There is a fundamental problem with AI that few seem to recognize, or maybe the noise from bots just prevents those that do recognize it from being heard. The problem is that most of our societal systems are based on sequential pipelines over time. Career development is a striking one. You need the entry level work to train the intermediate engineer, you need the intermediate work to train the senior level engineer, etc. Sequential pipelines are special in that nothing in, nothing out (after a lag). There is always people retiring, dying, burning out so that pipeline is always emptying.
As you have noticed, it doesn't matter if the work can't actually be replaced. There are strong arguments that the undecidable nature of some of the problem domains support views that this hype is just a falsehood.
The problem with AI is that in its effects, it creates the illusion of unlimited expert supply; and by extension demand goes to zero. In reality there is still need on both sides, but it degrades economic signaling and communication pathways we all rely on, raising costs paradoxically but predictably. Its not immediate because there is a lag but the whipsaw is already here with regards to IT System Engineers. The risk of collapse becomes greater when concentrated risk occurs in runaway positive feedback systems, because resources are in the hands of the few who don't understand the reality of their choices.
What few think deeply on is that most Intelligent and competent people have options because they keep their options open and are flexible. If they cannot find work because everyone has bought the AI hype of it taking over an entire industry where the decision-makers refuse to hire, they will abandon bad investments and retrain. You go where the work is to put food on the table.
The usual timeframe for this appears to be about 1-3 years. Decisionmakers don't realize that these people by proximity raise other people's performance beyond what they are normally capable of providing slack to pivot or prevent disaster. When they are gone, because its a psychologically sticky decision to leave a decade long career (or more) behind, the overall competency and the ability to find competent hires dramatically drops as brain drain sets in. Comparable events might be the spanish inquisition which forced people to leave an area geographically. Organizational and career based experience as a profession start losing practical knowledge around the 5 year mark, and about 1/4 industry cycles out every 10 years depending on labor shocks. If you don't regularly practice it, you lose the knowledge, and when fewer people have that knowledge it concentrates risk of loss over time.
One could naively blame the decision makers for these things, but ironically they are not at fault either. They simply are stupidly following the signalling of optimizing for profit, which is determined by the banks through privileged non-fractional reserve loans that decouple the need to act.
Based on what I know of systems and economics, the only predictable/closest thing that would describe everything we are seeing, like the chaos, whipsaws, and resource exhaustion cycles (derived from Ponzi stage 3, but also exhibited by gradually worsening shortages) are the Economic Calculation Problem as described by Mises in his works on Socialism.
Markets fail when their function of signaling fails, which can occur when there is cooperation, lack of price discovery, and rent-seeking fueled through slave labor (extracted indirectly via inflation). There is a path to a non-market which fails classically under socialism in short order, where state aparatus entities fueled by runaway money printing (unconstrained), outcompete legitimate business creating moats sieving wealth into few hands and preventing any new business from competing based on cost function constraints and distortion.
That is a positive feedback system that eventually as money loses its core properties, runs away leading to economic crisis or socio-economic collapse. These things don't happen overnight, but there is also no way to stop them once cascading failures have started and these types of problems don't have visibility in advance. If one looks at the history, this started with Social Security in the 50s (which if terminated following WW2 would have been fine), but didn't become irreversible until the 1970s, with the rise of the petrodollar.
All in all, the boomers really screwed their children's future. Resource exhaustion cycles don't end well for the majority.
The talent pipeline point is the one that keeps me up at night. You can’t skip the intermediate years and still produce senior engineers. And the people who can leave, will leave first. That part isn’t theoretical for me, I’ve watched it happen.
I’m less sure about the broader economic framing, but the core observation about what happens when you treat skill development as a cost center you can optimize away in a quarter… that’s real.
There is a fundamental problem with AI that few seem to recognize, or maybe the noise from bots just prevents those that do recognize it from being heard. The problem is that most of our societal systems are based on sequential pipelines over time. Career development is a striking one. You need the entry level work to train the intermediate engineer, you need the intermediate work to train the senior level engineer, etc. Sequential pipelines are special in that nothing in, nothing out (after a lag). There is always people retiring, dying, burning out so that pipeline is always emptying.
As you have noticed, it doesn't matter if the work can't actually be replaced. There are strong arguments that the undecidable nature of some of the problem domains support views that this hype is just a falsehood.
The problem with AI is that in its effects, it creates the illusion of unlimited expert supply; and by extension demand goes to zero. In reality there is still need on both sides, but it degrades economic signaling and communication pathways we all rely on, raising costs paradoxically but predictably. Its not immediate because there is a lag but the whipsaw is already here with regards to IT System Engineers. The risk of collapse becomes greater when concentrated risk occurs in runaway positive feedback systems, because resources are in the hands of the few who don't understand the reality of their choices.
What few think deeply on is that most Intelligent and competent people have options because they keep their options open and are flexible. If they cannot find work because everyone has bought the AI hype of it taking over an entire industry where the decision-makers refuse to hire, they will abandon bad investments and retrain. You go where the work is to put food on the table.
The usual timeframe for this appears to be about 1-3 years. Decisionmakers don't realize that these people by proximity raise other people's performance beyond what they are normally capable of providing slack to pivot or prevent disaster. When they are gone, because its a psychologically sticky decision to leave a decade long career (or more) behind, the overall competency and the ability to find competent hires dramatically drops as brain drain sets in. Comparable events might be the spanish inquisition which forced people to leave an area geographically. Organizational and career based experience as a profession start losing practical knowledge around the 5 year mark, and about 1/4 industry cycles out every 10 years depending on labor shocks. If you don't regularly practice it, you lose the knowledge, and when fewer people have that knowledge it concentrates risk of loss over time.
One could naively blame the decision makers for these things, but ironically they are not at fault either. They simply are stupidly following the signalling of optimizing for profit, which is determined by the banks through privileged non-fractional reserve loans that decouple the need to act.
Based on what I know of systems and economics, the only predictable/closest thing that would describe everything we are seeing, like the chaos, whipsaws, and resource exhaustion cycles (derived from Ponzi stage 3, but also exhibited by gradually worsening shortages) are the Economic Calculation Problem as described by Mises in his works on Socialism.
Markets fail when their function of signaling fails, which can occur when there is cooperation, lack of price discovery, and rent-seeking fueled through slave labor (extracted indirectly via inflation). There is a path to a non-market which fails classically under socialism in short order, where state aparatus entities fueled by runaway money printing (unconstrained), outcompete legitimate business creating moats sieving wealth into few hands and preventing any new business from competing based on cost function constraints and distortion.
That is a positive feedback system that eventually as money loses its core properties, runs away leading to economic crisis or socio-economic collapse. These things don't happen overnight, but there is also no way to stop them once cascading failures have started and these types of problems don't have visibility in advance. If one looks at the history, this started with Social Security in the 50s (which if terminated following WW2 would have been fine), but didn't become irreversible until the 1970s, with the rise of the petrodollar.
All in all, the boomers really screwed their children's future. Resource exhaustion cycles don't end well for the majority.
The talent pipeline point is the one that keeps me up at night. You can’t skip the intermediate years and still produce senior engineers. And the people who can leave, will leave first. That part isn’t theoretical for me, I’ve watched it happen.
I’m less sure about the broader economic framing, but the core observation about what happens when you treat skill development as a cost center you can optimize away in a quarter… that’s real.