Mx. Chara Aznable (they/them) of PnictogenMar 8

There's a heap of reasons I have practically no temptation to use an #LLM or "generative #AI" (i.e. fake AI) for any purpose. The insult to intelligence as a concept, the appropriation of artificial intelligence (a genuine interest of mine! for decades!) and its corruption into a mere buzzword and slang term for stochastic generation of an endless stream of trash in a mocking mimicry of human creativity, is one of the big reasons.

The elites of the #technology sector, the corporate executives and managers and the snobbish #programming nerds hoping to get rich quick through some computer tricks, don't just seem to have a low and bigoted opinion of humanity in general. They also seem to have an abysmal opinion about what #intelligence and creativity are.

Show threadMx. Chara Aznable (they/them) of Pnictogen

I'm not even getting into the #tech geeks' unhealthful and racist fascination with crap like intelligence testing, IQ and multiple-choice test scores, and the fallacy that genius (or stupidity) are solely determined through #genetics and have nothing to do with learning or nurture. Yes, these clowns tend to act sometimes like "intelligence" is little more than a number or a test score, really a kind of social credit rating that's always imagined as being stamped indelibly on people's genes.

I'm more interested right now in talking about how these #computer geeks seem to regard the exercise of intelligence. What does it mean to them, on a day to day basis, to be "smart"?

Again and again it seems like it's typical for the elites of #technology (and of #business in general) to think that there's nothing smarter in the world than breaking rules and getting away with it. If you're really good at cheating people that's proof of intelligence, and if you're the victim of cheating then you're just "dumb" and "moronic" for not anticipating it—or for refusing to cheat oneself.

Mar 8 at 10:29pmWeb
Show threadMx. Chara Aznable (they/them) of PnictogenMar 8

This speaks to the general techbro attitude towards #rationality as meaning something very different from what I think of as being rational. To me, one is rational if one thinks and speaks in a reasoned way and is able to defend one's statements and actions rationally without resorting to fallacies or contradictions.

But to the #technology elite crowd, to be "rational" is mostly about maximizing one's rewards. It would be "irrational" not to take advantage of some obvious and facile means of cheating large numbers of people out of money, for example, because it's "rational" to maximize one's personal hoard of money and conversely it would be disadvantageous and "irrational" not to grab such an opportunity, and thus be the loser of the game of life—the loser of #evolution itself, according to the particular way in which techbros view evolution.

Show threadMx. Chara Aznable (they/them) of PnictogenMar 8

I can think of these people as "irrational" myself, motivated clearly by emotional and fallacious notions about #science, but here's the paradox about #rationality: one's reason must always be based, ultimately, on a set of premises which one is forced to take more or less on faith, as the starting-point to one's thinking. Consider the simple example of Euclid's parallel postulate, for example: one cannot prove the parallel postulate in terms of the previous axioms in Euclid's Elements, so one is actually free to replace the postulate with another and thus obtain a consistent but "non-Euclidean" system of geometry.

I have my own set of essentially unprovable notions about how the Cosmos works and how people ought to behave ethically, and the techbros are free to reject ALL those premises. I can regard them as "irrational" but that's according to my premises, not theirs. Conversely they're free to look at my own presumptions as ridiculous ones, and prioritize their own.

Show threadMx. Chara Aznable (they/them) of PnictogenMar 8

Plainly their utilitarian approach to #rationality informs how they view #intelligence and their fixation on "generative AI" as the summa qua non of computing technology: it would be irrational by their lights to reject the use of a computing device that's:

(a) such a conceptually simple and brute-force method for approximating human behavior and creativity, requiring very little from the #tech entrepreneurs besides extravagant hoarding of computer hardware and #data

(b) so readily tailored to produce output within a desired specification, merely by tuning the training data fed into the #LLM or other "generative AI" thingummy

(c) so easily slotted into #business models precisely because the chief virtues of #LLMs and similar devices is their predictability of output and their brute-force ability to generate massive quantities of output.

To them it's a ridiculously obvious proposition: rent a building, pack it with GPUs, steal a mass of training data from somewhere and hey presto it's like printing #money. Who would be stupid and "dumb" enough NOT to do these things?!

Show threadMx. Chara Aznable (they/them) of PnictogenMar 8

Notice that NOTHING about what I've just written about the manifest advantages of "generative #AI" to the canny #technology entrepreneur or startup huckster has anything to do with whether the #LLM or other "generative" device is actually capable of thinking or reasoning. That's simply not even a relevant issue to these people, not really, even if making outrageous claims about the boundless and godlike #intelligence of their machinery is central to their marketing (and, perhaps, also central to their fondest mystical hopes about the future of technology.)

