A labor market in rough balance is a strange place to look for an AI-driven jobs collapse. The full-time typist has disappeared, but very few jobs are like that—fully automatable by “AI”…
"The full-time typist has disappeared" - never seen any statistic to demonstrate that this "automation" actually brought any cost benefits.
In ancient history, I was teamed with a secretary/typist who did my typing, managed my email triage and my diary. She and I both worked a 7-hour day and usually finished all the work. I earned around 3x her wages. So, everything got done for 4x "unit wage".
In less ancient history - I designed complex systems, documented them and wrote long technical documents and reports. I did my own typing. I never finished all the work, or achieved the 180wpm single pass typing efficiency of my secretary.
Instead, I constantly rescheduled my workload to make time for typing, correction, formatting, triage and diary and my work throughput dropped by around 50%. Fortunately, so did that of all my peers and colleagues, for the same reason, and the reduced efficiency went unnoticed while everything cost 6x "unit wage" x1.50% (rescheduling allowance).
The customer bore the cost increase, the business services bore the efficiency drop and lots of expensive workers spent lots of expensive time maintaining their skills at using word processors and the IT department budget rocketed. Someone managed to persuade the CFO that this was cost effective. Perhaps he just liked playing with word processors.
The automation bit should perhaps have stopped at teaching the typists to use word processors, because that's where the efficiency gains stopped.
AI strikes me as being much the same. Lots of expensive specialist will waste lots of time learning to construct AI prompts that work and to use the automated prompters that will follow and someone will persuade the CFO that it was a smart move to fire the graduates who were really good at AI prompting and tool use at 1/4 the cost.
Back in the day, executive secretaries were very much like sergeant majors in the British army of the 1800s: people who were vastly more qualified and effective than their putative bosses, but were ring-fenced out of the higher titular positions because of their sex (in the case of U.S. executive) or because of their class (in the case of the British army— due to the post-Cromwell rule that only landlords' close relatives were allowed to command soldiers). Those days were going even before the advent of the personal computer. That, at least, is how I see it. I may be wrong.
I was in reserve army boot camp in the 50s—like the Y locker room only less hazardous. It was amusing to watch the grizzled old sergeants pretend to defer to the ROTC secound looies while in fact running the whole show. Worked best when both signs knew it was a charade.
:-) :: I remember what my father told me about being at OCS at Quantico: "Your job is to do what your sergeant tells you to do, and look brave for as long as you live".
True, but it doesn't alter the fact that swathes of non-executive secretaries who actually just typed, managed diaries and fielded calls for tier-4 managers and professionals were classified as redundant whilst the workload they supported was persistently rescheduled to make time for tier-4 workers to play with PowerPoint and Word at 3x the cost and 50% of the throughput.
They answered the phone and took messages and typed my reports and papers and made all my travel arrangements. Now will ai do what they did back in the before times before we all became sysadmins to our computers? Hard to see them as much fun as flesh and blood women. I miss flirting with them as they were young or not young and beautiful. I miss knowing about their lives, husbands and children. I miss sending them Christmas cards. Congratulating them on the birth or graduation of a child or in some cases a grandchild.
I view missing them as another example of the enshittification of the job environment.
We are herd animals so I don't see ai assistants being part of the herd.
I understand. The ambiance of office work has changed since we could type our own stuff & send it out electronically, instead of dictate & send it out hard copy. At the same time, we now get more done in less time than before. The real damage has been done by voice mail (whatever it's called) when you need help from some institution & you have to negotiate that computer driven "assistant."
Fifty-five years ago I was a senior advisor and action officer on the technology side of the Office of the Secretary of Defense, supported by a deputy, a staff of 10-12 engineering professionals and a terrific executive assistant (who really helped me look good), with two or three full-time typists. The engineers produced a lot of papers and reports, written out in longhand, which the typists turned into final form under the guidance of the executive assistant. First I got everyone a typewriter and convinced them they could type faster even with two fingers than they could write, and turn out better draft copy in the process, resulting in small but nontrivial increment in volume and quality of output, as well as freeing more time for research. Then I managed to get Wangs for the executive assistant and typists, bringing added gains. Finally, in a coup, I got the new wonders, PCs, for every desk, allowing me (among other things) to repurpose the typists as research assistants. To integrate the hardware into an effective system took a certain amount of management innovation and effort but the improvements in quality and quantity were very perceptible.
There were several other offices I knew about where similar hardware did not bring comparable results. The difference was in management. I had been a teen computer nerd before there were computers (outside of a very few experimental machines), had my own (very primitive) machine at home, and had spent fifteen years imagining how "datamation" might be used; the people running the other offices had not.
I would guess that there is a certain sprinkling of people who are mentally prepared to make effective use of AI in their fields, but that in most cases it will take a lot of (often unwelcome) learning and/or replacement of leaders to get truly good use out of the technology.
