The Odds Are Good, But the Goods Are Odd

Over at Accelerating Future the conversation has arrived at that inevitable moment, when in the name of sanity I find myself pointing out things like, "Uh, guys, you do realize that all of you are going to die some day, right?" To this I receive responses that declaring humans to be mortal is "begging the question."

(It is of course especially rich to hear this complaint that I am indulging an informal fallacy here, given that almost every textbook that teaches syllogistic logic begins with the example of a categorical syllogism with the major premise "All humans are mortal." -- it goes on to say, "Socrates is human, therefore Socrates is mortal," right? Remember that? Begging the question, apparently, and all because a handful of boys with toys who can't distinguish science from science fiction declares by fiat that this is "in question." Who knew they were so powerful, dictating the terms of logical discourse from their bunker in the Robot Cult compound in the Oort Cloud or in a Sunnyvale basement next to the Wendy's or wherever it is. But I digress.)

I happen to believe that grasping the fact of your own mortality is something like a bare minimum precondition for assuming a position of basic sanity and adult responsibility in the world. For Robot Cultists this seems to amount instead to "defeatism" and rankest irrationality. I expect all that sort of thing, actually. But what I have been rather surprised to hear from them is not just that they don't concede their constitutive finitude (No limits, man! Extreme! sl4, dood! and so on), but that they like to assert things like there is an 85% probability that they are non-mortal -- or a more skeptical hard-nosed realist among the Robot Cultists might assign a mere 15% chance to his personal techno-immortalization. I'm not a scientist like these guys, and so I cannot join in these reindeer games, much is the pity.

But, honestly, Where are these “odds” coming from anyway, that declare anything other than mortality a certainty? Quite apart from the arrant absurdity of these highly unconvincingly "sciency" drag-balls, indulging this sort of speculation doesn’t even begin to take us to questions of the same kind I have already found myself asking about superlative claims about an “artificial” intelligence that doesn’t seem enough like anything else in the world we describe as “intelligence” to warrant the description even in the abstract, quite apart from questions of actual technoscientific implausibility connected with these claims…

Namely, what might it even mean to say of a self so prolonged as to look “immortal” that it was also narratively coherent enough to seem the same life in any meaningful sense? And given the centrality of digitization and roboticization in so many of these scenarios (very much continuous with the dis-embodying of “intelligence” in the “superintelligence” claims) is a life “lived” in a digitized pseudo-body or virtual body “alive” in a relevant sense of the term?

These aren’t actually claims about the plausibility of interestingly different imagined versions of complex future software, when all is said done, but claims about whether we here and now can use language that means something in particular here and now to describe radically different sorts of things that don’t exist, to which claims are attributed despite their non-existence based in part on the things these things mean here and now, but based at other times on refusals of what these things mean here and now?

How does the importation of such terms do the heavy lifting of plausibility production for the wish-fulfillment fantasies of superlative futurology without exhibiting much willingness to pay the price of the ticket in the way of actual consistency with the terms themselves? How coherent are these discourses conceptually speaking, quite apart from their obvious marginality to the science they like to themselves as the consummation of?

And whatever their incoherence on their own terms how does the work of these discourses actually function in the present to enact a kind of brutalization of terms like “intelligence,” and “life,” and “scarcity,” and “freedom” on which we depend to communicate fragile experiences that actually need testimony to remain secure in the world, and all in exchange for selling hyperbolic wish-fulfillment crap that won’t happen and wouldn’t deliver quite what it promises anyway and actually doesn’t even make sense on its own term if it’s submitted to anything like critical scrutiny?

Meanwhile, "Thom" and "Roko" have had something of a spat. Thom declares: This technology should come about tomorrow, if you ask me. It’s not that dangerous as SIAI says and it’s a good chance we will have it before 2020. Dale’s talk is just irrelevant. "Roko" demurs: I think you’re engaging in wishful thinking now. (Only now?!) "Roko" then exhorts me to enter the fray on his side, Dale -- see this? This is an incorrect position! Attack it!!!

Of course, the problem by my lights is that “AI” doesn’t make sense on its own terms, and so handwaving about how dangerous this non-existing “it” “is” or “is not” are equally incoherent efforts, and indulging in angels on pinheads wish-fulfillment discussions of this kind functions, whatever “side” the “experts” assume within them, to evacuate the terms misappropriated for this nonsense of their actual content to the cost of sense and dignity.

I already know in advance that this isn’t going to count with "Thom" or "Roko" as an actual “argument” because I’m not a scientist who can pat them on the heads and make them feel like grown ups. But the whole point is that superlativity is not science, it is a discourse rendering highly selective and superficial appropriations of science into narratives of personal transcendental aspiration. What is unique and definitive in Robot God-talk is happening at the level of rhetoric, not science. Scientists can poke holes in what they’re doing, too, of course, but the substance of their difficulties is happening in a different arena, like it or not.

Soon enough, the spat subsides and they get to the serious business of calculating the Robot God odds again. One writes, "I would prefer not a Friendly Robot God, but a controlled one." Interesting, interesting. Let's convene a conference to consider the matter, gentlemen.

I think I could really get into this after all. When I go to Heaven I would prefer my angelic sex-slaves to look more like Roger Huerta than John Malkovich. Look! I’m a “scientist”!

But it is not to be. They are not fooled. Not "one of us." It's written all over my face, I suppose. "You are not of the body." One sadly opines, "I was hoping [actually employed real scientist and nanotechnologist] Richard Jones might provide an argument as to why SAI is impossible, but he hasn’t said anything ;-("

I'm not making that up, not even the sad-faced emoticon. Hours of laugh out loud funny, these guys are. I, too, am a big fan of Richard Jones, as it happens, but I am personally awaiting his knock down drag out argument as to why leprechauns are impossible at the nanoscale first.

You know, I’ll admit I still can’t for the life of me understand why people won’t just participate in the secular democratic politics of increasing public funding for and ensuring equitable access to the results of well-regulated public education and research and development improving healthcare and widening flourishing lifeway diversity, improving materials and renewable energy provision, working on actual security problems of weapons proliferation and network security and so on. The superlative shifts into robo-immortalization and Friendly Robot Gods are in many cases clearly symptomatic of problems that are better dealt with in therapeutic settings and manage only to sensationalize policy discourse to the cost of sense at the worst possible historical moment for it.