Master AI (Artificial Intelligence) Discussion/News Thread

[pablolie]

Following this discussion is interesting. Lots of fear and uncertainty - which if course is a pretty universal thing with AI.

I don't want to be repetitive, since many are making great points, but one thought I'd like to add: Don't some of you think a lot of the AI backlash is because now the threat to white collar jobs is perceived as very real? Is that the main difference compared to previous rounds of "efficiency"... like outsourcing and offshoring manufacturing... or who remembers switchboard or elevator operators? Isn't that the way market economy naturally works, always streamlining for efficiency? Is the AI threat turning white collar workers socialist? Can't quite recall white collar workers being up in arms like this when blue collar jobs were sacrificed left and right...

I find myself wondering if this "AI Revolution" is really that different from previous waves of productivity increase. Think about it how many office jobs were eliminated with the use of computers - typists, assistants, finance people doing stuff by hand... a lot of them.

And the other factor is... is the real threat to white collar jobs exaggerated because we live in an AI bubble? Many billions are being poured into AI marketing to companies... inevitably its real capabilities are over-marketed. If it's really that much of a no-brainer to spend billions on AI to profit, why is this the case (there are hundreds of articles like this, just one example):

MIT report: 95% of generative AI pilots at companies are failing | Fortune

There’s a stark difference in success rates between companies that purchase AI tools from vendors and those that build them internally.

fortune.com

 

[rdenney]

I don’t think the general concern is necessarily related to job loss, though that is certainly one concern. Lots of folks I know who are deeply concerned aren’t even in the job market and lots of proponents are in the job categories seemingly likely to be affected.

I think the general concern has to do with people delegating (or coercively losing) their agency to implacably non-moral software in ways that could affect their safety, privacy, or freedoms.

A lot of idealism seems to recur as a theme for the proponents. But I’m a little too much of a Calvinist to put much faith in human idealism

Rick “personal agency requires vigilant protection—always someone will want to take it away ‘for your own good’” Denney

 

[pablolie]

I don’t think the general concern is necessarily related to job loss, though that is certainly one concern. Lots of folks I know who are deeply concerned aren’t even in the job market and lots of proponents are in the job categories seemingly likely to be affected.

I think the general concern has to do with people delegating (or coercively losing) their agency to implacably non-moral software in ways that could affect their safety or their freedoms.

A lot of idealism seems to recur as a theme for the proponents. But I’m a little too much of a Calvinist to put much faith in human idealism

Rick “personal agency requires vigilant protection—always someone will to take it away ‘for your own good’” Denney

True all of it, but some of it must also be because people -even when they don't admit it- know that market economy -left to itself- favors productivity over natural balance (which include human considerations like ethics etc)... it's left to the human overseeers of market economy (and hence AI companies) to orchestrate things to include other elements in the balance.

I work in high tech, so invariably now it's part of my job to evangelize the advantages of our AI product suite too. A lot of it has been done forever in high tech, we have always talked endlessly about reducing complexity, making things easier to use and what not... which incidentally has always meant that you need fewer people to keep your IT infrastructure running.

What's a bit new to me in AI is that some of the potential use cases we are told to always evangelize as we talk to exisiting and potential customers kinda don't make any sense to me. I'll give you one example... in retail, one of the supposed killer use cases is video-based checkouts: a consumer walks into a store with many cameras. The cameras identify people coming in via face recognition and -I assume- WiFi device or NFT. The system knows the payment info of the people it identifies. Customers can hence just grab something and walk out without standing in line anywhere to pay. Great. Only... do *I* want cameras to recognize my face everywhere I go? Do I want that automatic link between face and payment info?

My personal take is we are in a bubble (duh), and the positive and negative of real AI deployment as it exists... hasn't yet been truly distilled. It's the phase of throwing spaghetti at the wall and see what sticks. This is typical part of the tech hype cycle, which in turn is just a variation of "Crossing the Chasm" and stuff like that. We don't yet have an idea of where the chips will fall with AI, just like we didn't with tulips or dot.com-stuff or cloud tech or big data etc etc. What we know is that
(a) we're definitely in a bubble, so don't invest your kids' college fund in Nvidia stock right now
(b) it's here to stay in some shape or form
(c) but it'll crash some before it emerges as a truly useful tool/technology
(d) we're nowhere near self-aware AI breakthroughs, that would require fundamentally new neural models

 

[Salt]

Thanks for (a) through (d), appreciated

 

[antcollinet]

Only... do *I* want cameras to recognize my face everywhere I go? Do I want that automatic link between face and payment info?

