Tech as the 7th biological kingdom?

Mon Nov 24 14:06:00 -0800 2008
Last editor Alan manage

Let's see... when it comes to life there's plants, and animals, and fungus, and a few others.. The question has arisen: is "technology" the 7th biological kingdom?

The link for the question is to an article on radar.oreilly.com: a political and for-profit private commercial forum all at once. The well connected and Kevin Kelly, "Senior Maverick" of Wired Magazine, makes the case for technology as life in a video interview there.

Briefly summarizing his points in the interview:

The net had a big impact on the elections. The form and function of web sites and other technology has great impact on the behavior of users. Technology can be regarded as a life form with its own "wants". We (humans) are the sexual organs of technology. A mode of analysis similar to Dawkins' "selfish gene" concept is worth applying to technology.

I hope people appreciate just the news above. I humbly submit one more thing:

I have a reply to both Kelly's reasoning and to its appearance on "radar.oreilly.com". My reply appears to have been censored from "radar.oreilly.com" and so, I offer it below the fold.


Um...

Isn't he really adopting a kind of ethical nihilism here by attributing to "technology" a teleology that denies the role of humans as the ultimate "deciders"? To unpack that a bit:

He talks at length about what technology "wants". For example, he says, technology wants clean water for the manufacture of chips. He says technology wants clean air (for unexplained reasons). And, yes, there's that quip about humans being "the sexual organs of technology".

Well, this is some very old school sophistry -- that word game goes way back. A little green alien landing on the planet 250,000 years ago might wonder, upon first glance at things, whether a web is part of a spider's toolkit for making more spiders or if it isn't the opposite: perhaps a spider is a web's way of making more webs.

Both are true in the most vacuous sense but once our little green alien observes the life cycles more clearly, surely the conclusion will arise that it's more natural to recognize that spiders make webs as part of the process of making more spiders (spider makes web, spider catches food, spider converts energy from food to make more spiders). The spider's web is a passive thing and the main thing it "wants" to do is fall apart: thus we see spiders constantly repairing their webs and making new ones.

There being a cycle there -- spiders and webs locked in an inter-dependent relationship -- we could poetically say that "the universe here 'wants' to make spiders and webs" but that hardly says anything more than "spiders and webs exist".

Technology, like spider webs, seems mostly to "want" to fall apart and stop functioning. A running internal combustion engine, for example, "wants" to keep running for a while while there is fuel coming in but it "wants" to use that fuel up, with wear and tear, then stop, then rust in place. The engine has no tendency -- no "want" -- to drive to a refueling station or, for that matter, to go exploring for oil. The engine, as any mechanic can tell you, wants to be rock.

Rocks don't have sex organs.

He makes a kind of comparison of his way of talking about the "wants" of technology to Dawkins' "selfish gene" concept and yet, if we take the most defensible version of the selfish gene concept, it's nothing like technology. A gene is a partial-control element within a metabolism such that -- and this gets to the key ethical point -- such that the metabolism is constrained by physical law to honor the gene's control. Taking a simple single-celled asexual life form, for example: in certain environments the cell will reproduce. There is no choice. The fundamental laws of physics assure it. It "wants" to reproduce in exactly the same sense that the universe "wants" the cell to be there in the first place: the existence of the cell is real (the universe "wants" precisely "what exists") and an aspect of what exists is an unavoidable, fated reproduction of the cell (in a suitable environment).

Technology is quite different. Its reproduction is a question of choice. We can only start calling the reproduction of technology a natural "want" of technology if we assume that humans have no choice over the matter. And that's why his position is one of ethical nihilism: Why, there's nothing we can do!

One can see how such nihilism would be attractive in the context of trade press like radar: the last thing you want to hear when you are trying to make money on the margins of cheerleading an industry is "all of our techno-niche here is a big mistake and we should kill it." But you are in no danger at all of having to contemplate such a question if you just assert from the beginning, perhaps helped with a little grey-beard sophistry, that, well, the technology is just going to do what it will do because it "wants" to so let's make the best of it. Ethical nihilism.

