

Apparently they have recently discovered some sort of worm that is the father and mother of everything. This worm that lived all those thousands of years ago gave birth to two lineages, one that turned into all kinds of insects, and the other that turned into us. What is most fascinating about this worm is the fact that it passed on to its offspring a fully developed central nervous system, which they lost.

What's so fascinating about a nervous worm, centrally or otherwise, you might ask. Well, in a nutshell, worms are usually the sort of slithering little buggers that can do anything with any part of themselves. That is to say, most of their cells are similar if not identical. That's why, for example, you can spend pleasant hours chopping them into small bits and watching each bit crawl away indifferent to its condition.

A central nervous system however implies a sophisticated development, where some cells have been assigned a specific purpose - that of co-ordinating the actions of all other cells. These cells do nothing else but co-ordinate. More importantly, they depend on other cells for their sustenance.

This is a bit of a limb for a worm to go out on, devoting bits of itself to specific purposes. I mean, it couldn't possibly know who is peering intently at it, knife in hand, waiting to cut off which bits of it. Sort of like that spaceship in the Hitchhiker's Guide that lost to a meteor strike that very bit of itself that was designed to detect meteor strikes.

Anyway, if I'm not mistaken, this worm like all others in evolutionary history reproduced asexually. If true, this is something with the complexity of a central nervous system, but that reproduces asexually. In other words, it has the ability to transfer very complex genetic information to its offspring which it can produce by itself. An independent, self-sustaining process. No wives needed.

I'll come back to this in a bit, but first let's go on with the specialisation theme awhile.

Animals of every species exhibit some form of specialisation or the other. Bees divide themselves into those that do all the physical work, those that take care of the eggs, and those that sit around getting fat. Humans are no different, if somewhat more sophisticated about these divisions.

But where humans go one step further is in not only dividing the work and specialising in it but also recruiting other things to help them in their work. It started with animals and has ended up with machines that save effort and time. In essence, machines that we build save us a lot of bother by doing things for us that would amount to hard physical labour of one sort or the other. It doesn't have to be mowing the lawn or ploughing the field, necessarily; even calculating third order differentials of fluid vectors is hard labour.

What this has meant is that, to go on with the specialisation theme, more of us have more roles to play with our brains. Scientists, for example, or even hedge fund managers, do not depend at all on their physical effort to either sustain themselves or be employed anywhere. What they have to sell is their mental effort, which can be done even sitting in a chair all day. Wherever physical effort is needed, either someone else does it for them or machines help out. Coffee machines, for example.

This is still at a rudimentary stage, but it isn't hard to imagine that one day machines could be built that are sophisticated enough to provide for most (if not all) of one's physical needs. Those who make food will have machines that do most of the work, those that make clothes etc. too. Everywhere, basically.

In short, all you will need is your brain. There is even an Asimov short story about these guys who land on a planet and all they see is brains lying around. The inhabitants of that planet have evolved so far that they don't have use for any other body part, just their brains. All other needs are met one way or the other.

Okay we don't have to go so far, but we can allow for the possibility that our physical needs can be met by machines. But who will build them? Those that do will not have their needs met, in our narrow context. If humans build one set of machines that build another that then does all the work, those first machines still have to be built. Here you're faced with that age-old stumbling block of a question: Who will build the builders?

Wouldn't it be nice if we had machines that could build themselves? That would give you all the time you needed to do what you wanted, while your needs could be conveniently met.

But 'machine' is a loaded word. It implies steel and electrical circuits and all of that. It also implies that it functions based on how you programme it. You can teach it to teach itself certain things and take decisions based on certain things, but to some extent you cannot break out of the limits set by programming. That is to say, if you build it then you programme it, and if it has been programmed then it probably won't be able to think for itself. This is a limitation of anything that we build and programme ourselves. It must build itself to be free from programming. So we're not talking about machines, exactly.

Wouldn't it be nice, then, if we could create something that had three characteristics? One, something that is physically strong, to do all the physical work that we so hate doing; two, that has a brain that approaches the sophistication of the human one, at least in its ability to think for itself and not need directions; and three, at the same time can build itself without needing anyone else? In other words, in reverse order, something that would solve the independence issue, the complexity issue, and the, er, looking after us issue too.

Well.

We could possibly do any one of the above three by itself. The problem will lie in solving these three issues simultaneously.

Unless of course there is a way for the human brain to reproduce itself. If it can, then perhaps there is a way to kill the last two problems with one stone. The first of course is a piece of cake. That is, if you get something as complex as the human brain to be able to reproduce itself, then you solve the complexity and the independence problem. Strength is a trifling matter.

At first, this would seem somewhat difficult, wouldn't it. You need reproduction to reproduce, and humans tend to, er, prefer one way of going about it. Test tube babies exist. Human cloning is another possibility. But the problem with test tube babies is that it is still sexual reproduction, and therefore still dependent on pairs of people - a process that is still not truly independent. The problem with cloning is that you will probably get exact copies, which would spell the end for evolutionary changes. Cloning something and yet allowing for random mutations seems somewhat improbable - but I leave that for biologists to overrule.

We are left, then, with a seemingly insurmountable problem. How on earth can something that approaches the complexity of a human brain possibly reproduce asexually? The only things that can are things like viruses or other primitive life-forms. Certainly nothing that is specialised in any way can reproduce asexually, can it?

Enter Earthworm James.

Suddenly a lowly worm may be an eventual answer to a problem we don't yet have.

Suddenly this far-fetched hypothetical evolutionary exercise in artificial intelligence finds a rekindling of hope.

Was it really worth coming back?

Update:
I've just been told by a knowledgeable someone that although worms are hermaphrodites, they do cross fertilise. Not quite asexual reproduction, then. Back to the drawing board, and your five minutes just got wasted. The kindled hope, extinguished.

Ah well. I'll leave this here as a lesson to myself.

Continue reading...

Tuesday, April 24, 2007

How Now

Okay then. Two months away. Heart grown fonder yet?

What's that? Where have I been? Oh, nowhere interesting. Home. Exams and things.

Well, no 'things'. Only exams.

Yes, econometrics. Panel data estimation techniques. There's something, apparently, that is called a differences-in-difference estimate. It didn't make any difference, though. I'll scrape through, just like in the rest. How's that for an estimate.

Anyway, so I'm done. Like, overall done. The pursuit of aimless academic degrees gets sidetracked by corporate culture, lots of money, and a new city. Master Splinter, Master Yoda. Now I am one too.

Master Rakanishu? No, surely not.

Quick update: I am moving my sorry self to Bombay in June. Gainfully employed somewhere or the other. Thinking about renaming the blog, but I don't suppose that will happen.

Differences-in-differences, but indifference of flatulent crapulence?

In deference to long sufferance this itinerant petulant utterance.

Exam hangover. Plis excus.

I'll be back more regularly, now that the ice has been re-broken.

Continue reading...