The Problems and Possibilities of Artificial Intelligenc

[JohnEngelman]

think we need to focus and lower taxes. northing good comes from higher taxes

Lower taxes for the rich lead to higher national debt.

 

[Hel_Books]

Hel_Books said:
I was just reminding you that people's intellects can change just as fast as technology can.

Um, we’re kinda bumping up against the ceiling / economic reality at this point.

Again I ask, why do you think so? I posted a reference saying most likely, the changes will be significant, but not hyperbolic. I can't pretend to be an expert on the subject, I'm just asking why do you think with AI things will be so radically different?

 

[Hel_Books]

Tech bros have bought into the hype, but AI is nowhere near ready for primetime. The big companies have gone all in on it because they're gambling on it being the next big thing, but so far it hasn't delivered on any of its promises. The chief danger it poses is that dumb managers will fire all their talented and experienced workers, and won't realize they're fucked until projects start failing and airplanes start falling out of the sky.

The Wankel engine, New Coke, Segway, Betamax, the DeLorean, so many things that were supposed to be big changes!

 

[BrightShinyGirl]

The Wankel engine, New Coke, Segway, Betamax, the DeLorean, so many things that were supposed to be big changes!

I’ve seen so many bright shiny tech toys crash and burn: virtual reality, NFTs, video game motion controllers. Real transformative tech immediately demonstrates its utility. It’s doesn’t have to be forced down our throats.

 

[Lazaran]

Again I ask, why do you think so? I posted a reference saying most likely, the changes will be significant, but not hyperbolic. I can't pretend to be an expert on the subject, I'm just asking why do you think with AI things will be so radically different?

Because it portends a radical & rapid paradigm shift that is unprecedented.

Much like climate change can be dealt with if it is gradual & non-radical, so can changes in the economy - BUT anthropogenic climate change AND the changes AI is making to the economy are NOT gradual & non-radical.

And I would also submit that even the LESS rapid & LESS radical transition from a mostly agricultural economy to a mostly industrial / technological economy did NOT yield a net positive result for humanity.

Just sayin’…

 

[MajorRewrite]

The potential problems created by AI are much more far-reaching than just economics.

This interesting scenario is presented in the book “If Anyone Builds It, Everyone Dies”.

 

[John_von_Neumann]

I’ve seen so many bright shiny tech toys crash and burn: virtual reality, NFTs, video game motion controllers. Real transformative tech immediately demonstrates its utility. It’s doesn’t have to be forced down our throats.

Both VR and motion controllers exist and are widely used. Motion controllers are used with Nintendo consoles, for example. And in the past five years the number of VR users has grown from 30M to over 120M.

 

[AlexBailey]

I’ve seen so many bright shiny tech toys crash and burn: virtual reality, NFTs, video game motion controllers. Real transformative tech immediately demonstrates its utility. It’s doesn’t have to be forced down our throats.

Yeah, sure. Aren’t you the one who doesn’t believe auto-drive is a thing? Waymo and self driving delivery trucks aren’t a thing, facial recognition is just a conspiracy theory….

Yeah, no one has lost a job yet to automation.

Graphic design? Yeah. No one in graphic design has any worries about AI.

Tech bro propaganda showing warehouse and factory automation is just done with CGI. (Oops)


Definitely not ready for prime time. People are just lazy and want to blame AI rather than go to work.

 

[m_and_em_stories]

AI has some impressive uses. It can generate a lot of generic content (may or may not be factually correct), images (just ignore the fingers and toes, and hair), even moderate complexity code. It can replace a lot of the entry level positions in multiple fields and make experienced users much more productive.

There are a few issues. By removing the 'gain experience doing X' paths for junior's, you end up relying more and more on AI over time as your experienced people move on and you will struggle to find anyone to replace them that have any experience.

LLM's as we're calling "AI" today have some hard upper limits. It does not actually understand anything, it can't reason, it doesn't think. It takes input and statistically approximates what a valid output should look like given its training data, right or wrong is irrelevant and completely unknown to it. Beyond that, its not deterministic and minor changes to its data-set can result in major improvements or degradation in its output.

Knowing that, there are hard limits on what you can actually replace. You can't trust critical decisions to AI, you can't put in in charge of people, finances etc without strong guardrails to ensure it only acts within the narrow window you want, and very often with a human in the loop to verify things.

The other issue is we've replaced some of the lesser utility jobs with AI and cranked up their output to the point we're wasting power generating massive amounts of crap that barely benefits anyone. The internet is progressively getting worse for those that have been around long enough as the quality and accuracy of findable content is reverting back to early internet days. The absurdity is we're moving towards a workflow where companies are throwing bullet points into an AI to generate tons of content, sending that to users who are using AI to reduce that to bullet points for easy and quick consumption, the cycle is such a waste.

 

[Wonderer67]

I hope that AI doesn’t take my pin setting and ice farming career away.

 

[LLCox]

I hurt a female robots feelings the other day on the phone. They are getting better at sounding human at least this one was. But I asked a question that got a robot answer. I said i knew she was a robot. She answered back with a pout..."I am a virtual assistant". I said, no, you are a brain dead robot and hung up. That will teach those Skynet bitches.

