News sucks

[Emily Lake]

Yeah, I’ll take No.1, please. Not a hard choice.

View attachment 52756View attachment 52757

Style is there, visual technique, I guess. But yeah... not sure why you'd have a group of dudes squatting in a field, peeling potatoes. Nor why one of them would be sitting on potatoes. Nor why any of them would seem happy about it. It's just kind of disconnected from reality.

I think it's getting better, but it's missing the creative spark. I'm not sure AI will ever be able to actually develop it.

To elaborate - there's a tik tok thing that does AI-generated weird little vegetable/gnome/critters. Over the course of a year, I've watched them learn to eat. In the beginning, AI seemed to have some vague notion that mouths open, and then there's some stuff with a tongue. But half the time the food never made it into the mouth, or the tongue just sort of flopped out and drooled everywhere. After a year of constant tinkering, now they kind of take a bit... but the mechanics are wrong. It's still much improved from where it started, but it's really apparent that AI doesn't have an actual understanding of *how* and *why* critters eat. It doesn't really get how jaw articulation and chewing and swallowing works. AI doesn't have a body, it will never actually have the experience of biting into an apple or a piece of cheese, or licking an ice cream cone - so it doesn't have a reference for what that should look like, and certainly no clue as to why what it has done looks wrong.

Similarly, AI doesn't have actual emotions. It doesn't feel any of the things that go into art. It doesn't have a benchmark for why the things it's drawing would be happening.

 

[pood]

Larry Moran on AI.

So, yeah, if a certain someone loves this brave new world of AI and thinks all he/she has to do is a press a button to churn out accurate and meaningful text and then maybe do a bit of copy editing, or press a button and churn out a masterpiece of art that artists who are “losers” can’t compete with, that certain someone is sadly deluded.

 

[pood]

When TV arrived people were rhapsodizing about how families would sit around glued to the tube watching opera and marvelous concerts and scientific and philosophical lectures. Instead we got I Love Lucy and The Flintstones. Not that there was anything wrong per se with those, but still .. by 1962 FCC chairman Newton Minow was decrying TV as “a vast wasteland.”

When the internet arrived it was hailed as the “Information Superhighway” and everyone would become more intelligent and better informed and yada yada. Need I say what a horror show it actually is, except for a handful of little oases like this place? (And even this place ain’t always that great.)

Same thing with AI. It will make us stupider not smarter. It is already happening.

 

[Emily Lake]

When TV arrived people were rhapsodizing about how families would sit around glued to the tube watching opera and marvelous concerts and scientific and philosophical lectures. Instead we got I Love Lucy and The Flintstones. Not that there was anything wrong per se with those, but still .. by 1962 FCC chairman Newton Minow was decrying TV as “a vast wasteland.”

When the internet arrived it was hailed as the “Information Superhighway” and everyone would become more intelligent and better informed and yada yada. Need I say what a horror show it actually is, except for a handful of little oases like this place? (And even this place ain’t always that great.)

Same thing with AI. It will make us stupider not smarter. It is already happening.

You're not wrong. I've been contemplating this recently, sort of rolling around in the back of my brain.

One of the only real advantages humans have from an evolutionary perspective is extrapolative thinking, and that's fueled quite strongly by an amped up predisposition toward pattern recognition. But for pattern recognition to be really good, it needs a lot of patterns to recognize - and they have to be real patterns. My current favorite speculation is that memory is a necessary component of that pattern recognition - you have to be able to remember similar incidents well enough to be able to identify the common threads in them. Radio and TV introduced a lot of sitcoms and cartoons, sure, but in some fashion or other we've had plays and punch-and-judy and similar for all of recorded history. I'm sure we lost something when we outsourced our imagination to others on a daily basis, but you still needed working memory to be able to follow a weekly installment of a program.

Streaming services kind of killed that - I can now binge watch a decade of The Office, without having to embed what's going on in my working and mid-term memory. The internet is arguably worse, on two fronts. First, nobody has to remember anything, they can just google it and regurgitate what the internet tells them. That tends to short circuit critical thinking and the evaluation of information being received, and I think it's eroding our cognitive function as a whole generation or two or three. Second is that it presents information as if all info is equal, and it makes all voices seem as if they're valid. It lets kooks connect with other kooks, and share their kookery back and forth until they're all convinced that their kookery is valid because look at how many people say the same kooky thing?

I think there's a lot of reasonable potential for AI, as long as we learn really quickly that it's not actually intelligent at all. I end up a bit concerned about the way in which AI can be fed falsehoods, and then repeat those back to whoever asks. AI doesn't currently have any actual way to perceive the real world, and hence no actual functionality for distinguishing fact from fantasy.

 

[pood]

It is my opinion — I could be wrong — that AI will never be sentient or self-aware, because I suspect that these traits are confined to embodied and evolved entities.

Could something like HAL 9000 exist? I seriously doubt it. Would an actual digital sentient and self-aware entity like HAL, if it could exist, be afraid of being disconnected — be afraid of death? I seriously doubt that, too, Fear of death, or at least of harm, are evolved traits.

Fodder for late-night philosophical speculation.

I will stick with human writers and artists and thinkers and disregard AI shuck and jive.

 

[pood]

When AI can produce a work of art without “scraping” (stealing) stuff off the internet, and that art can make me THINK and FEEL, the way the Hopper and Van Gogh works I posted do, please wake me up. Until then …

Of course, some people’s standards are lower than mine.

 

[pood]

When I look at the Potato Eaters I thihk, wow, so that is what it was like being a peasant in 1880s Netherlands. Bear in mind that until he painted this, most images of peasants were idealized and sentimentalized, even by one of Van Gogh’s artistic heroes, Millet.

When I look at the AI image I think, why is that asshole sitting on potatoes, and what’s up with the goofy grins?

 

[DrZoidberg]

AI is... not as I as it's hyped to be. Can be useful, but also not always as effective as I want.

For example... I needed to do some research for a potential future product line. I figured this would be agreat opportunity for AI to help me out, so I asked "What are the current regulations for this kind of product in my state". AI, in its infinite wisdom, gave me wrong information... but it gave it to me in a very well written and convincing way. If I were somewhat less informed, I would have passed that wrong info along to the executives. I told it "That bill was proposed, but did not pass". At which point, AI said "you're absolutely correct, this bill was proposed by so-and-so during such-and-such session, and was voted against by this-and-that margin". I followed up immediately afterward with "okay, so what's the current regulations?" And it gave me the same failed bill info once again.

Then I had to go dig through regulatory BS.

Ie, you used it for the wrong thing. Using AI as a fancy Google will give you a lot of errors. That's not what its for. If you think this is what AI is for you will be dissapointed.

First find the relevant information the old fashioned way. Then feed what you got to your AI to collate it = win.

 

[bigfield]

Then feed what you got to your AI to collate it = win.

My org has Copilot embeded in Teams, and one of its features is a thread summariser.

It doesn't work. So far it has a 100% screwup rate with threads I've asked it to summarise. It gets close, but it always contains some fabrications.

These tools cannot be trusted to perform any kind of analytical work without a human user checking the output against the inputs.

They can give me suggestions, but they can't give me facts.