I remain extremely skeptical. So far no productivity gains have appeared in the statistics (despite a truly enormous amount of capital expenditure) and there was also that study showing that programmers actually lost productivity (around 20%!) using LLMs when they thought they were gaining it. I don't think any dangers of the technology lie in economics but in its application to surveillance and repression, where its many errors are irrelevant because the repressive state/corporate complex doesn't actually care how many eggs it breaks making a dystopian omelette.
- word salad and image synthesis with which to flood the zone
- pretense, as in AI "buddies" or "dates"
2. In business, AI doesn't have to be very good at a task. It just has to be "good enough", and as Egg correctly points out, consolidation has produced enough monopolies and oligopolies that the bar for "good enough" is extremely low.
3. B misses the point of the paperclip exercise. The issue isn't inundation with paperclips. The issue is the acquisition of more and more capability and power (as in the MPP) in service of whatever goal the system has been given.
4. All that other humans have ever meant to the elites is labor and consumption. They do not see us as human at all. They don't see each other as human either. They compete with each other, holding everything as fair game. The Epstein class. They don't just prefer robots because of cost savings. They prefer robots because it further facilitates objectification of all of humanity.
The unraveling will come like a thief in the night.
At first you wonder where you put your ladder which you were using to clean out your gutters and left outside to resume cleaning the next weekend. When the next weekend arrives, you are looking for your ladder and realize it is gone. Someone has stolen it. But you don't know when it happened. A dumb example I know, but it happens. You may not even realize how useful something is because you are so comfortable and lack nothing, but others may see things differently. Scratch planting a vegetable garden in the back yard.
It's the unwinding of the miracle, the miracle of the gift of the former plentiful natural resources, fertile top soil, plentiful oceans, coal, oil, etc. bestowed upon humanity. One-time gifts.
The unwinding/unraveling of a miracle of modern industrial civilization achieved in the last two centuries came to me when I encountered a book by an alum from my college. After overcoming so much, it was all taken from her.
The next few generations will have to suffer through the great unwinding.
The New Zealand government has just announced 10% reductions in government organisations over the next 3 years (announced in our annual budget) with AI headlined as how these reductions will be achieved (we’ve already had 7% reductions in the last 2.5 years). Whether NZ has the capability to redesign workflows at that rate to be more efficient with AI embedded and fewer people is another question. But here comes the recession...
You mean the 1930s Great Depression (which I wasn’t around for!)? Or the rather smaller GFC? Which NZ got through reasonably well. The 1980s sucked big time with 22% interest rates and 12% unemployment. But then things went right for NZ in the 90s to 2000s. So it’s all going to look like downhill from there :(
Certainly could be, I was just a kid during the 80s stock market crash, and still young during the wild inflation of the 90s. But this is worse than the 'great recession', and let's be honest, it never stopped being a recession did it. The rug got pulled up.
There’s a reason most of the AI futurists have built in assumptions about Universal Basic Income. The whole capitalist system immediately breaks down if nobody has any money to spend. This isn’t especially complex. People have been running more realistic thought experiments than the paper clip one for quite a while.
Of course, we won’t be getting that UBI. But we’re still all-gas, no-brakes with the AI push, so what you describe is quite likely.
I’m still vacillating between being sure I’ll be replaced by AI in the next 12 months to cackling about how stupid and worthless it is. It can really surprise you in both directions.
The only thing I’m 100% sure of is that when companies feel confident they can replace us without too much of a drop in accuracy/customer service/etc. they will. “It will open up other avenues and create new jobs…” is a common thought. There might be some truth to it, but timing matters. Tipping points matter.
In my particular industry, the customer is enormous, public retirement plans. There’s some pretty stiff competition to acquire them and they know it. Our largest client actually has stipulations in their agreements with us that we CANNOT use AI in any part of our service for them.
For now.
But it could just turn into a bargaining chip in the future. “Fine, you can use AI agents for x,y, & z, but we want a discount.”
Right now, I know that big business thinks they’re just going to pocket the savings, but if the clients of these businesses are savvy, they can make sure that doesn’t happen.
I also think the public backlash against AI will open up marketing opportunities for any businesses who choose to eschew it in favor of a human workforce. Kind of like buying organic food or whatever. Buying American, etc.
Most companies these days clearly don't give a rat's ass about customer service, largely because people have nowhere else to go: thanks to the rollups that have been going nonstop for a couple decades now, competition for your dollars is largely a thing of the past.
Don't be too sure about UBI being a non-starter. Remember that Trump started the Covid stimulus checks and gave us two. Biden only gave us one and he was just following Trump's lead.