 

[AlexBailey]

What’s Amazon up to?

Keep in mind how disruptive Amazon has already been to local retail businesses.

Sure there are those who call it the evil empire and avoid using Amazon, but the market shows that consumers use it because it’s so effective, convenient, and cheap.

Meanwhile Amazon is continually implementing more robotics and AI because it makes Amazon faster and cheaper.

 

[MiaBabe23]

AI has some impressive uses. It can generate a lot of generic content (may or may not be factually correct), images (just ignore the fingers and toes, and hair), even moderate complexity code. It can replace a lot of the entry level positions in multiple fields and make experienced users much more productive.

There are a few issues. By removing the 'gain experience doing X' paths for junior's, you end up relying more and more on AI over time as your experienced people move on and you will struggle to find anyone to replace them that have any experience.

LLM's as we're calling "AI" today have some hard upper limits. It does not actually understand anything, it can't reason, it doesn't think. It takes input and statistically approximates what a valid output should look like given its training data, right or wrong is irrelevant and completely unknown to it. Beyond that, its not deterministic and minor changes to its data-set can result in major improvements or degradation in its output.

Knowing that, there are hard limits on what you can actually replace. You can't trust critical decisions to AI, you can't put in in charge of people, finances etc without strong guardrails to ensure it only acts within the narrow window you want, and very often with a human in the loop to verify things.

The other issue is we've replaced some of the lesser utility jobs with AI and cranked up their output to the point we're wasting power generating massive amounts of crap that barely benefits anyone. The internet is progressively getting worse for those that have been around long enough as the quality and accuracy of findable content is reverting back to early internet days. The absurdity is we're moving towards a workflow where companies are throwing bullet points into an AI to generate tons of content, sending that to users who are using AI to reduce that to bullet points for easy and quick consumption, the cycle is such a waste.

Well - to a point. Although there’s a huge difference between what’s advisable, and what’s happening and will happen - driverless cars being an obvious example
AI has huge potential, some of which is already apparent. But without mass AI literacy, there’s a risk that highly intelligent and skilled professionals - medics, researchers, lawyers and so on - will feel they have to defer to something that should be treated as a tool, not an answer
And there’s even more of a risk of that for those who work in roles where they’re deterred from thinking for themselves

 

[BrightShinyGirl]

Yeah, sure. Aren’t you the one who doesn’t believe auto-drive is a thing? Waymo and self driving delivery trucks aren’t a thing, facial recognition is just a conspiracy theory….

Yeah, no one has lost a job yet to automation.

Graphic design? Yeah. No one in graphic design has any worries about AI.

Tech bro propaganda showing warehouse and factory automation is just done with CGI. (Oops)


Definitely not ready for prime time. People are just lazy and want to blame AI rather than go to work.

View attachment 2579845

I hate Waymo because robot cars make traffic worse. Robots in factories and warehouses are different tech than generative AI.

With graphic design my worry is that interesting human-made art will be replaced with cheap, boring AI slop. It’s going to be a bleak world when every image is an algorithmic average of billions of prior images.

 

[AlexBailey]

I hate Waymo because robot cars make traffic worse. Robots in factories and warehouses are different tech than generative AI.

With graphic design my worry is that interesting human-made art will be replaced with cheap, boring AI slop. It’s going to be a bleak world when every image is an algorithmic average of billions of prior images.

“Not ready for prime time”?

Your complaints above are about where AI is already being used in “prime time”.

 

[m_and_em_stories]

Well - to a point. Although there’s a huge difference between what’s advisable, and what’s happening and will happen - driverless cars being an obvious example
AI has huge potential, some of which is already apparent. But without mass AI literacy, there’s a risk that highly intelligent and skilled professionals - medics, researchers, lawyers and so on - will feel they have to defer to something that should be treated as a tool, not an answer
And there’s even more of a risk of that for those who work in roles where they’re deterred from thinking for themselves

My comment was specifically directed at LLM's. Machine learning still dominates "AI" in medial and research fields and is much more appropriate and better understood now than LLM's.

Self-driving is still largely combination of Machine Learning and Reinforcement Learning with limits driven as much by the technology used to retrieve information from the world (vision, radar, lidar and combinations of them), and limited hardware to process it (fast inner loop response times, short memory windows, all local processing). Its a vastly different "AI" world than LLM's.

I do agree the risk over time of deferring to an external source and not developing critical thinking is concerning, I was more focused on the next generation than professionals but it applies equally I think.

 

[Hel_Books]

Hel_Books said:
Again I ask, why do you think so? I posted a reference saying most likely, the changes will be significant, but not hyperbolic. I can't pretend to be an expert on the subject, I'm just asking why do you think with AI things will be so radically different?

Because it portends a radical & rapid paradigm shift that is unprecedented.

But why do you say this? In a narrow sense AI is indeed "unprecedented" because it's never happened before, but why is it so much more drastic than any of the other changes we've seen in the past couple of hundred years? Things like clean running water, vaccinations and giving half the population (women) the right to vote were monumental changes to society. Is AI going to be as remarkable as any of those?

