AI is the answer even though we don't seem to know what the question is. Is it all down to increasing profits? It is so unbelievably short sighted, just on an energy basis alone. All goes to feed the machine. This is not inevitable; it is a conscious choice. This is a terrible form for society to take and many will be left out, left behind, and ultimately culled. There is no way this ends well. Literally killing the goose that lays the golden eggs. We are the eggs. You can't have a consumer based economy when there is no one to consume. It is the shrinking of the world in a way that benefits the wealthy. The rest of us will toil and die.
My goodness this was an excellent read. My disgust with these CEOs grows daily. It seems like the true rebellion is to meet this deluded desire for efficiency with slow, measured, and deliberate quality.
Time to reread zen and the art of motorcycle maintenance.
There is no way this level of job loss without a social safety net will be tolerated by the large majority of human. Hammers up!
Maybe the real AI was the labor laws we abolished along the way.
In all seriousness there are two main categories of economic utopia. One is full employment (like Thomas Moore), the other post scarcity (like Star Trek). If you use techno-futuristic narratives to sell your stuff it’s usually one or the other if you boil it down. It’s quite telling that apparently these narratives aren’t interesting enough and instead we are sold the machine overlord dystopia. Perhaps that’s because the marketing is in fact a play where the AI orgs are the main actors, “business leaders” are the choir chanting “oh woe” and workers are the audience.
Finally, if you look at the winners in this spiel it’s the businesses providing the hardware, infra, energy - they’re being paid real money to line up and power huge server racks. The valuations of the AI orgs themselves are purely speculative - one next gen Chinese open source model and they are dust.
As someone who uses Claude and other AI on the daily... all I can tell AI does is produce better pdfs and help me write MUCH better work instructions so my employees are freed up to actually talk to clients.
AI is not taking over. These idiot CEOs who think it replaces employees fail to see the big picture in their race for "market share".
Also, as someone who has been using AI for 2+ years... it gets dumber with age. I can't even imagine the amount of HUMAN work it takes to fix the LLM
Thank you very much, Brian. The CEOs you're describing seem entrenched in a mindset that views employees merely as means for achieving a corporate goal centered on maximizing short-term shareholder value and relentless technological advancement. This perspective, often seen as inevitable (which it is not), reflects outdated linear-causal thinking rooted in the 'supremacy of short-termism and economic efficiency.' This widespread approach can be simply described as 'massive trivialization' or ‘trivial systems thinking.’
However, I believe this mindset prevents them from recognizing their inability to (more) effectively manage the complexity and unpredictability of technologies like Recursive AI, which they have unleashed upon the world. In my view, they often struggle unsuccessfully to contain the 'complex, nontrivial genie' of AI technologies they unleashed (but I doubt this is even fully possible). Many appear unaware of their own blindness to this issue and its often highly destructive social and ecological implications. However, feel free to correct me if you have a different perspective.
In this context, I also like to reference Heinz von Foerster's core ethical principle: 'I always try to act so as to increase the (total) number of choices' for as many people worldwide as possible, both now and in the future. I am aware, however, that this way of thinking is unfortunately still often po-pooed these days. (My website: freeDPerception.com)
I'm reminded of a quote from Marx's Capital, about how capitalists and workers are formallyequal under bourgeois law, but this conceals their different roles. These CEOs are all saying publicly 'my job is going to be impacted too! We will both/all have to adapt!" Privately, they know that their job is not going to be touched: no one is going to cede board seats to AI agents.
Anyway, here is the quote:
"Something changes, so it seems, in the physiognomy of our dramatis personae upon leaving this sphere of simple circulation or commodity exchange, the sphere on which the vulgar free trader bases his views, concepts, and the standards he uses to judge a society of capital and wage labor. The former money owner now strides ahead as a capitalist, while the owner of the labor-power follows him as his worker, one grinning self-importantly, eager to do business, the other wary and reluctant to continue, like someone who has brought his own hide to the market and now can’t expect anything other than … the tannery."
Tech bosses are indeed very eager for this 'apocalypse', and workers across the board expect to get flayed.
A better approach is to run the companies using AI...ie. AI-CEO. This way you can put all the parameters in to successfully run a company, keep employees happy, maintain profit margin, production rates etc without making dramatic and unrealistic decisions based on hype. Most reasons why companies fail is because of poor leadership decisions.
As a senior software engineer who uses AI daily, I can say with certainty that it is not 10x-ing my job.
It makes writing boilerplate stuff faster and makes it easier to ask certain kinds of specific technical questions. It takes some of the tedium out of code reviews as well.
It’s certainly handy, but it’s not fundamentally making me more productive. Maybe saves me 10-20 minutes a day, which is not nothing but it’s also not a 10x speed up.
I am concerned about people entering the field and feeling like they need to rely too much on AI or else they will fall behind.