The problem is that the people in positions of power are very much interested in hoovering up all the wealth for themselves. Short sighted? Obviously, but quarterly earnings reports have really messed up our brains.
And if we want to indulge in some dark conspiratorial thinking, it’s probably a mistake to think the oligarchs don’t understand the polycrisis. A global population reset so they can begin round two of their real life version of the game Civilization might be exactly what they want. Can’t have 10 billion gobblers scarfing down all their food and breathing up the rich man’s air, can we?
The world is reaping the benefits of a cultish worship of our psychotic economy and its byproduct, unchecked growth. The emergence of AI may or may not hasten the final breath of the perverted form of capitalism we’ve bowed down to. We don’t need more wealth we need to develop a sustainable program of wealth distribution. Instead of looking toward productivity gains quarter over quarter over quarter and on and on we must strive to create an economy that makes sure everyone is included in society. Let’s get AI to work figuring that out and close the book on this era of predatory, pernicious greed.
Destroying the planetary ecosystem that we depend on to live, simply because our economy must always "grow": this is an example of the kind of goal-oriented tunnel vision that the paperclip metaphor is about. Funny, we didn't need AI to doom ourselves exactly that way.
I am now reading Sunil Amrith’s book “The Burning Earth: An environmental history of the last 500 years.” As someone who reads and writes about environmental history, I’ve been stunned by the longtime exploitation of the many by the powerful few. Much of what we call “progress” has been accompanied by destruction. That’s obvious. Yet, the use of slavery and forced labor at such pervasive, persistent and inhumane levels makes me think that these atrocities are feature, not a bug. At some point, the hyper-wealthy few will need someone to serve their needs. Could that mean the return of work camps and forced labor of the formerly employed?
America’s Homeland Security/Defence/War , driven by the wealthy one percenter tech bros are the ones pushing AI technology, and as America has the highest number of firearms held by private individuals (there being estimated to be roughly 120.5 guns for every 100 people) I think what you’ve described is highly unlikely to happen. However having said that, this could/will happen, if they rapidly regress to a 17th century feudal lifestyle, most probably along with the rest of the world.
But how could that possibly happen, well to understand that I recommend you read two books, the first as a primer, Alice J Friedemann’s (nonfiction) book “When Trucks Stop Running” swiftly followed by Kurt Dahl’s (facts woven into fiction) “An American Famine”. What follows is a short synopsis of mine:
As America's fragile empire crumbles under the weight of insatiable consumption and ecological ruin, “An American Famine” unleashes a nightmarish chronicle of societal disintegration that claws at the soul. What begins as an insidious downward spiral. Supply chains fracturing, farmlands withering into dust, and cities swelling with the desperate and the destitute; escalates into a cataclysmic year, where the grid fails, marauding gangs descend upon the starving masses, and the veneer of civilisation shreds to reveal the primal beast within humanity. Through a mosaic of shattered voices; widowed mothers clawing through rubble for scraps, once respectable men turned cannibals in the flickering light of burning suburbs, and hollow-eyed survivors bartering their last shreds of humanity. Kurt Dahl paints a harrowing tableau of anarchy's embrace, where famine doesn't just empty bellies but devours the very essence of the American dream, leaving only echoes of screams in the endless barren night.
As you might detect it’s a dark and haunting story, one that’s not for those of a sensitive disposition. The book opens with a quote by Thomas Hobbes one we should all heed, “Hell is the truth seen too late”.
Anyone who believes they could survive in such a dire situation, this book will test whether they truly have what it takes to survive🤔
I think the dark forces who actually control America and the West have a plan, as depicted in the Georgia Guidestones, to drastically reduce the human population...They have ensconced abortion as a right, resulting in 100 million+ babies not born..they used the carefully planned Covid scam to release a deadly "vaccine" with generation spanning mRNA that is now killing babies who weren't inoculated...The ventilator and Remdesvir treatments took a heavy toll on senior citizens, the "useless eaters" they hate, and a financial drag on the State...Efficiency demands that you work hard until you can't, then die as soon as possible...only the billionaire class lives to a very old age, e.g. the Rockefeller's...and these "data centers" will steal massive amounts of power and fresh water, which are already in short supply in the West.....
That sounds extremely plausible. On the other hand, ancient scripture in both the pyramids of Gizeh and the caves of Ryleh predict a peaceful takeover by ancient beings that will stop humanity from destroying itself, getting rid of all human Kings and Emperors in the process . Who to believe now?
