I look at the same information but take away the opposite lesson.
In our fossil-fuel-powered techno-industrial economy, most "jobs" are already social constructs. A job used to be working 100 hours a week on the factory floor from age fourteen until death, now in the post-Covid era it's twenty hours a week in your own bedroom messing around with spreadsheets. Vast categories of jobs already exist entirely due to bureaucratic fiat, or to facilitate social status games: think of the entire education-industrial complex, the half-dozen people who need to stamp the paperwork (and collect a small fee) for you to buy a house, or the 50k commercial pilots to fly planes perfectly capable of flying themselves.
Even for the job categories ostensibly most affected by large language models, the technology won't eliminate jobs, on net, any more than the coffee maker exterminated the barista selling $5 cups of coffee. Some jobs will be retained or created for regulatory reasons, others will continue or even expand because human-generated output will always be valued over AI spam.
I think I've been reading you a lot longer than that but I don't remember that Colinvaux review. Your summary of it makes it sound a lot like Peter Turchin's theory of elite overproduction, but like 30+ years before Turchin.
If you haven't read it I recommend you check out Philip Bobbitt's The Shield of Achilles. The first half is a very interesting survey of the coevolution of constitutional law and technology (the second half is predictive and in my opinion not worth much) -- of particular relevance here is the relationship between mas conscription and democracy, and after reading it and thinking of the automation trends just then beginning to develop in warfare, I realized that soon the powers that be would simply not need us anymore and democratic privileges, such as they are, would gradually be withdrawn. What you're talking about seems a lot like the "civil" side of that coin to me.
I look at the same information but take away the opposite lesson.
In our fossil-fuel-powered techno-industrial economy, most "jobs" are already social constructs. A job used to be working 100 hours a week on the factory floor from age fourteen until death, now in the post-Covid era it's twenty hours a week in your own bedroom messing around with spreadsheets. Vast categories of jobs already exist entirely due to bureaucratic fiat, or to facilitate social status games: think of the entire education-industrial complex, the half-dozen people who need to stamp the paperwork (and collect a small fee) for you to buy a house, or the 50k commercial pilots to fly planes perfectly capable of flying themselves.
Even for the job categories ostensibly most affected by large language models, the technology won't eliminate jobs, on net, any more than the coffee maker exterminated the barista selling $5 cups of coffee. Some jobs will be retained or created for regulatory reasons, others will continue or even expand because human-generated output will always be valued over AI spam.
I think I've been reading you a lot longer than that but I don't remember that Colinvaux review. Your summary of it makes it sound a lot like Peter Turchin's theory of elite overproduction, but like 30+ years before Turchin.
If you haven't read it I recommend you check out Philip Bobbitt's The Shield of Achilles. The first half is a very interesting survey of the coevolution of constitutional law and technology (the second half is predictive and in my opinion not worth much) -- of particular relevance here is the relationship between mas conscription and democracy, and after reading it and thinking of the automation trends just then beginning to develop in warfare, I realized that soon the powers that be would simply not need us anymore and democratic privileges, such as they are, would gradually be withdrawn. What you're talking about seems a lot like the "civil" side of that coin to me.