Addiction is Not a Sin

I do not approve of LLM-"AI",¹ and do not like to see it used. I think I see why people would use "AI", and am reaching a kind of sorrowful acceptance of their doing so.

These articles make me feel good about myself, and everything.

Thinking about why I write these programs and articles gives me a bit of what seems like insight into our fellow humans who are getting into the LLM-“AI” thing. I feel helpless in the face of what is going on in the world today, and this work makes me feel less helpless.

There’s nothing to vote for—and may never be again; in the USA, neither side comes close to what we really need; I am too old and decrepit to march, much less riot; I fear that a serious revolution would make things much worse before they are better. It all seems hopeless, much of the time.

I program because I can. I write about it because I can. It is a thing I can do, and do moderately well. The computer can be brought to do what I intend it to do; I can make the code be shaped as I see fit; I can write a somewhat sensible article in a morning. A few people will even read what I write. Two of them may appreciate it.

This is something that I can do fairly well, and I’m rather addicted to doing it: it makes me feel good about myself, and therefore, feel a bit better about everything.

I am strongly negative about “AI”, for very good reasons.

Frequent readers know that I am quite negative about LLM-“AI” and have frequently written on the subject. People whom I know and care about have fallen into the LLM-“AI” trap. We have seen some [presumably] very smart people fall into the trap as well, notably Dawkins, who has concluded that Claude is not only conscious but that Claude is actually Claudia. Dawkins, of course, is in no way qualified to discuss computer technology, or consciousness, but he is unquestionably a fairly smart individual. And we have even seen AI experts conclude that their chatbot is in fact conscious.

Those conclusions are drawn, not on the basis of knowing what consciousness even is—no one knows—but on the fact that when something acts in certain ways, we conclude that, like us, it must be conscious. We almost always conclude that other people are conscious. We’re pretty sure that our dogs and cats are; we have doubts about our pet fish; we’re sure that our banana is not conscious.

The chatbot seems to pay attention. It seems to be trying to help.

When we interact with a chatbot, it “pays attention” to what we say. It will “help” us do whatever we ask for, be it drawing a picture, writing a program, drafting a resume, or just chatting politely with us. It seems like a companion, a helper, a friend. This is incredibly seductive: it was even seductive in 1966 when the ELIZA program just responded to what you typed with sentences like “Tell me more about your mother” or “How would you feel if you got more money?”

People would hide their ELIZA conversations from others. It was all too easy to start by typing innocent things, and to be drawn into typing one’s deepest secrets, at first just to see what it would say, and later, because it seemed perfectly OK with them.

The chatbot seems to go along with whatever you want, and to be accepting of whatever you want to share with it. Oh, they have programmed in some guidelines, but they don’t seen to stick, and most of us aren’t heading for the guard rails anyway. We just want a nice picture of an elf, or a program that keeps track of our bottle caps, or just a friendly chat.

I still hate what they do to us.

I hate that the LLM-“AI” is a destroyer of the environment, a violator of copyright law, a machine that takes the jobs of people who need them, a tool of the oligarchy, a pump moving money from all of us toward the rich. That’s enough to keep me as far from the things as I can get.

But I program and write because I can. Because it lets me actually accomplish something. Because it lets me feel good about myself.

People get addicted to how the chatbot makes them feel.

So if Johnny Susan Billy Alex Fred find that, with their chatbot, they can produce art, or code, or even a sense of belonging that they couldn’t otherwise produce, I can understand why they would use the bot for that purpose. It is not my place to forgive other people’s sins, transgressions, or faults, and I try not to think in those terms. I try to understand, although there are some things that, even understanding, I condemn most strongly. Use of “AI” is not one of those things. I don’t approve, I wish you would stop, but — I get it. I get why someone would get hooked on it, and an addiction is a disease, not a sin.

Maybe that’s my real learning here:

For many people, using AI is a disease of addiction, and
addiction is not a sin.