I found more interesting his earlier comments (prompted by your interview questions) about "setting defaults". He looks at examples like "opt-out vs. opt-in" or the rules that give structure to the editing process of Wikipedia and exclaims some fascination about how those defaults and that structure "nudge behavior" of crowds.

That's not sophistry it's only trite but it does refer to empirical facts. It refers to "what the universe wants". People know this well from time-motion studies, from marketing studies, from studies of the human factors of user interfaces, and on and on.

Yet it's an uncomfortable area to get into here, at "Web 2.0" central, precisely because to the extent we take it seriously it isn't ethical nihilism but ethical problematics. Perhaps it is wrong to not be deeply critical of, say, Wikipedia precisely because of the power imbalances that are reinforced by the structure given to it by its elites. Perhaps spinning flattering yarns about its evolution is wrong because it encourages powerful people to invest huge sums in trying to make more of the same, or similar.

But there, we are no longer talking about the "wants" of technology at all but rather of the wants and relations among people. When we start doing that we have to start recognizing that much of technology is in fact employed as human-on-human weaponry and that the dynamics of its creation, promotion, spread, and acceptance are all-too-human questions.

But if we go down that path we might start hesitating. We might trying to resist and substitute. We might try to extinguish a given technology and replace it on ethical grounds.

So much simpler if we can just suck up to power and explain away their actions by saying "Why, they're powerless! It's just what technology 'wants' to do!"

-t

censorship vs. tech as the 7th biological kingdom?

Mon Nov 24 14:41:23 -0800 2008
manage

Reply hasn't been censored, and has been posted at link as stated.

censorship vs. tech as the 7th biological kingdom?
Mon Nov 24 14:46:47 -0800 2008
manage

While wishing to still honor our apparent "silent gentleman's agreement" to leave each other alone for what remains of the 120 days I hope you will forgive me for replying directly just to say that, yes, the comment has now appeared. After a long delay (e.g., a later comment in another thread appeared first) and after the comment first appeared here (though that doesn't prove cause and effect).

-t

censorship vs. tech as the 7th biological kingdom?
Mon Nov 24 14:51:50 -0800 2008
manage

I broke it first.  Only to note that it showed up though.  Back to being silent (and that was really hard on a couple of posts you made on the weekend & today).

censorship vs. tech as the 7th biological kingdom?
Mon Nov 24 14:56:00 -0800 2008
manage

Was just testing. I'm comfortable with where you've been. I hope likewise. Heads up that the Conway proof I've been referring too and the C. Hitchens take on the civility of theism have been on my mind lately, not least in how they relate to the (no traditional religion involved) problematics of tech industry hegemony (as exemplified by O'Reilly) so be aware that (the bit about John Lennon aside) I'm not trying to poke you (and I thought that poke was gentle enough to see where we are and I'm not upset about the outcome -- and don't feel you'll offend if you want to get in some last digs under that topic, I owe you that. I appreciate the recent change -- thank you.)

-t

Arg! Neither gene nor machines "want"

Mon Nov 24 14:56:24 -0800 2008
manage

Don't anthropomorphize machines - they hate that!

Neither genes nor machines "want" anything: wanting implies self-motivation. My server downstairs doesn't "want" to rebuild its RAID around the failed drive, it does so because it really has no choice in the matter: the code for mdadm says "if dead_drive and spare_available then rebuild_array(with spare)". Likewise, the gene in my personal genome which says my hair shall be 6'2" (unfortunately, I am 6'3") doesn't "want" me to be bald on top, it simply is doing what it does.

Yes, in a way a chicken is an egg's way of making more eggs, but an egg doesn't "want" to make any more eggs. The rooster, on the other hand....