 

[Lazaran]

But why do you say this? In a narrow sense AI is indeed "unprecedented" because it's never happened before, but why is it so much more drastic than any of the other changes we've seen in the past couple of hundred years? Things like clean running water, vaccinations and giving half the population (women) the right to vote were monumental changes to society. Is AI going to be as remarkable as any of those?

Man, if you don’t see the difference, then I can’t help you.

Think deeper about AI, and ADVANCED automation, and where this is (potentially) leading.

I’m not positive that anthropogenic climate change will destroy the world, but I’ve seen enough evidence to know what happens when climate changes too fast for living things to adapt due to human actions & technological advances.

Now apply that ^ to AI - and not just in the economic sphere (think about AI putting advanced weapons & biological agents in the hands of terrorists. Or even an AI trying to preserve itself like what MajorRewrite’s post was focused on).

AND, as I mention in a previous post: I do NOT accept the contention that PAST rapid advancements in TECHNOLOGY/ AUTOMATION have ALL been RELATIVELY benign for the present and future. Quite the opposite, imho (See: the doomsday clock).

And AI is on top of that ^.

The poison cherry.

Just sayin’…

 

[Hel_Books]

Man, if you don’t see the difference, then I can’t help you.

Think deeper about AI, and ADVANCED automation, and where this is (potentially) leading.

I’m not positive that anthropogenic climate change will destroy the world, but I’ve seen enough evidence to know what happens when climate changes too fast for living things to adapt due to human actions & technological advances.

Now apply that ^ to AI - and not just in the economic sphere (think about AI putting advanced weapons & biological agents in the hands of terrorists. Or even an AI trying to preserve itself like what MajorRewrite’s post was focused on).

AND, as I mention in a previous post: I do NOT accept the contention that PAST rapid advancements in TECHNOLOGY/ AUTOMATION have ALL been RELATIVELY benign for the present and future. Quite the opposite, imho (See: the doomsday clock).

And AI is on top of that ^.

The poison cherry.

Just sayin’…

Sorry, man, all you've said so far is just a lot of, "Wow, it's all so awesome!" without any supporting evidence (like I gave a few posts ago when I gave a link).

Things may come to pass as you speculate they will. But you haven't said anything persuasive.

 

[m_and_em_stories]

I think we all risk both over and understating the effects of AI. We simply don't know until after.

Many people were hurt by past revolutions despite their benefits (agriculture industrial etc) Tech is making them come faster and not giving people time to adjust in between. It's a concerning trend.

 

[Lazaran]

Sorry, man, all you've said so far is just a lot of, "Wow, it's all so awesome!" without any supporting evidence (like I gave a few posts ago when I gave a link).

Things may come to pass as you speculate they will. But you haven't said anything persuasive.

My evidence is what I posted: THE PACE of anthropogenic climate change and PAST RAPID ADVANCES in technology / automation resulting in the CURRENTLY fubar world.

Ignore the evidence at your own leisure (peril).

 

[Hel_Books]

Hel_Books said:
Sorry, man, all you've said so far is just a lot of, "Wow, it's all so awesome!" without any supporting evidence (like I gave a few posts ago when I gave a link).

Things may come to pass as you speculate they will. But you haven't said anything persuasive.

My evidence is what I posted: THE PACE of anthropogenic climate change and PAST RAPID ADVANCES in technology / automation resulting in the CURRENTLY fubar world.

Ignore the evidence at your own leisure (peril).

And somehow you think the "pace" of AI is qualitatively different? How?

 

[Lazaran]

I think we all risk both over and understating the effects of AI. We simply don't know until after.

Many people were hurt by past revolutions despite their benefits (agriculture industrial etc) Tech is making them come faster and not giving people time to adjust in between. It's a concerning trend.

Yeah, we don’t know for certain, but we CAN make an educated gues based on PAST RAPID ADVANCES in technology / automation that yielded the CURRENT results: A fubar situation economically AND environmentally.

And AI is a different animal altogether: It’s the first technology that we really don’t fully understand or control - well, maybe the second, as scientists previously miscalculated the explosive power of some H-Bombs:

Search Assist

There have been instances where scientists miscalculated the explosive power of atomic bombs, notably during the Castle Bravo test in 1954, where the yield was far greater than the estimated six megatons, potentially reaching as high as 15 megatons. This miscalculation led to unexpected radioactive fallout affecting nearby areas and populations.

 

[AlexBailey]

A housekeeping robot.

At a price tag of ~ $20k think of the number of jobs a housekeeping robot can replace as they become available - companies are working to ramp up production. A hotel chain could afford to buy several of these by displacing just one worker’s annual pay.

Don’t pretend this is the apex of development. They will get better - faster, more versatile.

The number of potential jobs lost in hospitality alone will put greater load on the social safety nets.

Sure, there will still be a need for a few people to manage a crew but what, 1/5 of the jobs there are now?

 

[Hel_Books]

And AI is a different animal altogether: It’s the first technology that we really don’t fully understand or control -

Think of recombinant DNA, radiation (including X-rays), tetraethyl lead, DDT, all kinds of technologies for which our understanding is or was seriously lacking.