As an avid, long time reader of this blog, I am surprised and disappointed that you have come so close to grasping the truth behind the problem but have fallen for the techbro mythology/propaganda of AI supremacy.
The truth of the matter is that the functionality of LLMs-- which are still legendarily buggy-- is little more than a smokescreen by the investment class (the Epstein set et al) to disenfranchise the hordes of salary class (middle class) office fauna just as they have done to the wage class (blue collars) already.
Good point. When Reagan closed the door to blue-collar workers, the so-called "professional class" took no notice. When Bush Junior subjected the US economy to two (2) recessions and the GFR, the white-collar workers suddenly started whining.
I got a good laugh out of this article. Thanks B. But I have been living with soul-crushing uncertainty since 1970, so the paranoia about AI is a little thing. Hardly noticeable.
“Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them.”
The elephant in the room - energy . AI uses vast amounts of energy and that's about to get severely restricted . The human brain uses very little and performs better . AI's are restricted by their programming and are results focussed, which is only good for calculation . When AI's are used for long range predictions they fall short . The other issue is the old GIGO situation . The "training" data contains opinion and assumptions which will contaminate the result . If we were actually paying the true cost for what we received from AI's we would be doing our own research . Since AI's are free , that makes you the product.
in inevitable discussions about AI, one is left with the frightening certainty that a vast majority of people see it as some kind of ''physical entity'' with a capacity for ''physical endeavour''
(AI will think of something---is typical)
i find myself unable to dislodge this mode of thinking
AI makes sense if the hidden paradigm is to allow the 300 to kill circa 8 billion and leave a slave labour army to do the scut work AI cannot do to keep the ruling elite fed and powered by the Earth’s scarce resources. In a world of scarce resources, consumers are a problemmatic competitor, not a market.
I remain extremely skeptical. So far no productivity gains have appeared in the statistics (despite a truly enormous amount of capital expenditure) and there was also that study showing that programmers actually lost productivity (around 20%!) using LLMs when they thought they were gaining it. I don't think any dangers of the technology lie in economics but in its application to surveillance and repression, where its many errors are irrelevant because the repressive state/corporate complex doesn't actually care how many eggs it breaks making a dystopian omelette.
A few thoughts...
1. What is AI especially good at?
- surveillance, either military or domestic
- targeting, whether missiles or ads
- word salad and image synthesis with which to flood the zone
- pretense, as in AI "buddies" or "dates"
2. In business, AI doesn't have to be very good at a task. It just has to be "good enough", and as Egg correctly points out, consolidation has produced enough monopolies and oligopolies that the bar for "good enough" is extremely low.
3. B misses the point of the paperclip exercise. The issue isn't inundation with paperclips. The issue is the acquisition of more and more capability and power (as in the MPP) in service of whatever goal the system has been given.
4. All that other humans have ever meant to the elites is labor and consumption. They do not see us as human at all. They don't see each other as human either. They compete with each other, holding everything as fair game. The Epstein class. They don't just prefer robots because of cost savings. They prefer robots because it further facilitates objectification of all of humanity.
And on that cheerful note, have a nice day.
The unraveling will come like a thief in the night.
At first you wonder where you put your ladder which you were using to clean out your gutters and left outside to resume cleaning the next weekend. When the next weekend arrives, you are looking for your ladder and realize it is gone. Someone has stolen it. But you don't know when it happened. A dumb example I know, but it happens. You may not even realize how useful something is because you are so comfortable and lack nothing, but others may see things differently. Scratch planting a vegetable garden in the back yard.
It's the unwinding of the miracle, the miracle of the gift of the former plentiful natural resources, fertile top soil, plentiful oceans, coal, oil, etc. bestowed upon humanity. One-time gifts.
The unwinding/unraveling of a miracle of modern industrial civilization achieved in the last two centuries came to me when I encountered a book by an alum from my college. After overcoming so much, it was all taken from her.
The next few generations will have to suffer through the great unwinding.
https://www.penguinrandomhouse.com/books/576723/the-unwinding-of-the-miracle-by-julie-yip-williams/
The New Zealand government has just announced 10% reductions in government organisations over the next 3 years (announced in our annual budget) with AI headlined as how these reductions will be achieved (we’ve already had 7% reductions in the last 2.5 years). Whether NZ has the capability to redesign workflows at that rate to be more efficient with AI embedded and fewer people is another question. But here comes the recession...
The recession is already here for NZ... and has been for a couple of years. I don't care what the headline says, it's never been worse in my lifetime.
That could be a comment on the brevity of your lifetime?