Yes, you can apply some of the concepts of "natural" selection, "mutation", and "reproduction" to technology, and to some extent they make sense. But to take that to the step of saying that Windows "wants" to continue to exist is silly.

Now, one fine day we may actually see self-willed machines, but I would assert that in most cases that is exactly what we DON'T want: I don't want my car saying "Screw you I'm not going out! I don't care if you are out of milk, you should have picked that up on the way home last night. It's snowing, they are salting the roads, and I am NOT going to risk rusting out for your pathetic meatbag self!" Knight Rider be damned - it would be more of a pain in the ass arguing with KITT than just pointing the wheel down the road and hitting the gas.

Indeed, many of the good Dr. Asimov's robot stories were about the lengths to which US Robotics went to PREVENT robots from exercising "free will" and "self motivation" by the constraints of the Three Laws of Robotics. Unfortunately, Asimov knew that any attempt to have simple, rigid laws for behavior would lead to catastrophe - Godel's completeness theorem tells you any set of laws that are complete are going to be inconsistent, and any consistent laws are going to be incomplete.

So, for the foreseeable future, let's not ascribe intention to things without it: LOLCats may "WANT", I may "want", Dr. House may "want", but neither my genome (nor any part thereof) nor my computer "want" anything.

Arg! Neither gene nor machines "want"
Mon Nov 24 15:40:23 -0800 2008
manage

I used to work on a thing called SCATS (Sydney Coordinated Adaptive Traffic System). At one level it just processes data: traffic counts, button presses, configuration changes, etc. But it clearly has wants. They are defined in statistical terms. The system spends its time satisfying its wants. Raising alarms when parts of it need attention. Being tended to by adult humans.

I reckon SCATS was about as intelligent as my newborn son was in March 2002. He had wants too, but they were really programmed by his genes. In SCATS the wants are (or at least were then) hand assembled in PDP 11 machine code. Yes, its an old system. Like DNA it is pretty arcane and difficult to maintain if you don't know what you are doing.

I think there are points where human and machine intelligence connect. I know what my internal experience feels like but I don't know anything about yours. I assume that you are an intelligent human, different from software but that is really just a convention since we are now at the point where you really could be software and I wouldn't know the difference. There was a turing test competition recently where machines started coming out ahead.

I think the question soon will be "what do they want?". Seriously, what do dolphins want? We don't really know because we can't access their internal worlds, so we have to make assumptions based on their behaviour. Compared to machines, dolphins are almost like us.

Arg! Neither gene nor machines "want"
Mon Nov 24 16:31:01 -0800 2008
manage

See, it's hard to talk about because engineer to engineer I would use "want" in the same technical sense you are when you attribute "wants" to SCATS. A less esoteric example of that usage would be to talk about a non-true bicycle wheel, one warped in a particular way, and an engineer saying "If we mount it with this side on the left then the wheel 'wants' to turn left." That would mean that the alignment of physical forces is such that it keeps pulling to the left.

But "wants" isn't the right word to describe how "SCATS" controlled actual traffic. "Wants" is the right word to describe how "SCATS" controlled various digital signals under certain operating conditions. It's very different.

"Life" (in the sense of "the universe") tests the control elements of, say, one of Dawkins' genes by physics -- by the necessary impact on metabolism. SCATS, a software system, has but a necessary impact on digital signals -- not human traffic patterns. It "wants" digital signals (input and output) to satisfy certain constraints. It is "disinterested" beyond that. For example, SCATS software is quite happy to be never run at all. It is also quite happy to be run with false signals. It is also quite happy to produce output with no human meaning or an adverse human meaning. There is nothing *necessary* about SCATS role in our human lives other than our than that which arises from our (seeming) *choice* about whether and how to use SCATS.

You're mis-applying the judgement of "which 'wants' what" in a way that subtracts out the role of human choice -- and thus you wind up in the ethically nihilistic position that Kelly does. Stop theorizing so much about "want" and take its ordinary, informal meaning, if you want to keep ethical clarity.

