Sunday, December 14, 2008

Dangerous Fabricators

I happened to watch this today, only a day after writing this, and the two overlap.

I always get irritated at the beginning of these kinds of talks. It's so easy for people to get worked up about the potential horrors that theoretically await us. Perhaps unsurprisingly, Bill Joy isn't as blindly reactive as many of these kinds of people, but I thought I'd take the moment to give MY opinion on the matter.

The basic concern is that as individuals gain more capability to create higher-technology devices, someone will do something horrible. For example, once we have home printers for printing life forms (virii and bacteria first), what's to stop someone from printing up some new superdisease, causing an epidemic, and killing a billion people? The home printers for printing life forms are not scifi, they're... ten years away. Twenty at the most. Then we'll all die OMG!

Welllllll, the answer isn't easy, mostly because the question presumes things that are false are true. It supposes an incorrect social dynamic. It is in error about the way human minds work. I suppose you could call it a misunderstanding of basic memetics, except that basic memetics doesn't exist. It ALSO misunderstands the nature of science in a fundamental way. These two misunderstandings radically alter the dynamic of this kind of terrorism.

One thing that isn't an answer that I need to address is regulation. Regulation will not work when you're regulating home use. If we can do something in our house, it cannot be regulated without destroying the society you're trying to protect. Once I have a DNA printer, it's impossible to stop me, impossible to catch me, short of omnipresent monitoring from an insanely overpowered government.

Nothing less will accomplish anything, as you can clearly see from the explosion of music, video, and more flat-out illegal things that is saturating modern computers. These things - especially toxic data such as child porn - are controlled by law. Stringently. But they still range somewhere between available and omnipresent, especially if you have access to gray networks (such as Freenet) or can speak multiple languages (to search outside your government's jurisdiction).

Some people are happy to settle for omnipresent monitoring - a total loss of privacy - out of the stark fear that someday some asshole with a DNA printer will kill off half the world's population. Those people, as I mentioned, are operating under at least two false premises: they misunderstand human dynamics and they misunderstand scientific dynamics.

I'll cover the scientific dynamics first.

It's forty years in the future. I'm at home with my printer and, having been spurned by my erobot, I'm planning to destroy the world with a particularly clever little microbe that makes people explode violently. Like in video games.

I release it into the wild. What happens?

Well, if I released it TODAY, man, I'd kill off half the world. It would be aweso... ful. Yeah. Awful.

But I'm not releasing it today, I'm releasing it forty years from now. When the technology has been developed to allow a home user to build this sort of thing.

I can't predict precisely what the defense will be, but I know there will be defenses, because we always develop defenses in line with the technology. Sometimes those are nontechnical (such as the very political defenses against atomic weapons), but the majority of times they evolve as sister technologies or practices and grow to a level where we stop even considering the original technology a threat.

For example, a government worker can easily email billions of dollars of secrets to China and get paid to an anonymous account. Nobody the wiser. But it's not something people run around screaming about.

Because we've evolved defenses against it. Some are technical - email monitoring from secure sites, security clearance requirements, etc - and some are political. I don't really know what those are, but given that I could get around the technical requirements in half an hour, I presume they exist.

I'm going to list a few technical defenses that may evolve simultaneously with home DNA printers, but in honesty I think the biggest defense will be a sister technology allowing for a more... socially advanced?... human. I don't really want to go into it here, but it's pretty clear that this level of technological change fundamentally changes human society.

Anyway, technical. I think these are the most likely:

1) Alarm vaccines. These are intelligent defenses, probably bacteria, that scan against a wide variety of known intruders and actually report on the kinds of microbial intruders you're experiencing. If they find a NEW one, you can see the alarms howl! And, of course, they immediately report the structure and activities of this new form immediately, as well as beginning basic defensive operations. Sort of like a biological Norton, if you want.

2) Airborne scanners. Same idea, but in boxes rather than people. This has the advantage of being easier to update and able to use technology that couldn't exist in people's bodies. However, it would also be less widespread and unable to accurately watch the progress of a new monster against a host. Plus side, they'd catch nonhuman-targeted microbes.

3) DNA Reverse Engineering. Get a sample of the new critter, run it through the reverse engineer, and you've got instant vaccine and cure. Release the hounds within an hour of detecting the thing in the first place.

4) Internal Weather Control. Instead of injecting a drug, we inject a bacteria that can manufacture the drugs we want. Moreover, the bacteria can be recalibrated if it turns out that cocktail isn't working.

Now these all sound hopelessly farfetched and futuristic, but you have to remember that we're talking about fighting off a microbe that I created on my home machine. They're not really more futuristic than that.

The point is that by the time we have home machines for this, we'll have developed defenses and responses to the threats the home machines create.

You can argue this isn't true - that we don't have defenses against nuclear bombs, for example - but those can't be created at home. Computers and the internet were once predicted to have catastrophic consequences for humanity back in the eighties, if you recall. "A hacker brought down wall street! OMG!"

But that passed because we developed safeguards, built them right into the infrastructure of the system. Built them right in because we WERE the system. The home users actually enabled the development of the defenses we use against the many horrors of the internet.

Same idea. Technology doesn't advance on its own. It advances for a large number of people and in tandem with a large number of brother and sister technologies.