You mean the 1930s Great Depression (which I wasn’t around for!)? Or the rather smaller GFC? Which NZ got through reasonably well. The 1980s sucked big time with 22% interest rates and 12% unemployment. But then things went right for NZ in the 90s to 2000s. So it’s all going to look like downhill from there :(
Certainly could be, I was just a kid during the 80s stock market crash, and still young during the wild inflation of the 90s. But this is worse than the 'great recession', and let's be honest, it never stopped being a recession did it. The rug got pulled up.
There’s a reason most of the AI futurists have built in assumptions about Universal Basic Income. The whole capitalist system immediately breaks down if nobody has any money to spend. This isn’t especially complex. People have been running more realistic thought experiments than the paper clip one for quite a while.
Of course, we won’t be getting that UBI. But we’re still all-gas, no-brakes with the AI push, so what you describe is quite likely.
I’m still vacillating between being sure I’ll be replaced by AI in the next 12 months to cackling about how stupid and worthless it is. It can really surprise you in both directions.
The only thing I’m 100% sure of is that when companies feel confident they can replace us without too much of a drop in accuracy/customer service/etc. they will. “It will open up other avenues and create new jobs…” is a common thought. There might be some truth to it, but timing matters. Tipping points matter.
I guess we’ll see.
In my particular industry, the customer is enormous, public retirement plans. There’s some pretty stiff competition to acquire them and they know it. Our largest client actually has stipulations in their agreements with us that we CANNOT use AI in any part of our service for them.
For now.
But it could just turn into a bargaining chip in the future. “Fine, you can use AI agents for x,y, & z, but we want a discount.”
Right now, I know that big business thinks they’re just going to pocket the savings, but if the clients of these businesses are savvy, they can make sure that doesn’t happen.
I also think the public backlash against AI will open up marketing opportunities for any businesses who choose to eschew it in favor of a human workforce. Kind of like buying organic food or whatever. Buying American, etc.
Again, guess we’ll see.
Most companies these days clearly don't give a rat's ass about customer service, largely because people have nowhere else to go: thanks to the rollups that have been going nonstop for a couple decades now, competition for your dollars is largely a thing of the past.
Don't be too sure about UBI being a non-starter. Remember that Trump started the Covid stimulus checks and gave us two. Biden only gave us one and he was just following Trump's lead.
Would LOVE to be wrong about it.
The problem is that the people in positions of power are very much interested in hoovering up all the wealth for themselves. Short sighted? Obviously, but quarterly earnings reports have really messed up our brains.
And if we want to indulge in some dark conspiratorial thinking, it’s probably a mistake to think the oligarchs don’t understand the polycrisis. A global population reset so they can begin round two of their real life version of the game Civilization might be exactly what they want. Can’t have 10 billion gobblers scarfing down all their food and breathing up the rich man’s air, can we?
The world is reaping the benefits of a cultish worship of our psychotic economy and its byproduct, unchecked growth. The emergence of AI may or may not hasten the final breath of the perverted form of capitalism we’ve bowed down to. We don’t need more wealth we need to develop a sustainable program of wealth distribution. Instead of looking toward productivity gains quarter over quarter over quarter and on and on we must strive to create an economy that makes sure everyone is included in society. Let’s get AI to work figuring that out and close the book on this era of predatory, pernicious greed.
Destroying the planetary ecosystem that we depend on to live, simply because our economy must always "grow": this is an example of the kind of goal-oriented tunnel vision that the paperclip metaphor is about. Funny, we didn't need AI to doom ourselves exactly that way.
I have a 5 part series on my Substack that addresses this issue https://brettstoltey.substack.com/p/ai-is-here-now-we-have-to-shape-it?
It's most useful as a military technology. Look at how Israel used it for targeting in the Gaza strip & Iran war.
In war you don't have to be right 99.99966% of the time like in a six sigma manufacturing environment. If you kill the wrong person its no big deal.
I am now reading Sunil Amrith’s book “The Burning Earth: An environmental history of the last 500 years.” As someone who reads and writes about environmental history, I’ve been stunned by the longtime exploitation of the many by the powerful few. Much of what we call “progress” has been accompanied by destruction. That’s obvious. Yet, the use of slavery and forced labor at such pervasive, persistent and inhumane levels makes me think that these atrocities are feature, not a bug. At some point, the hyper-wealthy few will need someone to serve their needs. Could that mean the return of work camps and forced labor of the formerly employed?