"Prestige" rather than "command" as a coordination system for East African Plains Apes: "I am an expert and this is how we get s**** done here" works in lots of cases when "this is how we do this because I am the boss" does not work well:
> **Will O'Neil**: Fifty-five years ago I was a senior advisor and action officer on the technology side of the Office of the Secretary of Defense, supported by a deputy, a staff of 10-12 engineering professionals and a terrific executive assistant (who really helped me look good), with two or three full-time typists.
> The engineers produced a lot of papers and reports, written out in longhand, which the typists turned into final form under the guidance of the executive assistant. First I got everyone a typewriter and convinced them they could type faster even with two fingers than they could write, and turn out better draft copy in the process, resulting in small but nontrivial increment in volume and quality of output, as well as freeing more time for research. Then I managed to get Wangs for the executive assistant and typists, bringing added gains.
> Finally, in a coup, I got the new wonders, PCs, for every desk, allowing me (among other things) to repurpose the typists as research assistants. To integrate the hardware into an effective system took a certain amount of management innovation and effort but the improvements in quality and quantity were very perceptible.
> There were several other offices I knew about where similar hardware did not bring comparable results. The difference was in management. I had been a teen computer nerd before there were computers (outside of a very few experimental machines), had my own (very primitive) machine at home, and had spent fifteen years imagining how "datamation" might be used; the people running the other offices had not.
> I would guess that there is a certain sprinkling of people who are mentally prepared to make effective use of AI in their fields, but that in most cases it will take a lot of (often unwelcome) learning and/or replacement of leaders to get truly good use out of the technology.
I do think a few formerly solid professions, may be starting a junior level jobs crisis. Particularly software. Senior staff can use AI as reasonably competent assistants. In the short term this will make it tough for recent garduates. LOnger term software related jobs might start to recover on account of the Jevon's paradox. This states that as an economic activity becomes cheaper, demand for it will rise enough that the total amount of labor will either stabilize or even increase.
Yes—a "junior-level jobs crisis" is a possible thing. But the junior-level jobs that actually let you do rather than just watch while you make coffee and such are the ones that are lest likely to vanish. And those might be the valuable ones:
> **Thomas Hewitt**: I do think a few formerly solid professions, may be starting a junior level jobs crisis. Particularly software. Senior staff can use AI as reasonably competent assistants. In the short term this will make it tough for recent garduates. LOnger term software related jobs might start to recover on account of the Jevon's paradox. This states that as an economic activity becomes cheaper, demand for it will rise enough that the total amount of labor will either stabilize or even increase.
"The full-time typist has disappeared" - never seen any statistic to demonstrate that this "automation" actually brought any cost benefits.
In ancient history, I was teamed with a secretary/typist who did my typing, managed my email triage and my diary. She and I both worked a 7-hour day and usually finished all the work. I earned around 3x her wages. So, everything got done for 4x "unit wage".
In less ancient history - I designed complex systems, documented them and wrote long technical documents and reports. I did my own typing. I never finished all the work, or achieved the 180wpm single pass typing efficiency of my secretary.
Instead, I constantly rescheduled my workload to make time for typing, correction, formatting, triage and diary and my work throughput dropped by around 50%. Fortunately, so did that of all my peers and colleagues, for the same reason, and the reduced efficiency went unnoticed while everything cost 6x "unit wage" x1.50% (rescheduling allowance).
The customer bore the cost increase, the business services bore the efficiency drop and lots of expensive workers spent lots of expensive time maintaining their skills at using word processors and the IT department budget rocketed. Someone managed to persuade the CFO that this was cost effective. Perhaps he just liked playing with word processors.
The automation bit should perhaps have stopped at teaching the typists to use word processors, because that's where the efficiency gains stopped.
AI strikes me as being much the same. Lots of expensive specialist will waste lots of time learning to construct AI prompts that work and to use the automated prompters that will follow and someone will persuade the CFO that it was a smart move to fire the graduates who were really good at AI prompting and tool use at 1/4 the cost.
Back in the day, executive secretaries were very much like sergeant majors in the British army of the 1800s: people who were vastly more qualified and effective than their putative bosses, but were ring-fenced out of the higher titular positions because of their sex (in the case of U.S. executive) or because of their class (in the case of the British army— due to the post-Cromwell rule that only landlords' close relatives were allowed to command soldiers). Those days were going even before the advent of the personal computer. That, at least, is how I see it. I may be wrong.
I was in reserve army boot camp in the 50s—like the Y locker room only less hazardous. It was amusing to watch the grizzled old sergeants pretend to defer to the ROTC secound looies while in fact running the whole show. Worked best when both signs knew it was a charade.
:-) :: I remember what my father told me about being at OCS at Quantico: "Your job is to do what your sergeant tells you to do, and look brave for as long as you live".
True, but it doesn't alter the fact that swathes of non-executive secretaries who actually just typed, managed diaries and fielded calls for tier-4 managers and professionals were classified as redundant whilst the workload they supported was persistently rescheduled to make time for tier-4 workers to play with PowerPoint and Word at 3x the cost and 50% of the throughput.
Damn I miss the secretaries!!!