And only code or instructions for things that are popular and mainstream where there's enough data to feed into the LLM so it has enough consistent material to work on. Ask it to provide instructions or code for something relatively obscure and all you get is hallucination.
what’s looney about all these execs thinking, is they’re basically in the throes of mania saying, effectively, "I like what Facebook did to the news, can I sign up to have it happen to *my* business???”
the AI platforms *themselves* are sucking all the value out of the business, and these companies are cheering to launch into the void!
As a has been technocrat, now a real entrepreneur I agree with you and there is this ancient, old analogy in Shangri-La to explain it in layman's terms.
If you've learned to use a hammer with a nail really really well then everything starts looking like a nail even if you're about to bang the hell out of a dynamite stick.
These lickorous gluttons, as I like to call them now, have created the biggest cloud of hot air I've seen in my more than four decades as an ICT professional.
As always, deftly written: we need to always consider that it may not be AI perse that has deterministic effects so much as that (as Chiang puts it) it "sharpens the knife blade of capitalism." Still, the possibility that individual humans are losing control (and that those managers really dont have choices) can't be dismissed entirely - especially when technology is deployed in capitalist systems.
I think the other bits of the statement are worth examining too.
Will LLMs speed up curing cancer? Possibly - (especially if you count non-LLM, ML techniques). Are we on the brink of it? No sign of that, especially since the research apparatus that could be using the tools is being dismantled in front of our eyes. (Note he has nothing to say on that.)
Balanced budgets? Sure, if there's 10% growth you could tax it to fix the budget. As above, not sure the reality of who is in power fits that. Efficiencies? See below.
10% growth? So far the main industries we're seeing actual changes in are the code generation and word generation ones. Can this feed through to overall efficiencies (putting aside for the moment questions about how real the improvements in code/word generation are) ? Possibly. Will it happen quickly? Almost certainly no. Will such efficiencies lead to 10% growth? No sign of any such size impact. Don't forget his own organisation hires coders who can show they aren't dependent on his own products.
As always, these statements by me relate to the current moment. It is possible someone will invent/unveil a new "AI" that has much greater capabilities, but right now it looks like we're being sold a future that is not here yet.
This will happen for a lot of people where there’s automation involved, but we will always need human moderators—sometimes CEOs can get their minds inflamed with their own thoughts—I’ve been working with AI since 2015, and lost a job over night training an early model at Google UX on contract. As a red shirt it’s easy to do this. But, I’ve learned over time: there is so much data out there that’s lacking context & meaning making from humans. And, there’s glut.
AI is the answer even though we don't seem to know what the question is. Is it all down to increasing profits? It is so unbelievably short sighted, just on an energy basis alone. All goes to feed the machine. This is not inevitable; it is a conscious choice. This is a terrible form for society to take and many will be left out, left behind, and ultimately culled. There is no way this ends well. Literally killing the goose that lays the golden eggs. We are the eggs. You can't have a consumer based economy when there is no one to consume. It is the shrinking of the world in a way that benefits the wealthy. The rest of us will toil and die.
Well put
My goodness this was an excellent read. My disgust with these CEOs grows daily. It seems like the true rebellion is to meet this deluded desire for efficiency with slow, measured, and deliberate quality.
Time to reread zen and the art of motorcycle maintenance.
There is no way this level of job loss without a social safety net will be tolerated by the large majority of human. Hammers up!
Maybe the real AI was the labor laws we abolished along the way.
In all seriousness there are two main categories of economic utopia. One is full employment (like Thomas Moore), the other post scarcity (like Star Trek). If you use techno-futuristic narratives to sell your stuff it’s usually one or the other if you boil it down. It’s quite telling that apparently these narratives aren’t interesting enough and instead we are sold the machine overlord dystopia. Perhaps that’s because the marketing is in fact a play where the AI orgs are the main actors, “business leaders” are the choir chanting “oh woe” and workers are the audience.
Finally, if you look at the winners in this spiel it’s the businesses providing the hardware, infra, energy - they’re being paid real money to line up and power huge server racks. The valuations of the AI orgs themselves are purely speculative - one next gen Chinese open source model and they are dust.
Mass unemployment will lead to societal collapse. UBI won't solve the problem.
Either let people work and get paid justly or a violent revolution against the tech and ai billionaires will be inevitable.
As someone who uses Claude and other AI on the daily... all I can tell AI does is produce better pdfs and help me write MUCH better work instructions so my employees are freed up to actually talk to clients.
AI is not taking over. These idiot CEOs who think it replaces employees fail to see the big picture in their race for "market share".
Also, as someone who has been using AI for 2+ years... it gets dumber with age. I can't even imagine the amount of HUMAN work it takes to fix the LLM
Thank you very much, Brian. The CEOs you're describing seem entrenched in a mindset that views employees merely as means for achieving a corporate goal centered on maximizing short-term shareholder value and relentless technological advancement. This perspective, often seen as inevitable (which it is not), reflects outdated linear-causal thinking rooted in the 'supremacy of short-termism and economic efficiency.' This widespread approach can be simply described as 'massive trivialization' or ‘trivial systems thinking.’
