One of the reasons I'm so skeptical of AI getting shoved into everything is that it is not forward-looking, it's just reinforcing all of the bad systems and processes we already have. It's not just that I miss a past, pre-AI world (though I do), it's that when I think of a genuinely better future, it mostly involves ripping out the systems that AI is trying its hardest to perpetuate.

AI is not the future, it's the present, repeated endlessly.

Another reason I'm skeptical of AI getting shoved into everything is that private infrastructure is a great way to concentrate wealth - just ask the railroad barons. The people pushing AI know this.

You take money that was being paid to millions of people, and pay it to one guy, who in turn employs a couple thousand guys.

If you think we have wealth inequality now, just wait...

@ricci
My hope is that these two things cancel.

If we can’t change any system because AI needs it stable, then all of our old interfaces will still work, and we avoid lock-in. If we innovate, the AI is locked out.

Maybe I’m missing something big tho. Like a making the old stuff even more Byzantine so we need AI to navigate it?

@ricci By “AI” do you mean LLM/GPT? If yes, absolutely agree. It’s a shame AI was completely conflated with this slop. While intelligence isn’t artificial and machines don’t learn, they’re great marketing tools to get grants…starting about the year I was born 👴🏼

@jadp This is a good question, I am mostly thinking about the way that AI is being *used* and *marketed*. Eg. still customer service reps, but now they're AI. Still roads jammed full of cars, but now they are driving themselves. Still putting pointless "hero images" on our blog posts but now they're AI. Issuing bullshit corporate "apology" letters but now they're AI.

Technologically, LLMs and stuff are really quite interesting but what people are doing with them or claiming they will do is pretty much just more of the same. But AI.

@ricci I see your point. Regardless of the underlying technology, mathematics, algorithms and heuristics, we’re using these new tools to do exactly what we have always done.

@jadp And that's kind of all they *can* do, by their nature.

At least LLMs, etc.

@ricci True. I’m much more excited by the work being done in causal inference. But I was excited by that in 1988, when the head of AI research where I was working, asked me to help her team understand Dr. Judea Pearl’s Bayesian Neural Networks. (I was the only Bayesian there, but I was using empirical Bayes and Bayesian Weighted Variables in Boolean trees for cause-consequence analysis.) But your point may still hold, even with different tools, we may still do what we have always done. /sigh
@jadp @ricci …we may still do what we have always done. /sigh…
and sigh again:
From the outside, the danger is great for those who own these "machines" toss in some of their own recipes for what is recommended to be repeated... business as usual
@ricci @jadp
THANKYOU. Some of its useful, but most of the uses it's being put to were already working fine without it and it's energy guzzling. Like they've found a solution and go looking for problems.
@ricci It's like an enormous, immortal cyberboomer, reinforcing the status quo of its training data. Forever.

@ricci while I think there are contexts where it's OK to antromorphise AI (as in "it's trying to..."), this isn't one of them.

There are actual human actors (with names, faces etc.) who have their interest in shaping the adoption of AI in some particular ways.

@PaniczGodek Yeah, fair, the agency here is human agency

@ricci

But... But... But.... If I don't use "AI" how will I ever demolish the environment while churning out mountains of garbage based on stolen content and forgetting how to actually do anything?!?!?!

"I used AI to....", is nothing more than, "Listen I'm not an asshole but....", for the 21st Century.

#FuckAI

@ricci very similar to how a primary objective of AVs / "self driving cars" is to prolong societies' dependence on cars.
@ricci Exactly! I dream of a world where we take the teetering tower of abstraction that is modern computing and collapse layers together from scratch wherever we can. A world where we build up stronger foundations and reinforce what's there. Not a world where we instead just add the most rickety, unpredictable, unreliable layer imaginable on top of it all and try to have it vibe away all the underlying cruft.
@ricci
I once described it as "If intelligence is the ability to come up with new ideas, AI is the ability to predict the past".
@ricci This is so eloquently said! AI will undoubtedly perpetuate all the -isms and -phobias that we’ve tried so hard to get rid of within our societal structures.

@ricci

The mechanisms that could be used to create an Economy of Abundance, instead being used to continue the Economy of Scarcity.

Short-sighted and petty.

Definitie lack of vision.

@ricci technology used to be so exciting. In the 90's we had probably the most rapid advance in consumer technology the world has ever seen, and it was *so cool*. Like you could be really optimistic about it, even if there were a couple red flags. Now whenever a new technology is developed, it's just the super-rich finding yet another way to make the world worse and squeeze more profits out of a decaying economy. In a sensible world I'd be able to be excited about AI and the future, but it's all just so bleak now.
@ricci
I hadn’t considered it explicitly, but you’re right. If someone releases a new, brilliant programming language tomorrow, adoption would be even more stunted than usual because the AI would lack a training set and be unable to help. It reinforces use of old systems.
@ricci and the past. Its input data is all of our history... Humanity has some seriously dark times.
@ricci … and incorrectly