And - are they really never going to have false positives. Seems ludicrous to expect that - even people don't recognise people reliably enough to automate taking money from them.

 

[PristineSound]

I will freely admit that in the first few years of ChatGPT, I wouldn't even give it any time of the day because I am not a fan of early technology adoption. It costs you more work and you are nothing but a beta tester. About 10 months to a year ago ago, I really started to use Copilot and ChatGPT, it was a revelation, it reminded me how I lived my life before and after the Internet.

Since, I can't go through one day without the use of AI, except when I am on vacation. As AI gets more and more sophisticated, it will be a defining moment of human achievement and technological advancement.

For anyone out there that haven't embraced AI yet, you risk being the elderly person who still used a dumb phone.

 

[rdenney]

And - are they really never going to have false positives. Seems ludicrous to expect that - even people don't recognise people reliably enough to automate taking money from them.

The Civil Rights Lawyer (YT channel) has already posted two videos about arrests of innocent people cause by facial recognition misidentification. In both cases, the LEOs took the computer match at face value (so to speak) with any additional verification (and both victims had an abundance of exculpatory identification), resulting in a false arrest and in one case extended incarceration. That indicates the flip side to the desire for increased productivity—laziness and willingness to accept the proffered AI product without verification. That has been a repeating theme in this long thread, of course.

Dr. Evil may be less scary than either Dr. Lazy or Dr. Credulously Apathetic.

Rick “no such thing as the easy button” Denney

 

[rdenney]

I will freely admit that in the first few years of ChatGPT, I wouldn't even give it any time of the day because I am not a fan of early technology adoption. It costs you more work and you are nothing but a beta tester. About 10 months to a year ago ago, I really started to use Copilot and ChatGPT, it was a revelation, it reminded me how I lived my life before and after the Internet.

Since, I can't go through one day without the use of AI, except when I am on vacation. As AI gets more and more sophisticated, it will be a defining moment of human achievement and technological advancement.

For anyone out there that haven't embraced AI yet, you risk being the elderly person who still used a dumb phone.

Being excited about its positive use does not obviate concern about its inevitable misuse.

Rick “true for most power tools” Denney

 

[PristineSound]

Being excited about its positive use does not obviate concern about its inevitable misuse.

Rick “true for most power tools” Denney

Agree. But to use the Internet as my earlier example, think about all the internet scammers, all the phishing, hacking, malware, ransomware.

Let's use aviation, it changed how warfare is conducted.

The same idea can be applied to medicine, think CRISPR gene editing.

The list goes on and on.

For me personally, I can't think of any time in history where I wouldn't want to be living in the more advanced times.

 

[Timcognito]

For anyone out there that haven't embraced AI yet, you risk being the elderly person who still used a dumb phone.

I think AI will liberate old people "Siri drive me to shop in town the that has the latest delivery of Pacific Salmon". People will get dumber and lazier just what the calculator did to math and ChapGPT is doing to writing.

https://www.sciencedirect.com/science/article/pii/S2590291125010186?via%3Dihub

 

[pablolie]

I think AI will liberate old people "Siri drive me to shop in town the that has the latest delivery of Pacific Salmon". People will get dumber and lazier just what the calculator did to math and ChapGPT is doing to writing.

https://www.sciencedirect.com/science/article/pii/S2590291125010186?via%3Dihub

Which in turn may provide those with a more creative and more free thinking mind with an advantage.

As an example, I seldom trust Yelp reviews for restaurants, and nor do I remotely trust the Michelin guide. One takes you to hyper popular silly places (the latest $30 Ramen place that tastes the exact same) while the other guides you to some pretentious BS place... I like to discover my fav stuff by myself and never trust anyone else entirely.

I have used ChatGPT for writing (not here) ... but only after I have created a positioning document it can not deviate from, and with very directive prompting every time. And lots of back and forth... and then I may rewrite most of it and feed it back as a new reference.

AI doesn't create. It does what we used to call "innovation through reassembly" in Web 2.0 days at much bigger scale.

 

[PristineSound]

People will get dumber and lazier just what the calculator did to math and ChapGPT is doing to writing.

Yet, we are getting smarter and more advanced as a civilization it's a paradox.

 

[pablolie]

Yet, we are getting smarter and more advanced as a civilization it's a paradox.

Not really. We're still apes wielding twig technology.

Douglas Adams and Mark Carwardine "LAST CHANCE TO SEE' TWIG TECHNOLOGY translation to English | Lingualeo https://share.google/x5yn0OlvfdaybjUO2

 

[Timcognito]

Which in turn may provide those with a more creative and more free thinking mind with an advantage.

