The Internet as the "Geekosystem"

Lev Grossman writes "Is the Internet alive? Of course not, silly. But as this article points out, in some ways it makes sense to study it as a living organism, or an ecosystem, in terms of its growth and structure. "

Comments

...

[Signal 11]

The network admin slashes valiantly at the Router! "Back, back evil filtering fiend!" It shoots packets at him, he hits with 3d6 damage, which is halved because he has a Wand of Guruhood. The netadmin strikes back, forcing the router to reboot.

Not the kind of alive you meant though, right?

Interesting points, bad analogy.

[argentus]

It seems to me that the "Darwinian Theory" that was presented at the conclusion of the article was a bit off... Way back when yahoo was a couple pages with lots of cool links, it was survival of the fittest. Now it's survival of the richest, or those that can spam the most, or target the best ads, or plant the most (irrelevant) keywords getting so much of the attention. Yahoo and friends are exceptions... Hold-overs from the Internet's more open past. Anything created today has a snowballs chance in hell unless they have the capital or engage in sketchy practices.

On the other hand, the idea of 4 clicks of separation is pretty neat, and true in most cases, too, I'd bet. The article's a bit fluffy, but after separating the wheat from the chaff there is some useful/interesting information in it.

There's certainly emergent behavior, though.

[dmorin]

I suppose it all depends on how you define "living thing", because the a-life people will certainly want to discuss it. A flock of birds is not necessarily a living entity unto itself, but it moves in its own way, responds to stimuli, and so on. A "glider" in a game of Conway's Life is really just an optical illusion, it's really one 6 cells that are either on or off independently and just happen to look like a little angle bracket marching across the screen. And then there's a bag of independent molecules all doing their own little job in order to produce......well, us. At what point did we shift from just being the emergent behavior of a bunch of cells into being something that really is alive?

Re:There's certainly emergent behavior, though.

[cyclopes]

This is probably just a mindless rant, but it may be that we are no more than the emergent behaviour of a bunch of cells. After all, "alive" is something that is only defined by the human mind. There is no such physical concept; on the molecular level everything is dead. A piece of metal is made up of atoms, many of them the same as those that make up the human body; what is the difference between the two? If an individual atom is dead, a human who is made entirely of atoms cannot be alive either. So the idea of life is not a purely physical concept. If it were, the definition would be physical, there would be "live" atoms and "dead" atoms. The idea of life is an abstraction which the human mind uses to interpret the physical world.

(Cyclopes rapidly descends into a half drooling state of philosophical abstraction.) If the state of being alive is a construct of the human mind, the internet may be alive. Eventually humans may accept the fact that computers can think and that networks are organisms, at that point they will be living things. Until then computers are just so much silicon and gray plastic.

...

[Signal 11]

Modern man's patterns of what researchers call information foraging turn out to be just as habitual as his ancestors': he follows the scent, hunts in packs and returns to familiar ground as often as possible.

*sniff* *sniff* Do you smell something?
Yeah, I do. What do you think it is?
*peering around the corner* Just a bunch of dead links, keep moving.

Internet Highway... or Highway Internet?

[Raul Acevedo]

I'm not sure who the original author of this is... there's a link to it here.

There it is again. Some fool ranting about the information superhighway. It's nothing like a superhighway. That's a rotten metaphor. But suppose the metaphor ran the other way. Suppose highways were like the Net...

A highway hundreds of lanes wide, most with pitfalls for potholes. Privately-operated bridges and overpasses. No highway patrol. A couple of rent-a-cops with broken whistles on bicycles. Five hundred-member vigilante posses with nuclear weapons. A minimum of 237 ramps on every intersection. No signs. Wanna get to Ensenada? Holler out the window at a passing truck to ask directions. Ad hoc traffic laws. Some lanes would vote to make use by a single-occupant-vehicle a capital offence on Monday through Friday between 7:00 and 9:00. Other lanes would just shoot you without a trial for talking on a car phone.

AOL would be a giant diesel-smoking bus with hundreds on board throwing dead wombats and rotten cabbage at other cars, most of which have been assembled at home from kits. Some are built around 2.5 horsepower lawnmower engines with a top speed of 9mph. Others burn nitro-glycerine and idle at 120.