However, I believe this mindset prevents them from recognizing their inability to (more) effectively manage the complexity and unpredictability of technologies like Recursive AI, which they have unleashed upon the world. In my view, they often struggle unsuccessfully to contain the 'complex, nontrivial genie' of AI technologies they unleashed (but I doubt this is even fully possible). Many appear unaware of their own blindness to this issue and its often highly destructive social and ecological implications. However, feel free to correct me if you have a different perspective.
In this context, I also like to reference Heinz von Foerster's core ethical principle: 'I always try to act so as to increase the (total) number of choices' for as many people worldwide as possible, both now and in the future. I am aware, however, that this way of thinking is unfortunately still often po-pooed these days. (My website: freeDPerception.com)
Thank you again.
I'm reminded of a quote from Marx's Capital, about how capitalists and workers are formallyequal under bourgeois law, but this conceals their different roles. These CEOs are all saying publicly 'my job is going to be impacted too! We will both/all have to adapt!" Privately, they know that their job is not going to be touched: no one is going to cede board seats to AI agents.
Anyway, here is the quote:
"Something changes, so it seems, in the physiognomy of our dramatis personae upon leaving this sphere of simple circulation or commodity exchange, the sphere on which the vulgar free trader bases his views, concepts, and the standards he uses to judge a society of capital and wage labor. The former money owner now strides ahead as a capitalist, while the owner of the labor-power follows him as his worker, one grinning self-importantly, eager to do business, the other wary and reluctant to continue, like someone who has brought his own hide to the market and now can’t expect anything other than … the tannery."
Tech bosses are indeed very eager for this 'apocalypse', and workers across the board expect to get flayed.
You might like the book Fully Automated Luxury Communism by Aaron Bastani. It looks like Marx is still relevant!
A better approach is to run the companies using AI...ie. AI-CEO. This way you can put all the parameters in to successfully run a company, keep employees happy, maintain profit margin, production rates etc without making dramatic and unrealistic decisions based on hype. Most reasons why companies fail is because of poor leadership decisions.
As a senior software engineer who uses AI daily, I can say with certainty that it is not 10x-ing my job.
It makes writing boilerplate stuff faster and makes it easier to ask certain kinds of specific technical questions. It takes some of the tedium out of code reviews as well.
It’s certainly handy, but it’s not fundamentally making me more productive. Maybe saves me 10-20 minutes a day, which is not nothing but it’s also not a 10x speed up.
I am concerned about people entering the field and feeling like they need to rely too much on AI or else they will fall behind.
And only code or instructions for things that are popular and mainstream where there's enough data to feed into the LLM so it has enough consistent material to work on. Ask it to provide instructions or code for something relatively obscure and all you get is hallucination.
what’s looney about all these execs thinking, is they’re basically in the throes of mania saying, effectively, "I like what Facebook did to the news, can I sign up to have it happen to *my* business???”
the AI platforms *themselves* are sucking all the value out of the business, and these companies are cheering to launch into the void!
As a has been technocrat, now a real entrepreneur I agree with you and there is this ancient, old analogy in Shangri-La to explain it in layman's terms.
If you've learned to use a hammer with a nail really really well then everything starts looking like a nail even if you're about to bang the hell out of a dynamite stick.
These lickorous gluttons, as I like to call them now, have created the biggest cloud of hot air I've seen in my more than four decades as an ICT professional.
It's a joke, really! 🤣
Great article, Brian.
As always, deftly written: we need to always consider that it may not be AI perse that has deterministic effects so much as that (as Chiang puts it) it "sharpens the knife blade of capitalism." Still, the possibility that individual humans are losing control (and that those managers really dont have choices) can't be dismissed entirely - especially when technology is deployed in capitalist systems.
I think the other bits of the statement are worth examining too.
Will LLMs speed up curing cancer? Possibly - (especially if you count non-LLM, ML techniques). Are we on the brink of it? No sign of that, especially since the research apparatus that could be using the tools is being dismantled in front of our eyes. (Note he has nothing to say on that.)
Balanced budgets? Sure, if there's 10% growth you could tax it to fix the budget. As above, not sure the reality of who is in power fits that. Efficiencies? See below.
10% growth? So far the main industries we're seeing actual changes in are the code generation and word generation ones. Can this feed through to overall efficiencies (putting aside for the moment questions about how real the improvements in code/word generation are) ? Possibly. Will it happen quickly? Almost certainly no. Will such efficiencies lead to 10% growth? No sign of any such size impact. Don't forget his own organisation hires coders who can show they aren't dependent on his own products.
As always, these statements by me relate to the current moment. It is possible someone will invent/unveil a new "AI" that has much greater capabilities, but right now it looks like we're being sold a future that is not here yet.
This will happen for a lot of people where there’s automation involved, but we will always need human moderators—sometimes CEOs can get their minds inflamed with their own thoughts—I’ve been working with AI since 2015, and lost a job over night training an early model at Google UX on contract. As a red shirt it’s easy to do this. But, I’ve learned over time: there is so much data out there that’s lacking context & meaning making from humans. And, there’s glut.
It's never been a better time to be a plumber 🪠