The high-tech huckster who wants to succeed in #business tends to bring this axiom to the table: they have faith that once some particular business model or form of technology is flooding the market, by force if necessary (e.g. legal coercion and "regulatory capture") then common humanity will simply accept it as a new fact of life and complaints will die away to a negligible fringe. "Stop whingeing, #LLMs are here to stay, like cars replacing horses," etc.

I happen to reject that premise, outright. I say that it should be possible and indeed desirable for humanity to rise up against some established form of technology or entrenched business model—reversing what the techbros think of as inevitable progress.

Show threadMx. Chara Aznable (they/them) of PnictogenMar 8

Another aspect to the utilitarian, rewards-driven way in which #tech geeks regard intelligence is that they have great faith in emergence, and optimizing things through sheer repetition and stochastic methods. Surely this fits in with their simplistic notion of biological #evolution as necessarily producing optimal organisms and structures by virtue of sufficiently prolonged competition.

They seem to forget that "random walk" algorithms are apt to converge upon local (not necessarily global) minima and maxima, and therefore one can't be assured that a sufficiently lengthy brute-force stochastic process will converge upon an optimal answer. Instead they seem to think, incorrectly, that the results of such a stochastic process must necessarily always be improving—even improving exponentially. After all, aren't they making exponentially growing amounts of #money by doing things in this way? (Or at least, exponentially accruing debts.)

Show threadMx. Chara Aznable (they/them) of PnictogenMar 8

And thus the world was treated a while back to Sam Altman, currently the most visible booster of "generative #AI", still regarded by tech #journalism as though his lies about #OpenAI and his empty promises for the future were automatically newsworthy, responding to the accusation that #LLMs were merely "stochastic parrots" by saying something like, "aren't we all stochastic parrots? lol" on Twitter.

There's a bizarre backhanded form of humility at work, a doublethinkful willingness from elitist techbros to acknowledge—but only in a "joking" sneakfaced way—that really, despite all their racist posturing about their own purportedly genius intelligence and superior genes, these people aren't very smart at all. They are people who operate in public life by extremely simple-minded rules: hoard money, chase popularity, and do whatever is necessary no matter how criminal or destructive in order to preserve their social position and maintain their grip on publicity and political influence.

They do whatever works, and they copy whatever they need to from people who are better thinkers—they are stochastic parrots, furiously pulling the same few tricks over and over again in a kind of random-walk algorithm that they hope will converge upon success.

Show threadMx. Chara Aznable (they/them) of PnictogenMar 8

Sam Altman is easily one of the most witless and bubble-headed lords of #technology and #computing I've ever run into, capable of making even #ElonMusk seem clever and inventive by comparison—but one must always be a bit careful about judging such people solely by their public behavior. Altman's probably ruthless and clever enough in private, when he needs to be. In public, however, Altman (much like Donald Trump) simply does not NEED to be smart, so long as he's still able to publicize himself as smart.

And that's roughly the same for the #LLM and "generative #AI". These things do not actually need to exercise any genuine intelligence, so long as they're consistently reported and chattered about as if they were intelligent.

The "intelligence" of these things, in the end, is simply a matter of persistence. Can #LLMs remain the talk of the town? Can they benefit indefinitely from gullible and starry-eyed reportage from @arstechnica and other tech journalism outlets which have been hopelessly corrupted by money and profit-seeking and vacuous techno-optimism? Then LLMs will continue to reckoned as the smartest things in the world.

Show threadMx. Chara Aznable (they/them) of PnictogenMar 8

All that matters to this crowd is results, or rather the ability to keep up a permanent illusion of producing results. Ultimately their "metric" for intelligence, as with everything else, is whether there's still big money to be made.

And nobody important really wants to think too hard about what happens if that particular illusion were to collapse. Global #capitalism has grown dependent upon maintaining the fiction that "wealth" can be accrued infinitely and exponentially, which means that it's now a social imperative to keep #technology speculation afloat whether or not the technology itself is sound.

Why, it would be irrational to behave otherwise! Question #OpenAI or Sam Altman too hard about what they're actually doing and the whole #economy could collapse! What sort of "moron" would want that to happen? Keeping the "generative #AI" bubble inflated...why, that's the only #intelligent thing to do.

~Chara of Pnictogen