Slashdot Mirror
The Los Alamos Bug
Kannappan writes "'You somehow have to forget everything you know about life', says Steen Rasmussen, a colleague of Norman Packard. Packard and his team are working on creating life artificially, nicknamed The Los Alamos bug (pdf). It will be created out of a molecule called Peptide Nucleic Acid(PNA), with a blend of three different factors crucial to life, viz. containment, heredity and metabolism. The researchers believe that the synthetic lives so created will have an enormous practical value in producing clean fuels, healing injured bodies and acting as tiny diagnosticians roaming our bodies."
We are certainly moving towards artificial intelligence. We actually have programs that can write themselves to a limited degree. And so we'll probably have artifical intelligence shortly. But making a new biological species with our level of intelligence sounds tough because of our limited knowledge of the brain.
I'd recommend you read things by Ray Kurzweil on this topic. In particular, "The Singularity" seems relevant. Apparently there is a short collection of essays by him online, but I don't know if it'll have what you're looking for.
Some people think that before DNA evolved, everything was done with RNA. Both hereditary information and the physical catalysts. Like proteins, RNA molecules can fold up into odd shapes and perform catalytic reactions. The only difference is that Protein based system work faster. The Ribosome, which converts RNA into Proteins is actually made from RNA, rather then proteins, and is almost exactly the same in all life.
autopr0n is like, down and stuff.
Imagine a bug that can convert cellulose to alcohol. Or eats dioxins and destroys them. Or generates hydrogen from sun and water. Wouldn't these be somewhat beneficial?
If you don't know where you are going, you will wind up somewhere else.
can we create *intelligence* from scratch?
I think I fundamentally disagree with you in saying that intelligence is hard to create, given life in the first place. At this point in time, science has (almost) undisputedly proved the theory of common descent. I pretty well believe that humans eventually came from single-celled organisms, and so does most of the world.
So assuming that is true, intelligence more or less created itself, through life, by a glorified trial-and-error system. Although it seems surprising at first, if you consider how many many different orananisms there are (were) at any given time, and how many trials (generations) there have been, it becomes much more down to Earth.
Actually I think even 'intelligence' today is still a glorified system of trial and error. Think about solving an elementary algebra problem. What's your first intuition (or was when you were learning)? Isolate the variable, etc? Hell no! Trial and error. It's intuitive and doesn't take much mental 'work'. Example: Mary and Sue have a combined age of 15. Mary is 5 years older than Sue. How old is Mary?
Spit this problem at an average 5th/6th grader and I promise you won't get anything along the lines of x + (x + 5) = 15. You'll just get 3 + 8 = [crossed out], 4 + 9 = [crossed out], 5 + 10 = 15 !! And that's how the problem is solved by a (we'll say) 10 year old.
Now, I know I don't seem to be really getting at anything big, but consider this: the average 10 year old has solved a LOT of 'problems' in his/her lifetime, from how to balance to stand up, how much food to eat so you aren't hungry anymore but don't throw up... I could go on forever, but I will call one example: pouring.
Is it hard to pour water from a pitcher into a cup? I'm pretty sure most of you have figured out how to do this reasonably well by now. To do this problem systematically is EXTREMELY difficult. I'll simplify the problem slightly and boil the problem down to two varibles: The height of water in the cup (we'll say % full), and the tilt on the pitcher (an angle between 0 and 180). There is ABSOLUTELY no simple, one-line algebraic equation to solve this one. You can't simply say, when the cup is 100% full, put the angle to zero. You have to correct for how much water is out of the pitcher already and is about to fall into the cup (a time delay), and also the time it takes to move the pitcher from say, 20 degrees to 0 degrees (more time delay). Even better, the flow of the water within the pitcher depends not only on the angular position (zeroth derivative), and the rate and acceleration (first and second derivatives), but also the "jerk" of the pitcher (third derivative of angular position). Wow. That's hard.
To solve this problem analytically, you would need a lot of math. A LOT. In fact, even more than we know today. Using LaPlace transforms and 3rd order differential equation solvers, this can be done, but even the DE solvers are written in trial-and-error form to some extent. If you've read this far, you're probably asking: What exactly am I getting at?
YOU ALREADY SOLVED THIS PROBLEM! Ever fill up a cup and not spill? Not bad. Basically, your mind (body?) has already found at least some solution to this problem without you knowing it. You have subconciously short-circuited hundreds of PhD's worth of math with a magic black-box of trial and error. Remember when you were a kid? You tilt the pitcher little and tilt it back. Not enough. You do it again. Not enough. You tilt the pitcher until the cup is full. Crap. Spilled it. Note to self: stop before the cup is full.
So there you have it. Our 'intelligence' has solved math problems than most college graduates could do (even with Maple) to save their lives. If it works, do it again, and if it doesn't work, do something different. That's all our 'intelligence' is.
I really don't think this whole 'intelligence' thing is a very novel concept at all.
Partial Credit: The Engineer's Best friend
"Well, the bridge didn't fall all the way down!"
Run that past me once again in English?
OK. "They're fscking cheating"
They're using PNA because it does fancy stuff "on its own", just because the out of it is soluble in oil, but the inside of it is repelled by oil and prefers water. So it goes up and down according to whether it's "single-stranded" or "two-stranded" (i.e. whether the inside is expopsed or not). You don't need the complex machinery of metabolic reactions which is necessary for "real" life to cut, assemble and move stuff around.
The whole thing is a fraud, at least if TFA from the New Scientist is an accurate description. Never mind that the genome is essentially random bits of PNA that don't code for any chemical machinery. TFA says that it does influence "metabolism" directly, through electromechanical influence. Wow, that leaves a lot of degrees of freedom for evolution to play with, doesn't it ? (Hint: no, it doesn't). I could mention the utter lack of self-regulation (that thing just grows and divides when it's too big, period), removing the essential computational component of life (wonder what Packard's friend Stuart Kauffman would say about that).
The worst part is the thermodynamics. Apparently all the reactions that occur within the bug are "downwards", degrading reactions. The bug doesn't relly "build" anything. The miracle of life lies precisely in its self-constructing aspect: life is able to couple downwards, energy-releasing reactions and upwards, constructive reactions so that the former "feed" the latter. Thus living systems really construct themselves. That "bug" just uses hand-tailored, pre-activated, energy-packed components which are fed to it by the experimenter and degrades them according to a carefully hand-defined pathway. Evolution of the inner processes is utterly impossible because, essentially, there is no real "inner process". It's just like fire - a downwards, energy releasing reaction without any self-regulation. .
If this thing is alive, then so were Sydney Fox' "protocells" from 40 years ago !
That thing is about as relevant to understanding life as Deep Blue was to understanding intelligence - i.e. it gives a good example of what life is not.
Thomas.