-t

Do androids dream of electric sheep?

Tue Nov 25 03:28:12 -0800 2008
manage

Tom I know you have seen Blade Runner. You quoted it in a different post. So I am curious to see what your answer is to Philip K Dick's question.

I certainly think they could "dream of electric sheep". If not now, then sooner than we expect.

For me, moral questions should not put our species at the centre. I don't think this means abandoning morals, but it may mean starting with fewer knowns.

Do androids dream of electric sheep?
Tue Nov 25 09:42:59 -0800 2008
manage

Tom I know you have seen Blade Runner. You quoted it in a different post. So I am curious to see what your answer is to Philip K Dick's question.

It's a great movie containing a great allegory about class. Classic noire trick of making the protagonist an anti-hero "investigator" and so we're assured there will be danger and conflict and lack of moral clarity yet we get to follow this guy as he travels across class barriers uncovering information about the overall structure of a heavily segmented society -- it's a screen-writing trick to generate an excuse to "tour" an imagined world. What a great metaphor that the replicants, in many ways morally, intellectually, and physically superior to normal humans are essentially slaves whose lives are owned by a corporation that strives (in vain) to deprive them of developing a full humanity in order to maximize the amount of labor that can be extracted from them. And we see the complicity and the slavery-of-a-different-kind of all of the normal humans who contribute to the production of these slaves. Of particular interest to technocrats should be the pathos of weakness of Sebastian, the cowardly anti-hero to Deckard's hero. Finally, for the necessary "twist of the knife" characteristic of noire, we see the conspiracy of the hyper-rich Dr. Tyrell and the hyper-armed Gaff which serves to (ambiguously but not encouragingly) locate our protagonist's position in the vast web of corruption and decadence his investigation uncovers: he is Gaff's blade, and Gaff's toy.

In the world of Blade Runner, everyone is doomed to unrealized potential and misery and oppression except Gaff and Tyrell, and even Tyrell turns out to be Gaff's pawn, in the end.

To answer your question: Dick's question ("Do androids dream...") is a question about the humanity of an oppressed slave class. It is an excerpt from a scholastic debate on the question: are slaves human? You can imagine a party for elites in Tyrell's pyramid where, over c*cktails, a friendly exchange goes "Well, isn't that the question, really? Are the slaves human?" "Surely they are," might offer the liberal poser, "for example, do they not dream like humans?" "Perhaps they dream but if they do, would they not dream only the dreams of slaves?"

So, Dick's question isn't really about the potential of machines. After all, the replicants in Blade Runner aren't exactly machines, at all, they are altered humans -- they're biologically engineered, not assembled from inert ingredients. In a sense they're bred, not made.


Do I think that machines can develop human consciousness? No. I see the evidence as indicating that human consciousness is thoroughly "embodied" -- it's a unique property of a human body. All the "subtle stuff" of metabolism and of physical presence in the world looks to me like it is inseparable from conscious experience and it is beyond simulation for reasons of complexity if no others.

Do I think that machines can have something that might reasonably be called "non-human consciousness"? Yes. I see the evidence as indicating that consciousness of some form is best understood as a property of the universe itself -- even elementary particles have a parcel of it. But it's also a mistake to over-interpolate that and wrongly anthropomorphize. For example, looking at the telephone on my desk, I don't think that for the phone there is some kind of self-aware "I am a phone" consciousness there. Conscious experience, it seems most likely to me, is the universe's experience of wave function collapse -- it is a set of quantum computations. There is no other physical phenomenon I'm aware of that has a chance to explain it's simultaneously monadic and spacially/temporally-extensive nature. "Qualia" are indivisible and yet "contain" information that isn't localized in space and time -- that describes an entanglement collapsing.

If that hypothesis is right it suggests that computers, as an essential part of their operation, are host to many, tiny, "non-humanly conscious" experiences but this hardly suggests that there is a "computer consciousness" or that we could build a machine with anything even vaguely resembling a human consciousness.

