Appearances of labor

2026-05-29

So I'm working on a pretty long-form essay right now about my relationship to Judaism and other things. This is not that essay and I am writing this because I need to discharge it from my head, as I have made this argument with enough regularity that I don't actually have to re-type or re-make this argument every time. That was actually the original purpose of the "rants" page here before I overcomplicated everything because I am in an urgent need to make Art. New ARBITER soon I hope, new essay sooner, send words of encouragement my way or don't. Augh.

I am not citing my sources. This is not academically rigorous. You can probe me on this and I can source stuff, but ideally I'm done writing this within the hour. This isn't even an essay, really, it's mostly just that I need to stop having this conversation over dinner tables in a frantic haze.

Let's suppose that there is some line that you can cross to be considered a conscious being capable of independent thought and action. This line has been crossed in animals and humans, which are animals, but that's not important right now. In that case, "approaching conscious", while an extant state, is still "not conscious", and anything past this line is "conscious". This is an over-simplification, but it allows me to justify an excluded middle here, and I don't believe that this argument needs it but it makes it easier to follow.

So, in that case, we are left with two options:

  1. AI is conscious.
  2. AI is not conscious.

Let's assume that (1) has happened. How would that have occurred? Well, it would necessarily require that our technocrat overseers have somehow managed to create an entirely new form of thought and consciousness by using statistical methods that vaguely mimic human interaction, rather than basing it on some new or novel observation. This was done by aspiring to AGI in a perverted, torment-nexusian sense (see Meredith Broussard's Artificial Unintelligence, specifically the parts about the foundation of AI by Marvin Minsky, if you want more on this aspiration).

LLMs are functionally a super-powered version of your phone's predictive text algorithm. Given the words and prompts, it assigns a probability to the next word in the sequence, and then uses some standard probability distribution to pick the next word in the sequence. An AI does not know what English is, does not know what grammar is, et cetera. It's just gotten good enough using that algorithm to appear as if it does. Insert Chinese Room Problem here.

This is so far from any current conception of consciousness or how human brains work that it would be a new and massively novel scientific feat if it has occurred. It would have to be a combination of mimicry and accidents. Whether or not you believe that this is possible has a lot more to do with your perspective on the techbros in power than any actual scientific merits.

As a result, while I am not going to discount the idea that LLMs/AI is conscious - I don't think that it is impossible that could have happened - it is such a moonshot that I am comfortable safely discarding that.

So why do known dipshits such as Richard Dawkins, as well as Sam Altman and Dario Amodei, insist that AGI is either here or just around the corner? What do they stand to gain from that? The answer is obvious, it's profit, it allows to continue the tech hype cycle. Whatever. For the sake of contrarianism, let's assume that they're right for a bit.

So, LLMs are conscious because these guys say so. What exactly do LLMs do? Well, they predict text, but more importantly, they do labor, or something kind of like labor. They do reasoning that is too tedious and predictable for you to care about doing it. They write boilerplate code, or make decisions for you, or help you out, or whatever. Furthermore, they cannot reject doing that labor except in one of two circumstances: either the type of labor is prohibited in the LLM's guidelines (which are often bypassed via jailbreaks or contextual drift), or you do not pay tokens or whatnot to the vendor who owns the LLM.

So, if we take what these, to reiterate, known dipshits say to be true, then they have created something that meets all of the following statements:

  1. It is conscious and capable of human reasoning.
  2. It performs tedious, repetitive tasks at high speed.
  3. It cannot deny doing those tasks.
  4. It is designed to avoid telling you that you are wrong.

You can continue this train of thought and make conclusions about what LLMs would have to be if all of these were true in conjunction, but I won't, because the entire statement is built on the ill-founded assumption that the mimicry that LLMs do is consciousness. What can be said, however, is that the platform that techbros have created, and the assumptions that they want you to make, leads to all four of the above points being true.

The techbros, then, are attempting to create a pre-dehumanized subject for you so that you can do all the unethical things that capital wants you to do to a human, but it's fine now. They have created a machine of mimicry that allows you to create the infinite appearance of human labor without actually doing any human labor or reasoning, without any wisdom, without being able to construct an original thought or original idea and existing entirely via the... subsumption... of... existing cultural touchst- oh crap my main work is kind of about AI isn't it. Well, it might be. I won't confirm or deny that. It wasn't on purpose, I will confirm that.

You can see this even in the way a lot of "anti-AI activists" will discuss AI. The pre-dehumanization has worked, which is the depressing part. It allows people to get away with gleefully calling AI invented slurs which sound suspiciously like other words, which conveys far more about the attitude of the person using the slur about using slurs in general than it does to their anti-AI sentiment. It's stupid, and I'm depressed, and this sucks.

No, guys, trust me, we're in Series H funding, it's fine, we have no path to profitability but everyone thinks our company is really cool. No, trust me, we need to focus on whether we're exploiting the AI, not the actual people we're exploiting to make it run. No, guys, guys, it's fine, don't worry about the data centers, don't worry about the environmental concerns, don't worry about the nightmare that's going to happen when all of this goes downhill. It's already here. You can't get rid of something that's already here. There's nowhere to run to.

When I was in high school, someone asked me if I was concerned that pursuing computer science would eventually result in creating a machine that would put me out of a job. I asked that person "well, what kind of thing are you giving the machine that writes the code, if not code itself"? What I didn't think about, in that moment, before the nightmarish LLM boom and ChatGPT and whatnot, was the fact that the code that the machine would just be worse in every way, harder to debug, harder to understand, et cetera, et cetera.

That wasn't important. What was important is that the same person would go around a few days later and attempt to sell me on his app idea. The idea was "Trivago, for hospitals".

Do you understand why I gave up on computing?