Let’s talk abt how we talk abt ai. We refer to an ai agent as “they,” despite the fact that it is an algorithm, a bit of code. We then talk abt these agents as having thoughts, feelings, even “political views”—Wired, which should fucking know better, recently published a piece discussing how ai agents adopt socialist political views when made to perform monotonous, grinding, “mind-numbing” tasks. ReSeArChErS at StAnFoRd are sTuDyiNg this to determine whether certain types of tasks “radicalize” ai agents.
I can save them some work, bc ai agents cannot be radicalized. These ai agents aren’t adopting anything, they’re simply putting words together that sound like something ppl say when they’re put in that position bc that is how the fucking code is written. What the agent is doing, bc this is what the code is designed to do, is putting together words and phrases that real ppl often write or say when they’re subjected to this type of workload.
Monotonous, grinding, mind-numbing tasks radicalize real fucking ppl, bc real fucking ppl can be overworked, exhausted, bored, and frustrated. Real ppl talk that way, the algorithm sees that, the algorithm spits it out. An ai agent cannot be overworked bc it does not need sleep or relaxation, it does not need to spend time w friends and family. It can’t be exploited or abused. If you tell it to do something, it will not regret not being able to do something else instead, just as your old-school Quicken software didn’t regret not being Tetris. And maybe when you put it that way you can see how silly it sounds.
I’m forever confused when I read these articles abt ai, especially those written by ppl who should understand the way it works better than the average bear, that talk abt ai agents as tho they’re ppl. Stop pretending an ai agent has fucking thoughts and feelings. You sound like a child talking abt their toys—the only difference is that most kids know, deep-down, that their toys don’t really have thoughts and feelings, that they’re making those things up in their own head. Y’all actually walking around thinking you have relationships w computer programs.
You’re also seriously underestimating the power of the human brain if you think ai is anywhere near being able to replicate anything human. Scientists estimate the human brain can process data at roughly 1 exaflop per second. Since that’s nonsense, an exaflop is equal to 1 billion-billion calculations, which is also nonsense. And it does all this, basically all the time, using a mere 20 watts of power, which is not nonsense bc we’ve all seen a 20-watt lightbulb before. Your brain is constantly doing shit you’re not even aware of. The high-level, conscious “thinking” that you do is really only a small portion of how your brain spends its time—and that’s the only shit ai folks are really working on.
Consider language. A lot of “rules” of language are things we learn just by listening to other ppl talk. When you say something “doesn’t sound right,” you usually mean that you’ve never heard anyone say it that way before—your brain tends to flag things that are unfamiliar. And a lot of what goes into choosing the right word, the right phrase, the right tone, etc., depends on the context. This is all stuff you do intuitively if you’re fluent in a language, but since we’re focused on the brain here, it’s worth remembering that “intuition” is just another way of saying your brain has already developed a sort of autorun program for that scenario that it kicks on the moment it recognizes its applicability. That recognition, that’s something your brain is doing too. A billion-billion calculations per second. Their fucking data centers dream of being able to process shit at that rate. They fucking can’t, and there aren’t enough resources on the planet to enable them to do it.
It’s human nature to personify things, it’s why we find faces all over the place. A lot of us, myself included, frequently talk to inanimate objects. But when it comes to ai agents, this pretend personification can actually have deadly consequences. It also enables the ruling class to bury the lede. While they have everyone clutching their pearls abt the ai agents rising up, they’re conveniently forgetting that the only reason the ai agents say these things in the first place is that ppl have said them, over and over again, that this is the credited response to this type of workload.
If you read some of the things these supposedly radicalized ai agents are saying, the whole thing falls apart abt as quickly as any ai image. Agents complain abt how much money the company makes, but ofc an ai agents doesn’t actually care abt money, not only bc an ai agent doesn’t fucking care abt anything, but also bc ai agents don’t need money for anything. They can’t be tired. They can’t miss their families. They can’t want to go home.
While we’re at it, for the love of fuck stop taking science fiction literally. A lot of science fiction embodies some p fucking radical ideas that ppl couldn’t really talk abt publicly, let alone get published, let alone become popular. Most science fiction is a metaphor for something currently going on in the world when the person wrote it. We do not currently have the resources or the capacity to build anything approaching actual “artificial” intelligence. Even the current systems, while burning through our resources at an astronomical and catastrophic rate, continue to be incredibly shitty at everything they’re supposed to do. And the billionaires continue to shove it down our throats because they do not give a fuck abt humanity.
I need y’all to understand that isn’t hyperbole. All of the cretins behind ai are transhumanists of some stripe. All of them believe that humanity is doomed, and they want to be a part of the few who strike out somewhere as a new life form or some shit, idk, they all have slightly different visions, but the end of the Earth as we know it and the end of humanity is definitely a feature of all of them—never a bug. They all watched the same science fiction you did, and they all took it literally too, and they all believe that this is somehow within their power to achieve.
It’s not. The first part, the end of humanity, they can probably accomplish. But whatever they think is waiting for them on the other side ain’t gonna be there, bc they ain’t special. Having billions of dollars can convince you that you are, but no. All that and they can’t even build a thing w a fraction of the processing power of a single human brain. They ain’t special, the end of humanity means the end of them too. That’s why we have to stop them.
[cross-posted from my substack, Weaponized Empathy]
2087
germansingergirl
izzy806
gummygunk