America’s Homeland Security/Defence/War , driven by the wealthy one percenter tech bros are the ones pushing AI technology, and as America has the highest number of firearms held by private individuals (there being estimated to be roughly 120.5 guns for every 100 people) I think what you’ve described is highly unlikely to happen. However having said that, this could/will happen, if they rapidly regress to a 17th century feudal lifestyle, most probably along with the rest of the world.
But how could that possibly happen, well to understand that I recommend you read two books, the first as a primer, Alice J Friedemann’s (nonfiction) book “When Trucks Stop Running” swiftly followed by Kurt Dahl’s (facts woven into fiction) “An American Famine”. What follows is a short synopsis of mine:
As America's fragile empire crumbles under the weight of insatiable consumption and ecological ruin, “An American Famine” unleashes a nightmarish chronicle of societal disintegration that claws at the soul. What begins as an insidious downward spiral. Supply chains fracturing, farmlands withering into dust, and cities swelling with the desperate and the destitute; escalates into a cataclysmic year, where the grid fails, marauding gangs descend upon the starving masses, and the veneer of civilisation shreds to reveal the primal beast within humanity. Through a mosaic of shattered voices; widowed mothers clawing through rubble for scraps, once respectable men turned cannibals in the flickering light of burning suburbs, and hollow-eyed survivors bartering their last shreds of humanity. Kurt Dahl paints a harrowing tableau of anarchy's embrace, where famine doesn't just empty bellies but devours the very essence of the American dream, leaving only echoes of screams in the endless barren night.
As you might detect it’s a dark and haunting story, one that’s not for those of a sensitive disposition. The book opens with a quote by Thomas Hobbes one we should all heed, “Hell is the truth seen too late”.
Anyone who believes they could survive in such a dire situation, this book will test whether they truly have what it takes to survive🤔
I think the dark forces who actually control America and the West have a plan, as depicted in the Georgia Guidestones, to drastically reduce the human population...They have ensconced abortion as a right, resulting in 100 million+ babies not born..they used the carefully planned Covid scam to release a deadly "vaccine" with generation spanning mRNA that is now killing babies who weren't inoculated...The ventilator and Remdesvir treatments took a heavy toll on senior citizens, the "useless eaters" they hate, and a financial drag on the State...Efficiency demands that you work hard until you can't, then die as soon as possible...only the billionaire class lives to a very old age, e.g. the Rockefeller's...and these "data centers" will steal massive amounts of power and fresh water, which are already in short supply in the West.....
That sounds extremely plausible. On the other hand, ancient scripture in both the pyramids of Gizeh and the caves of Ryleh predict a peaceful takeover by ancient beings that will stop humanity from destroying itself, getting rid of all human Kings and Emperors in the process . Who to believe now?
As an avid, long time reader of this blog, I am surprised and disappointed that you have come so close to grasping the truth behind the problem but have fallen for the techbro mythology/propaganda of AI supremacy.
The truth of the matter is that the functionality of LLMs-- which are still legendarily buggy-- is little more than a smokescreen by the investment class (the Epstein set et al) to disenfranchise the hordes of salary class (middle class) office fauna just as they have done to the wage class (blue collars) already.
Good point. When Reagan closed the door to blue-collar workers, the so-called "professional class" took no notice. When Bush Junior subjected the US economy to two (2) recessions and the GFR, the white-collar workers suddenly started whining.
I got a good laugh out of this article. Thanks B. But I have been living with soul-crushing uncertainty since 1970, so the paranoia about AI is a little thing. Hardly noticeable.
“Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them.”
Dune, Frank Herbert
The elephant in the room - energy . AI uses vast amounts of energy and that's about to get severely restricted . The human brain uses very little and performs better . AI's are restricted by their programming and are results focussed, which is only good for calculation . When AI's are used for long range predictions they fall short . The other issue is the old GIGO situation . The "training" data contains opinion and assumptions which will contaminate the result . If we were actually paying the true cost for what we received from AI's we would be doing our own research . Since AI's are free , that makes you the product.
You had me at: "the real danger in implementing AI: the utter pointlessness of it all"
in inevitable discussions about AI, one is left with the frightening certainty that a vast majority of people see it as some kind of ''physical entity'' with a capacity for ''physical endeavour''
(AI will think of something---is typical)
i find myself unable to dislodge this mode of thinking
AI makes sense if the hidden paradigm is to allow the 300 to kill circa 8 billion and leave a slave labour army to do the scut work AI cannot do to keep the ruling elite fed and powered by the Earth’s scarce resources. In a world of scarce resources, consumers are a problemmatic competitor, not a market.