Technology

Good luck. Even David Attenborrough can't help but anthropomorphize. People will feel sorry for a picture of a dot separated from a cluster of other dots. The play by AI companies is that it's human nature for us to want to give just about every damn thing human qualities. I'd explain more but as I write this my smoke alarm is beeping a low battery warning, and I need to go put the poor dear out of its misery.

I think we should start by not following this marketing speak. The sentence "AI isn't intelligent" makes no sense. What we mean is "LLMs aren't intelligent".

Anyone pretending AI has intelligence is a fucking idiot.

I've never been fooled by their claims of it being intelligent.

Its basically an overly complicated series of if/then statements that try to guess the next series of inputs.

The other thing that most people don't focus on is how we train LLMs.

We're basically building something like a spider tailed viper. A spider tailed viper is a kind of snake that has a growth on its tail that looks a lot like a spider. It wiggles it around so it looks like a spider, convincing birds they've found a snack, and when the bird gets close enough the snake strikes and eats the bird.

Now, I'm not saying we're building something that is designed to kill us. But, I am saying that we're putting enormous effort into building something that can fool us into thinking it's intelligent. We're not trying to build something that can do something intelligent. We're instead trying to build something that mimics intelligence.

What we're effectively doing is looking at this thing that mimics a spider, and trying harder and harder to tweak its design so that it looks more and more realistic. What's crazy about that is that we're not building this to fool a predator so that we're not in danger. We're not doing it to fool prey, so we can catch and eat them more easily. We're doing it so we can fool ourselves.

It's like if, instead of a spider-tailed snake, a snake evolved a bird-like tail, and evolution kept tweaking the design so that the tail was more and more likely to fool the snake so it would bite its own tail. Except, evolution doesn't work like that because a snake that ignored actual prey and instead insisted on attacking its own tail would be an evolutionary dead end. Only a truly stupid species like humans would intentionally design something that wasn't intelligent but mimicked intelligence well enough that other humans preferred it to actual information and knowledge.

So why is a real “thinking” AI likely impossible? Because it’s bodiless. It has no senses, no flesh, no nerves, no pain, no pleasure.

This is not a good argument.

Steve Gibson on his podcast, Security Now!, recently suggested that we should call it "Simulated Intelligence". I tend to agree.

This article is written in such a heavy ChatGPT style that it's hard to read. Asking a question and then immediately answering it? That's AI-speak.

The idea that RAGs "extend their memory" is also complete bullshit. We literally just finally build working search engine, but instead of using a nice interface for it we only let chatbots use them.

People who don't like "AI" should check out the newsletter and / or podcast of Ed Zitron. He goes hard on the topic.

My thing is that I don’t think most humans are much more than this. We too regurgitate what we have absorbed in the past. Our brains are not hard logic engines but “best guess” boxes and they base those guesses on past experience and probability of success. We make choices before we are aware of them and then apply rationalizations after the fact to back them up - is that true “reasoning?”

It’s similar to the debate about self driving cars. Are they perfectly safe? No, but have you seen human drivers???