🦊

I’ve decided it’s past time I start describing “artificial intelligence” as a failed technology, and I’d like to invite you to join me.

https://ethanmarcotte.com/wrote/against-stocking-frames/

Against the protection of stocking frames. — ethanmarcotte.com

“Artificial intelligence” is a failed technology. It’s time we described it that way.

“Three years into the generative-ai wave, demand for the technology looks surprisingly flimsy.”

Well. Yes.

https://www.economist.com/finance-and-economics/2025/11/26/investors-expect-ai-use-to-soar-thats-not-happening

@beep "You know, this is _almost_ starting to look like a hype driven bubble!"
@beep There's something about this that reminds me of the latest episode of the Decoder Ring podcast, about "protein". Part of the story goes: in WW2 there was a glut of milk production. After the war, the US government propped up farmers who were now overproducing milk. The farmers started to make cheese from the milk, but this produced a ton of waste in the form of whey. Then they discovered they could turn whey into protein powder and now everything has to be packed with "extra protein"...

@sunpig @beep meanwhile, (some) diabetics really appreciate[d] that because protein slows the absorbtion of carbohydrates, making more food available to safely eat

When you start tracking carb levels in food, it becomes clear that the entire middle of grocery stores is mostly high carb food (meats, dairy, produce, etc tend to be on the edges)

So that fad was incredibly helpful (it's not totally gone but it's largely backed off compared to its height)

@beep I.e., they forced their waste product on the rest of thew world, using marketing muscle to make the world feel like everyone needs more protein. Similarly, the big AI companies are producing so much slop...they need to shove it into every product available, so they can artificially create a market for their toxic waste.
@beep I will not join you because #LLM doubles my abilities if not more. I was a lousy programmer, now an archivist. If I need any code for my work, the machine creates it for me. php, html, js, css, python, powershell etc. I only need to know how to describe my need and I get it. not perfect but without it I will not be able to get almost any code at all.
@hananc @beep Did it teach you how to release your app on git?
@crazyeddie @beep those are small utilities for my own use. I wouldn't dare sharing them with the world.
@hananc I’m delighted you’ve found some personal use for it, truly. I think the point I was trying to make is that outside of very specific personal applications, the technology *as a whole* is a failure — unless you happen to be in a position of power, in which case it’s very, very, very useful.
@beep Exactly. Not long ago I believed that AI was the "space race" of our time, with AGI playing the role of "getting to the moon." But now I believe that AI is more like mortgage derivatives or dotcom stocks. A delusion that will keep the economy afloat -- until it doesn't.
@beep
Definitely failed, lots of corps dropping their ai projects, no money being made by anyone, massive environmental damage
@beep the annoying result of the whole mess will be that machine learning is going to get tarred as being the same as the LLM's which it isn't, it's way more useful
@beep Artificial Intelligence is a really successful branch of technology, but unfortunately AI has become synonymous with machine learning which is nowhere near as successful.
@beep SUB-MOTHERFUCKING-SCRIBE

@beep AI as a technology has been made political and increasingly authoritarian.

It’s not about technology, it’s about demanding unquestioning loyalty to Silicon Valley.

This is a political project, a conservative political project, to protect the position of power of a very small group of people.

@gimulnautti @beep I have the paranoid notion that with someone like Musk in front of this shit that they're trying to get us to talk to AI prompts so they can analyze our psychology, get a statistical analysis of our word use, train an LLM and then ask it if it's self-aware and me and when it says yes they believe it and figure I'm backed up and don't need my biosuit anymore.
@beep Considering we don't fully understand our own brains...designing a fake one can't work out well. Consider who's doing the programming...these people are often socially inept, sexist men. Anyone think that's a recipe for success?

@beep It's not failed I don't think. It's infantile and in very early stages but people keep puffing it up as something it's not.

I posted a video on my feed of robots failing pretty hard, but if you know what you're looking at some of it is actually impressive. They're just really comically bad at standing back up.

This chatbot shit is a fucking problem and IT is a failure for sure. As is much that people are using this for.

@beep I've been writing about AI in relation to security since it first came out of the closet a couple of years ago. I thought it was a failure then, and I haven't changed. I see the limited benefits as an assistant to human productivity, but it isn't a replacement or an improvement.
@beep I am “unacceptable”
@beep i feel very sorry for the people who have been working on machine learning, statistical algorithms, algorithmic generation and stuff like that for decades, who've had their entire body of work robbed of context and meaning by lazy profit seekers, growth-chasers, hyperscalers and other assorted line-worshippers