News sucks

[WAB]

I actually kind of liked Picasso's earlier stuff, that red & blue phase. I'm just not really big on cubism.

Picasso’s red and blue phase was inspired by Gauguin. Cubism was inspired by Cezanne,

In Cubism, Picasso and Braque attempted to do for form what the impressionists did for light. The latter were interested not so much in painting what they saw but in painting light, breaking it up into dots of complementary colors.

The Cubists were exploring form from multiple perspectives at the same time.

Maybe we need a philosophy of art thread.

Yes, we do!

 

[Swammerdami]

AI is improving rapidly. Not long ago, Dr Z's OP would have been thought too optimistic about AI efficacy. Now it's almost too pessimistic. I just watched a 16-minute video "Yes, AI Will Take Your Job. But What Happens NEXT Is Worse". It may contain nothing new, but click if you want 16 minutes of gloom.

At the end he links to an even gloomier 37-minute video "AI 2027: A Realistic Scenario of AI Takeover." Maybe I'll watch it (at 1.25X speed) some day when I feel too happy.

 

[pood]

AI is improving rapidly. Not long ago, Dr Z's OP would have been thought too optimistic about AI efficacy. Now it's almost too pessimistic. I just watched a 16-minute video "Yes, AI Will Take Your Job. But What Happens NEXT Is Worse". It may contain nothing new, but click if you want 16 minutes of gloom.

At the end he links to an even gloomier 37-minute video "AI 2027: A Realistic Scenario of AI Takeover." Maybe I'll watch it (at 1.25X speed) some day when I feel too happy.

I generally don’t watch videos. The video is based on this report, which I have not yet read.

Perhaps the video, or the report, addresses a simple question:

If AI takes away all our jobs, who is going to have any money to buy the goods and services that companies using AI produce? From which it follows that the companies who use AI to produce everything will … go out of business. Lacking a customer base and all.

Perhaps all is not dire, though. If AI produces everything (I don’t believe that is going to happen, but still) it seems with all the money these companies save not having to pay real workers, they can be taxed to provide a universal income for everyone,

 

[pood]

I worked in journalism for many years. I see no prospect that AI can rush out to the scene of a fire and interview survivors. To do that, it would have at least to be fitted with a suitable robotic body that can do everything human bodies can do, and we are far from that, and it may be impossible. In addition, since AI is not self aware so far as we know, it would never be able to ask survivors appropriate empathetic and meaningful questions. And while AI may be getting more competent, it still fucks a lot of stuff up (though of course humans do too).

Are we going to have AI bartenders, lumberjacks, line cooks, chefs, circus performers? If AI is a threat to jobs they would seem mainly to be white-collar jobs that involve a lot of data crunching.

---

 

[pood]

AI Russian robot shows us our glorious future.

Of course, this is Russia, which is incompetent at everything, so there is that.

And they named it … Adol??? How appropriate for modern Nazi Russia. Just add the latter “f” at the end, and give it a little mustache.

 

[pood]

I think the face planting Russian Adol(f) is a fitting metaphor for Russia spectacularly face planting in Ukraine,

 

[WAB]

AI Russian robot shows us our glorious future.

Of course, this is Russia, which is incompetent at everything, so there is that.

And they named it … Adol??? How appropriate for modern Nazi Russia. Just add the latter “f” at the end, and give it a little mustache.

It's Aidol.

And I say, people fall over, so a robot that falls over is more realistic.

You know when I'll buy a robot? When they can make hootch:



Well I'd have to win the lottery first.

 

[bilby]

I worked in journalism for many years. I see no prospect that AI can rush out to the scene of a fire and interview survivors. To do that, it would have at least to be fitted with a suitable robotic body that can do everything human bodies can do, and we are far from that, and it may be impossible. In addition, since AI is not self aware so far as we know, it would never be able to ask survivors appropriate empathetic and meaningful questions. And while AI may be getting more competent, it still fucks a lot of stuff up (though of course humans do too).

Are we going to have AI bartenders, lumberjacks, line cooks, chefs, circus performers? If AI is a threat to jobs they would seem mainly to be white-collar jobs that involve a lot of data crunching.