They answered the phone and took messages and typed my reports and papers and made all my travel arrangements. Now will ai do what they did back in the before times before we all became sysadmins to our computers? Hard to see them as much fun as flesh and blood women. I miss flirting with them as they were young or not young and beautiful. I miss knowing about their lives, husbands and children. I miss sending them Christmas cards. Congratulating them on the birth or graduation of a child or in some cases a grandchild.
I view missing them as another example of the enshittification of the job environment.
We are herd animals so I don't see ai assistants being part of the herd.
The "herd animals" part of this is very important, I think—which is why we go to coffeeshops with our laptops. Or I do, at least.
I understand. The ambiance of office work has changed since we could type our own stuff & send it out electronically, instead of dictate & send it out hard copy. At the same time, we now get more done in less time than before. The real damage has been done by voice mail (whatever it's called) when you need help from some institution & you have to negotiate that computer driven "assistant."
Fifty-five years ago I was a senior advisor and action officer on the technology side of the Office of the Secretary of Defense, supported by a deputy, a staff of 10-12 engineering professionals and a terrific executive assistant (who really helped me look good), with two or three full-time typists. The engineers produced a lot of papers and reports, written out in longhand, which the typists turned into final form under the guidance of the executive assistant. First I got everyone a typewriter and convinced them they could type faster even with two fingers than they could write, and turn out better draft copy in the process, resulting in small but nontrivial increment in volume and quality of output, as well as freeing more time for research. Then I managed to get Wangs for the executive assistant and typists, bringing added gains. Finally, in a coup, I got the new wonders, PCs, for every desk, allowing me (among other things) to repurpose the typists as research assistants. To integrate the hardware into an effective system took a certain amount of management innovation and effort but the improvements in quality and quantity were very perceptible.
There were several other offices I knew about where similar hardware did not bring comparable results. The difference was in management. I had been a teen computer nerd before there were computers (outside of a very few experimental machines), had my own (very primitive) machine at home, and had spent fifteen years imagining how "datamation" might be used; the people running the other offices had not.
I would guess that there is a certain sprinkling of people who are mentally prepared to make effective use of AI in their fields, but that in most cases it will take a lot of (often unwelcome) learning and/or replacement of leaders to get truly good use out of the technology.
"Prestige" rather than "command" as a coordination system for East African Plains Apes: "I am an expert and this is how we get s**** done here" works in lots of cases when "this is how we do this because I am the boss" does not work well:
> **Will O'Neil**: Fifty-five years ago I was a senior advisor and action officer on the technology side of the Office of the Secretary of Defense, supported by a deputy, a staff of 10-12 engineering professionals and a terrific executive assistant (who really helped me look good), with two or three full-time typists.
> The engineers produced a lot of papers and reports, written out in longhand, which the typists turned into final form under the guidance of the executive assistant. First I got everyone a typewriter and convinced them they could type faster even with two fingers than they could write, and turn out better draft copy in the process, resulting in small but nontrivial increment in volume and quality of output, as well as freeing more time for research. Then I managed to get Wangs for the executive assistant and typists, bringing added gains.
> Finally, in a coup, I got the new wonders, PCs, for every desk, allowing me (among other things) to repurpose the typists as research assistants. To integrate the hardware into an effective system took a certain amount of management innovation and effort but the improvements in quality and quantity were very perceptible.
> There were several other offices I knew about where similar hardware did not bring comparable results. The difference was in management. I had been a teen computer nerd before there were computers (outside of a very few experimental machines), had my own (very primitive) machine at home, and had spent fifteen years imagining how "datamation" might be used; the people running the other offices had not.
> I would guess that there is a certain sprinkling of people who are mentally prepared to make effective use of AI in their fields, but that in most cases it will take a lot of (often unwelcome) learning and/or replacement of leaders to get truly good use out of the technology.
You typically need a bit of both, but command without effective persuasion rarely works well in complex tasks, very rarely indeed.
I do think a few formerly solid professions, may be starting a junior level jobs crisis. Particularly software. Senior staff can use AI as reasonably competent assistants. In the short term this will make it tough for recent garduates. LOnger term software related jobs might start to recover on account of the Jevon's paradox. This states that as an economic activity becomes cheaper, demand for it will rise enough that the total amount of labor will either stabilize or even increase.
Yes—a "junior-level jobs crisis" is a possible thing. But the junior-level jobs that actually let you do rather than just watch while you make coffee and such are the ones that are lest likely to vanish. And those might be the valuable ones:
> **Thomas Hewitt**: I do think a few formerly solid professions, may be starting a junior level jobs crisis. Particularly software. Senior staff can use AI as reasonably competent assistants. In the short term this will make it tough for recent garduates. LOnger term software related jobs might start to recover on account of the Jevon's paradox. This states that as an economic activity becomes cheaper, demand for it will rise enough that the total amount of labor will either stabilize or even increase.
======
<https://braddelong.substack.com/p/the-ai-jobs-apocalypse-is-not-in>