As an example, I seldom trust Yelp reviews for restaurants, and nor do I remotely trust the Michelin guide. One takes you to hyper popular silly places (the latest $30 Ramen place that tastes the exact same) while the other guides you to some pretentious BS place... I like to discover my fav stuff by myself and never trust anyone else entirely.

I have used ChatGPT for writing (not here) ... but only after I have created a positioning document it can not deviate from, and with very directive prompting every time. And lots of back and forth... and then I may rewrite most of it and feed it back as a new reference.

AI doesn't create. It does what we used to call "innovation through reassembly" in Web 2.0 days at much bigger scale.

All of us will take advantage of AI but most use it like a smart phone to simplify life's challenges and pleasures rather than actively pursuing them. I use a lot of power tools in my woodwork for speed and accuracy but my hand tools give more pleasure and understanding of the material if I use them. I learned Solid Works CAD but find I am more creative if I start by a hand drawing first. The study I posted ends with this

Future teaching strategies should aim to harness the benefits of AI without sacrificing the cognitive engagement and productive struggle required for durable learning. This means using AI to complement, not replace, challenging learning activities, avoiding scenarios where the AI becomes a mere cognitive crutch. In the age of AI, the core principles of human learning are not outdated; in fact, they are more important than ever to uphold.

Yet, we are getting smarter and more advanced as a civilization it's a paradox.

That's correct now we can shoot down a $20k drone with $500k missile. Pablo stated it above, some of ethos and ethics are going out with the bathwater.

 

[PristineSound]

Not really. We're still apes wielding twig technology.

Douglas Adams and Mark Carwardine "LAST CHANCE TO SEE' TWIG TECHNOLOGY translation to English | Lingualeo https://share.google/x5yn0OlvfdaybjUO2

Compared to an advanced alien civilization, yes, we are. Compared to even just 5 decades ago, there is no way I would want to live 5 decades ago knowing what we have today. Just the advances in medicine alone is enough reason to live in preset or future time. We found a gene therapy (2 actually) that completely cures sickle cell permanently with one treatment as an example.

 

[Somafunk]

For anyone out there that haven't embraced AI yet, you risk being the elderly person who still used a dumb phone.

It offers absolutely no possible benefit for myself in my life, that’s why I have no use nor interest in it, apart from allowing the utter dicks like Altman and his ilk to boil ma piss with their bullshit and data centre bollox - that can get in the fu**ing sea


I have an iPhone 17 but have the ai stuff switched off, I have an Viwoods Aipaper reader (phone sized e-ink reader) with multiple ai bots included but have no use for them as I just use it for reading

I live in the Scottish countryside, lived/worked in outdoors manual jobs in this area, was a mtb guide/skill coach throughout uk/europe from 2000 to 2013 when I had to quit due to leg/muscle weakness (didn’t know it was ms). Now stuck on my arse at home with SPMS so at the age of 50 I doubt my last couple years will be aided by a.i. - if it can roll me a joint I’m all in though,

Unless it comes up with a cure for SPMS/demyelination/brain damage then I’m not fussed about it

 

[pablolie]

Yet, we are getting smarter and more advanced as a civilization it's a paradox.

To me the paradox is we're predictably heading into a world where the drive for further productivity and the drive for higher collective quality of life do not longer go hand in hand... but collide and conflict.

I am not sure if that's going to happen in my lifetime, but clearly there will be a problem.

Sure producing and selling more goods at ever lower development and manufacturing seems a formula that has been working well (albeit not always)... but the problem with constant linear improvements is that even mathematically they inevitably end up looping the loop. You can't grow forever in any environment with limited resources... and if you keep disenfranchising people you will run out of customers that can buy your stuff eventually. The model of an exclusively AI led economy doesn't seem in any way sustainable or in any way logically solid.

As some of us train AI models... shouldn't we be compensated for their automated output of our input? Doesn't that mean my intellectual property is churning out revenue far more than ever before? Shouldn't my AI model count as me still. working for whatever company I created it for? Sure patents did some of it, but AI is operationally profitable, while IP is kind of a one off licensing stuff for patents that do not evolve or change.

When manufacturing automation is where we have arrived at... shouldn't manufacturing facilities pay a tax or other community contributions to help pay for the inevitable unemployment created in the area?

I personally find it bizarre when companies lay off people yet claim it's all good because they are investing 50B in AI Infrastructure... check Oracle for one.

 

[JustAnandaDourEyedDude]

Only... do *I* want cameras to recognize my face everywhere I go?

in ways that could affect their safety, privacy, or freedoms

Being excited about its positive use does not obviate concern about its inevitable misuse.