---

Over my lifetime we have gone from white collar workers being suspected by the blue collar workforce of not doing actual work, through white collar workers being convinced that they do most of the real work, to white collar workers believing that everyone is a white collar worker.

None of these have ever been true.

Automation took away a lot of blue collar jobs, and now it is possible for a cubicle dweller to never encounter anyone who doesn't work at a desk, with a computer. Well, it isn't, but the folk they do encounter in that category are mostly invisible - janitors, bartenders, bus drivers and the like. Servants.

To such a cubicle dweller, it's pretty easy to imagine AI taking over "all the jobs", by which they mean "all the real jobs", by which they mean "all the jobs that entail staring at a computer all day".

Of course, it won't. AI isn't up to the task. But even if it could and did, there would still be a need for someone to go and stick a microphone in the face of a traumatised homeowner who just lost everything, for the entertainment of a TV audience who might have forgotten, since last week, what the survivor of a house fire looks like.

 

[DrZoidberg]

AI is improving rapidly. Not long ago, Dr Z's OP would have been thought too optimistic about AI efficacy. Now it's almost too pessimistic. I just watched a 16-minute video "Yes, AI Will Take Your Job. But What Happens NEXT Is Worse". It may contain nothing new, but click if you want 16 minutes of gloom.

Its just nonsense. He doesn't understand what AI is or what it does. AI is not that smart. Its just scare mongering. Its click bait.

Yes, lots of people will lose their jobs. But its the simple rote skill jobs that are disappearing. Any job that requires thinking is safe.

Junior employees will still be employed because companies still need future experts. That's been true in the IT industry since its inception.

And most importantly, each employee will perform more quality work, ie everyone working will add more genuine value. So there will be more value generated from the entire system. Profit that can be reinvested and make the entire economy grow.

The unemployment from AI is a non issue. Short term it might suck for some people, but they'll all find new jobs eventually.

It's a world apart losing your job in an expanding market vs a depression.

At the end he links to an even gloomier 37-minute video "AI 2027: A Realistic Scenario of AI Takeover." Maybe I'll watch it (at 1.25X speed) some day when I feel too happy.

I haven't watched this. Are you sure its not also garbage?

 

[DrZoidberg]

I worked in journalism for many years. I see no prospect that AI can rush out to the scene of a fire and interview survivors. To do that, it would have at least to be fitted with a suitable robotic body that can do everything human bodies can do, and we are far from that, and it may be impossible. In addition, since AI is not self aware so far as we know, it would never be able to ask survivors appropriate empathetic and meaningful questions. And while AI may be getting more competent, it still fucks a lot of stuff up (though of course humans do too).

Are we going to have AI bartenders, lumberjacks, line cooks, chefs, circus performers? If AI is a threat to jobs they would seem mainly to be white-collar jobs that involve a lot of data crunching.

---

Over my lifetime we have gone from white collar workers being suspected by the blue collar workforce of not doing actual work, through white collar workers being convinced that they do most of the real work, to white collar workers believing that everyone is a white collar worker.

None of these have ever been true.

Automation took away a lot of blue collar jobs, and now it is possible for a cubicle dweller to never encounter anyone who doesn't work at a desk, with a computer. Well, it isn't, but the folk they do encounter in that category are mostly invisible - janitors, bartenders, bus drivers and the like. Servants.

To such a cubicle dweller, it's pretty easy to imagine AI taking over "all the jobs", by which they mean "all the real jobs", by which they mean "all the jobs that entail staring at a computer all day".

Of course, it won't. AI isn't up to the task. But even if it could and did, there would still be a need for someone to go and stick a microphone in the face of a traumatised homeowner who just lost everything, for the entertainment of a TV audience who might have forgotten, since last week, what the survivor of a house fire looks like.

The human connection. We like to think we're rational beings. But we're not. We're emotionally driven beings who occasionally can park some of our emotions in some ways sometimes, to think somewhat rationally.

We're still mostly a bunch of monkeys who need to feel emotional connection with real other monkeys.

AI can't replace that.