No registration plates. World War II nose art instead. Terrifying paintings of huge teeth or vampire eagles. Bumper-mounted machine guns. Flip somebody the finger on this highway and get a white phosphorus grenade up your exhaust. Flatbed trucks cruise around with anti-aircraft missile batteries to shoot down the traffic helicopter. Little kids on tricycles with squirtguns filled with hydrochloric acid switch lanes without warning.

No off ramps. None.

Now that's the way to run an interstate highway system.

I much prefer this analogy to all these fancy shmansy theories. :) Either that, or the old "How the Internet is Like a Penis"...
----------

Useful metaphors, not much else

[vlax]

It doesn't surprise me that an information system the size of the Internet would have some unpredicted emergent properties. Stanislaw Lem, in his Summa Technologiae in 1962, predicted that biology would be the main source of engineering inspiration in the 21st century, and clearly this is coming true.

However, don't mistake a metaphor for a truth. They do not propose any kind of unified framework for analysing the 'Net, nor can they. They are simply looking to biology to inspire analytical methods.

Examining usage logs of 120,000 sites, Huberman and Adamic discovered that the distribution of visitors follows a universal power law -- better known as winner-takes-all. This is a world as viciously inequitable as the real one; the most popular 5% of websites get the lion's share -- 75% -- of all Internet traffic.

They missed an important implication of the power law. Increasingly, we should see metasearch systems parasitising the most commonly viewed sites - so long as IP law doesn't prevent it.

I'd like to see some useful predictions come out of there analysis, but I don't see any.

I'm not convinced that disk space restrictions are the major cause of the Darwinian distribution of file lifespans, as the article asserts in the second last paragraph.

Their discussion of an immune system for the web seems pretty speculative, and as they point out elsewhere in the article, monoculture systems are not sufficiently robust. A monoculture immune apparatus (as they propose) probably wouldn't be adequate either.

The point about monoculture is the best one they make. Melissa would have been impossible to propagate, or at least much less damaging, if Windows wasn't so widespread. You would think we had learned this lesson during the Internet Worm fiasco back in the late 80's.

Bail on the word "e-cology." Lem would probably call it "webological analysis", but I think something more greco-latin is in order. Gnostography maybe? Araneastics? Cognostofluxology?

Life vs. self-modifying system

[Cuthalion]

I think you'll find that many of the laws "of biology" tend to really be much more general, and are instead general laws of diverse complicated systems. Biology happens to be an interesting and easy to study niche in this larger field.

Survival of the fittest?

You have a cardboard box with a bunch of things made out of legos in it. Shake the whole box a lot. The ones that doesn't break are what's left over. If that stuff can get reproduced somehow (by itself, or by anything else), "natural" selection happens.

This happens with wholes (organisms & web sites) as well as parts (genes & memes/paradigms) - if the part causes the whole to break, that part won't be very common. We don't see a lot of humans with the "dead" gene.

Nothing comes free, even existence. That's what makes this whole thing work. (in other words, your website is in a cardboard box getting banged against legos)

For internet entities, the cost of existence is bandwidth & server space. Human interest is what it costs to cover these needs. Whether people are interested enough to pay the internet bill because the entity is neat or useful or lucrative is irrelevant.

Existence for humans is normal activity, as well as healing wounds - general metabolism. This cost is paid by an influx of chemical energy (food).

Biological things expend energy getting food, Internet things expend energy getting people interested. If either one of those entities's costs of existance exceeds it's resources, the data pipe will be shut down, so to speak.

Reproduction?

Q: What's the best way to learn HTML?
A: View->Source

In biological systems the notion of parenthood is pretty clear-cut. In memetic systems, however, it can be very difficult to see where ideas come from. But don't tell me that everybody who's implemented a web-based shopping cart thought of the idea themselves.

There are differences, sure. Darwinian vs. Lemarkian evolution.. One or two parents vs dozens or hundreds of 'parents'.

But what's important is that the environment has only limited resources (food, eyeballs), there is some kind of non-exact reproduction (cells divide, ideas get solen), combined with a non-zero cost for existance. Given those constraints, you're pretty much guaranteed to get an ecosystem, or something similar to it.

Is it (the internet) life? I don't care. If it is, great. If it's not, make a new word that means the same thing as "life" without requiring the processes be biological in nature. Good luck getting people to use it.

Cheswick's Maps (links to)

[jbum]

The article didn't appear to supply links to Bill Cheswick's (very interesting) Internet maps, so I thought I'd provide some:

The Internet Mapping Project

Peacock Maps (buy one for xmas!)

-- jbum