For anyone out there that haven't embraced AI yet

Some companies are embracing AI in more ways than one. In one approach, some companies are revisiting the algorithmic pricing and dynamic pricing and personalized pricing that Instacart tried, this time with AI assistance.

In many countries, consumers have gotten used to prices on commodities and other goods being relatively slow-changing and also being the same for all consumers in a store or e-commerce website at any given time. This is not true in all countries: for example tourists in some places are sometimes faced with tourist prices higher than what the locals pay. I watched a YT video a while ago (cannot remember the title or channel) in which the content creator explained that historically the asking price depended on the seller's assessment of your willingness to pay (WTP).

I watched a recently-posted YT video in which a local TV reporter said that Walmart is planning to introduce algorithmic pricing in its brick-and-mortar stores. With identification of individual consumers and with information collection about consumers increasing by the day, more companies are experimenting with how AI can help them find the maximum WTP of individual consumers in real time. I imagine driving up to a gas pump and having to pay a higher price per gallon because the pump's AI quizzes my car and learns that I do not have enough gas left in my fuel tank to reach the nearest competing gas station (which would anyway charge me even more). This may or may not be a misuse of AI in the sense of being illegal, but it would violate my expectation of privacy/anonymity/non-discrimination. From now on, I am going to wear tattered old clothes and tousle my hair when I visit stores.

Here is timely discussion of algorithmic pricing by CNN's Clare Duffy and Consumer Reports' Grace Gedye:

 

[pablolie]

Some companies are embracing AI in more ways than one. In one approach, some companies are revisiting the algorithmic pricing and dynamic pricing and personalized pricing that Instacart tried, this time with AI assistance.

In many countries, consumers have gotten used to prices on commodities and other goods being relatively slow-changing and also being the same for all consumers in a store or e-commerce website at any given time. This is not true in all countries: for example tourists in some places are sometimes faced with tourist prices higher than what the locals pay. I watched a YT video a while ago (cannot remember the title or channel) in which the content creator explained that historically the asking price depended on the seller's assessment of your willingness to pay (WTP).

I watched a recently-posted YT video in which a local TV reporter said that Walmart is planning to introduce algorithmic pricing in its brick-and-mortar stores. With identification of individual consumers and with information collection about consumers increasing by the day, more companies are experimenting with how AI can help them find the maximum WTP of individual consumers in real time. I imagine driving up to a gas pump and having to pay a higher price per gallon because the pump's AI quizzes my car and learns that I do not have enough gas left in my fuel tank to reach the nearest competing gas station (which would anyway charge me even more). This may or may not be a misuse of AI in the sense of being illegal, but it would violate my expectation of privacy/anonymity/non-discrimination. From now on, I am going to wear tattered old clothes and tousle my hair when I visit stores.

Here is timely discussion of algorithmic pricing by CNN's Clare Duffy and Consumer Reports' Grace Gedye:

It's illegal to gouge prices based on emergency needs. It's called "profiteering" and I doubt it's a legitimate AI use case But you are right to bring it up, as an example of how it can easily be misused.

 

[PristineSound]

To me the paradox is we're predictably heading into a world where the drive for further productivity and the drive for higher collective quality of life do not longer go hand in hand... but collide and conflict.

I am not sure if that's going to happen in my lifetime, but clearly there will be a problem.

Sure producing and selling more goods at ever lower development and manufacturing seems a formula that has been working well (albeit not always)... but the problem with constant linear improvements is that even mathematically they inevitably end up looping the loop. You can't grow forever in any environment with limited resources... and if you keep disenfranchising people you will run out of customers that can buy your stuff eventually. The model of an exclusively AI led economy doesn't seem in any way sustainable or in any way logically solid.

As some of us train AI models... shouldn't we be compensated for their automated output of our input? Doesn't that mean my intellectual property is churning out revenue far more than ever before? Shouldn't my AI model count as me still. working for whatever company I created it for? Sure patents did some of it, but AI is operationally profitable, while IP is kind of a one off licensing stuff for patents that do not evolve or change.

When manufacturing automation is where we have arrived at... shouldn't manufacturing facilities pay a tax or other community contributions to help pay for the inevitable unemployment created in the area?

I personally find it bizarre when companies lay off people yet claim it's all good because they are investing 50B in AI Infrastructure... check Oracle for one.

There will be short term and medium term pain, that is for sure. Then there will be a complete paradigm shift in our economy. I suspect that all will work less a week and the quality of life will be better. There will be new types of jobs that we can't even imagine about with this new economy.