Alas, calling this view a "hypothesis" is a stretch: it's not empirically testable. It's not a scientific hypothesis -- more like an exegesis on the conceptual implications of some well established science.

-t

p.s.: Apparently you can not spell out the word "c*cktails" on Technocrat.

Arg! Neither gene nor machines "want"
Mon Nov 24 15:59:41 -0800 2008
manage

Evolution in biology occured without self-motivation, unless you're willing to attribute self-motivation to bacteria and single-celled organisms.

I believe his use of the term "want" may be closer to "evolutionary pressure", or "tendency".  Self-motivation and conscious free will may come later, whether we desire it or not.

And I would hesitate to anthromoporphize the desires of a technological intelligence as you did with the car analogy.  The outlook on "life" would be vastly different if you could clone your mind in the blink of an eye, or if you could have perfect-state backups.

As far as technology being the 7th kingdom of life, you'd have to give a fairly broad definition of "life".  I don't believe any technology yet fits the definition of "alive", though have no problem allocating it a kingdom if/when that comes to pass.

Arg! Neither gene nor machines "want"
Mon Nov 24 16:33:15 -0800 2008
manage

A sphere at the top of a smooth hill "wants" (in the sense Kelly is using) to roll downhill.

That's a reasonable, if a bit technical use of the term "wants."

But he gets from there to "Technology wants clean water" and I'm trying to help pinpoint the multitude of errors he makes along the way -- and more importantly the social consequences (evil ethical nihilism) that follow from that line of thinking.

-t

"want" - an operational definition

Mon Nov 24 17:42:20 -0800 2008
manage

It seems several people here have different definitions of "want", and so we cannot have a meaningful discussion because we aren't using the same terms.

So let me present a definition of "want" that shows, in my mind, the difference between, say, Firefox as it processes this post vs. me as I process this post.

Firefox follows the rules programmed into it. I hit the "'" key, I get a "'" character. Now let's suppose that for some silly reason the font this text box used didn't define a "'" character. Firefox isn't going to stop, look around, and on it's own initiative change the font we are using. It's not going to pull down a font from the 'Net that has "'" characters. It's not going to stress over the fact that it cannot display a "'" character. It will happily show the standard "no character defined" box and move on, because it doesn't "WANT" to show anything.

I, on the other hand, have wants. If I "want" to get a "'", I'll find a way. I'll change character sets, or find a character that looks like "'", or otherwise seek out some way to meet that want above and beyond what I have been "programmed" to do.

Consider a dog in a fenced-in back yard. That dog "wants" out - it will spend every free second it has to find a way out, digging, jumping, climbing, clawing. If you've never tried to keep a dog in a fence, you don't understand how creative they can be in satisfying that "want".

That's the difference between "want" and "operationg per natural rules. A ball on the top of a fencepost doesn't "want" to fall down - it's not going to try to find a creative way to get down. It will fall, or not, based upon whether it is in a state of (meta-)stable equilibrium or unstable equilibrium.

And that is why I'm not sure I "want" my furnace to "want" to keep the house warm. I don't want my furnace getting "creative" about how to heat the house. If I run out of propane, I want it to stop working, not to decide "Hey, I can keep the house warm if I seize up my blower motor and run it hot."

Self-willed machines, in most cases, are a bad idea.

censorship vs. tech as the 7th biological kingdom?

Tue Nov 25 06:21:34 -0800 2008
manage

In a way, the argument that 'technology wants...' is a mistake in confusing swarm behavior (which may show intelligence) with cognitive behavior (which possesses genuine intelligence).  As you observe, any process has prerequisites.  One could just as easily say that plagues 'want' humans (or other living vectors) to exist so that they (the germs) might exist.

Ultimately, any set of processes could only be considered life if it is the most complex set of causitive processes in its environment; other process sets would be (in part at least) effects, not causes.