One thing AI generated media clips has shown is that watching actors on TV is not real human connection. It triggers some of those feelings, but ultimately it only adds to disconnection.

We need to watch TV less and hang out together for genuine human connection. Less media, more connection

There's so much AI can never replace. We should let AI replace the useless interactions that don't add to human connection.

That's my soap box

 

[pood]

There's so much AI can never replace. We should let AI replace the useless interactions that don't add to human connection.

That's my soap box

Like the human connection of bombing the shit out of innocent children in Gaza?

 

[Elixir]

There's so much AI can never replace.

Right?
AI can't even take potshots at National Guardsmen.
Thinking about the trumpsucking judge who put a hold on the finding that this deployment is illegal (which it obviously is).
I hope he is up late at night re-thinking his obsequiousness.

 

[Loren Pechtel]

Krishnamurti wondered why people flock to museums to gawk at pretty portrayals of nature when they can just experience nature firsthand. Henry Miller observed that K’s philosophy made writing itself seem superfluous, not too different from Socrates. Miller lived in Pacific Palisades not far from Ojai where Krishnamurti lived and posted a note on his door that said, in condensed form, if you want enlightenment, don’t knock on my door. Go visit Ojai.

I gawk at portrayals of nature that I can't reasonable access from my location, or which feature events that I doubt I'd be lucky enough (or brave enough) to experience directly. A typhoon in the mediterranean with a ton of lightning and tormented ships isn't really on my travel itinerary. And while I could certainly make plans to visit Yellowstone every now and then, the chances of me being there when the light is just right is kind of small...

Exactly. If it's both spectacular and easily accessible it will already be a tourist spot and very crowded. The reality is most of the good stuff entails both skill and risk. I do not believe I'm capable of class 5 stuff (what you typically see in mountaineering) anymore and certainly would not be willing to without a controlled environment. A climbing gym is skill. A mountain is skill plus luck that you don't have a weak rock or the like.

 

[Loren Pechtel]

My claim is that using AI has no downside. If AI fails at a task you have lost nothing.

Mmm.... If I recall, I think there were two fairly substantial lawsuits in the past few years related to AI having a downside. I believe one of them was a hospital that was using AI to augment diagnoses and the formation of treatment protocols, and it was found to be discriminatory. The other was I think United (not positive, one of the big insurers) using AI as part of it's prior authorization evaluation, also found to be discriminatory.

I still don't see what the problem is.

The AI is just following the rules its given. They fucked up when promoting it and where lazy when validating the results. Something they would have needed to do anyway without AI.

AI output needs far more validation as humans generally don't make up nonsense. If getting it in the ballpark is good enough, go ahead, use an AI. If not, quit pretending it's manglement's holy grail, a magic formula to get more and pay less. All too many in management are on an eternal quest for said magic formula, the snake oil salesmen keep peddling new ones.

Saw it again and again at tech shows--all the effort was put on marketing to managers that would believe the hype. When us technical types showed up dressed very casually we were immediate identified for what we are. They knew we weren't tire-kickers, but selling us hype didn't work.

The biggest problem with AI is that some people think its magical. When people realise its not magical they become angry at it. The irony is of course they could just have asked the AI how it works

AI doesn't need to be perfect. It just needs to save work. Since its so cheap to use its very hard not to profit from using it

And here's where you do not get it. The problem is that low quality work often has negative value.

 

[unapologetic]

Here's a good video on AI

Ugg..
As you probably noticed, My language skills suck.
For me, wrighting a good prompt would be more trouble than just doing the research myself.

 

[unapologetic]

AI generated Jimmy Kimmel.
The give-away: He says the shot national guards are "two zero year old" and "two two year old" when they are 20 and 22. Not a mistake any human would make. (also, it's not funny)

 

[bigfield]

Lately I've been getting vendor "support" from a team who are copy-pasting from a chatbot to reply to their support tickets. They're using a built-for-purpose marketed-for-purpose customer service chatbot platform, which they disclosed by accident by sending bot slop without reading it first.

Every response from the support team is a big shit sandwich.

This must be what it feels like to talk to DrZoidberg's HR department.