The OTHER issue here is a fundamental misunderstanding of humans. And this is perhaps a more serious misunderstanding.


Nearly every crime ON THE PLANET is perpetrated by someone who spends a great deal of time near the victim. Likely it's a member of their family. Kidnapped children? It's usually a family member. Abuse? It's usually a family member or a trusted friend. Theft? It's often someone you know, the rest of the time it's usually someone who lives relatively nearby. Mugging? Usually muggers stay pretty close to home or their primary hang out, believe it or not.

Believe it or not, humans shit where we eat. Criminals commit most of their crimes within spitting distance of their normal zone of activity.

There are exceptions. For example, a lot of people knock over convenience stores. But it's usually a convenience store they've been in before or at least driven by a bunch. The convenience store makes a particularly appetizing target, so I guess it's understandable.

"But that doesn't hold true for terrorists and serial murderers... does it?"

Yyyyeah, it does. To a large extent.

It seems the majority of terrorist attacks happen in the same city as the terrorist lives. Nearly all of the rest happen in the same nation. You'll notice that most nations that suffer terrorist attacks have a higher percentage of immigrants from so-called "at risk cultures".

Our very own terrorist attack was a black swan. A particularly aggressive gambit using a particularly appetizing method against a particularly appetizing target.

I need to be clear. I'm not arguing against immigration or whatever else your personal concerns are. I'm explaining that terrorists do not simply pop over on a boat and blow you up. They blow themselves up near where they live, in the same way that muggers tend to mug people on the same set of streets, often within walking distance (or bus distance) of their home.

The same is true for our theoretical disease-criminal of the future. He is operating not out of coordinated misanthropy, but out of greed or UNcoordinated misanthropy. He won't cause any more damage that criminals today do. In fact, he'll probably cause less, because we'll be operating under augmented reality, which will make us very community-driven organisms. But let's assume the same level of misanthropy as today. Not a big threat.

"How can you be so sure? Isn't in DANGEROUS?!?!?!"

Well, hell, I can put together a mean chemical cocktail out of the crap under my sink and kill seventy people on a crammed-full red line train. Anyone can. But it hasn't happened, because that's not how humans work. We just don't work like that.

There are some amazing exceptions, most of them American. The unabomber, for example. But to call these uncommon is an understatement. They are so rare that we still remember him, even though he only killed three people. THREE PEOPLE! Similarly, some of you remember the Japanese guy who attempted pretty much what I described in the last paragraph. But more people are harmed in bowling-related accidents than these kind of coordinated criminal misanthropy. Perhaps it's because anyone whose brain is working so poorly as to WANT to do these things can't possibly do them WELL. I'd like to think that if I went off to kill people, I'd do it quite a lot better... but maybe there's something inherent in the act, some kind of mental screw-up? Perhaps there's some other reason. But... these people seem incapable of doing massive harm, even though they already have the tools to do so.

There are exceptions. Black-swan events like 9/11. That's what the technical defenses are for.

Also, there are some memetic factors worth considering. For example, suicide rates go up when there's lots of news coverage of suicides. School shootings went up when the media started going nuts about school shootings. (In fact, they basically didn't exist until the media went nuts.)

These are worth considering, but it's not something that we can stop progress for.

In fact, you can't stop progress at all.

Even if I haven't convinced you, in twenty years you'll find me sipping lemonade as I genetically engineer a green puppy.


Brog said...

What about computer viruses though? Sure, these aren't capable of causing serious damage these days because of the technical defenses, but the minor damage they cause to the computer illiterate is still significant. And the people who write these are therefore capable of this kind of damage.
Translating this into the biological sphere is unpleasant. People who crash aren't so easily rebooted or reformated. Not everyone keeps their immune system up to date, or is careful about running strange genes they recieve. Even people who know what they're doing still might need to filter out a few penis-enlargement viruses every week. And then there's the zombie farms..

(I'm not entirely serious, just having some fun with idle speculation.)

Craig Perko said...

Computer viruses are not much of a problem compared to what people thought they would be. And they're actually a good example: nearly everyone in America has the capability to create computer viruses... but we don't. They're remarkably rare for the fact that the entire world's supply is enough to mildly inconvenience us.

Greg Tannahill said...

Individual human beings already have the ability to kill off very large numbers of our fellow human beings using such things as large-scale explosives, deliberately started fires, poisons introduced into food and water supplies.

These things just don't happen that often. They require the unique combination of not just the will to kill many humans, but the sustaining of that will over a prolonged period of planning, and a degree of intelligence, dedication and follow-through which rarely corresponds to the profiles most likely to commit these acts. And it needs to happen without the individual engaging in the early warning behaviour of petty crime and mental illness which would make them the subject of close scrutiny by authority.

The real danger is when the possibilty for large-scale destruction lines up with the ability to cause it on a whim. That's the only bracket of danger I think we need to be discussing in this context, really.

Craig Perko said...

I agree. It should be noted that, technically speaking, the capacity to create computer viruses and explosive devices is available to everyone, but it doesn't happen often because the knowledge takes some effort to learn and the act some time to do.

I don't realistically see creating real viruses to ever be easier than creating computer viruses!