Here we go again. The greatest shortage today, as before, is not in our analysis, it is in our collective humanity.
The Nazis collected and kept plenty of data on their subjects, especially their Jewish ones. And they drove all the waste out of their killing machines. They were as tech-forward as anyone.
Drucker saw this in person. He was present at his university the day of the Nazi takeover. His account is chilling to read, here in The Atlantic, or in his autobiography "Adventures of a Bystander," which is all about people, his primary source:
Thanks for this rescue of Drucker, who along with John Kenneth Galbraith, seems to have been flushed from collective memory just when their ideas might help think through the problems of ensuring this new social technology is used to manage organizations rather than control or replace workers.
How strange for Prince to choose Drucker for this exercise! I wonder what prompt led him or his pet writer there.
Not only Drucker and Galbraith, but W. Edwards Deming as well, who probably understood the use of measurement better than nearly anyone writing about management.
There is more than a whiff of "liberating" AI from oversight in this chap Prince's view, rather than the apparent premise of "liberating" businesses from "superfluous" middle management in the interest of "productivity".
He can also be using "Drucker" to justify what he already wanted to do, rather than what Drucker actually wanted him to do. Same as you can always grab a phrase out of the Bible or Orwell to "justify" stuff. At least he acknowledged having to think about it. But it is very hard to get a long existant idea out of one's head.
Liked "That would be one of the points of reading."
Ideas are like the other addictive substances: the person has to want to change. A lot of people went to the meetings only because the court ordered them to.
This is a wonderfully precise distinction: measurement can support judgment, but it cannot replace it. The danger is not that AI measures too little, but that leaders start believing measurement is the same thing as management.
It's remarkable that this guy absolutely swears that he's put a lot of thought and reading into this decision, and then makes these huge sweeping generalizations that belie a gaping incuriousness.
He obviously has spent no time with his middle managers, understanding what they actually do or how they make the business succeed.
The first person we have to manage is ourself. Peter Drucker was someone who knew and did that. If you read his autobiography it is mostly about the interesting people he met. And he seemed to be a magnet for interesting people.
There is no getting around the need to find out what is really going on. In war the Generals have to go to the front lines to find out, because the chain of command will not get the true picture to them. Same problem in companies.
You can think the problem to death, and read all the books, and look at all the numbers on your screen, but you still have to go and find out.
Be where it happens.
Otherwise, to borrow a line from the great comedian, Warren Buffett, when the tide goes out we get to find out who's been swimming naked.
This is where many senior managers in tech companies get stuck, though. They think that because they have to go to the front to find out "what's really going on" that the middle management structure must be broken. That's not the case. The middle managers know about and are balancing many factors that the senior managers don't see. If their priorities are misaligned that's usually the senior managers fault, probably because they are spending too much time micromanaging their most important problem.
It absolutely is true that sometimes senior managers have to focus on some specific problem that's in the span of control of someone it would normally have been delegated to. But that's because shorter lines of communication help with alignment. The problem I've seen again and again is senior management does not even try to get wider alignment and so thinks they need to micromanage everything. The actual military has much better processes, has done for a very long time, and would not fall into this trap.
No argument from me about the danger of micromanagement.
From my experience, the top managers who ignore the middle managers also ignore everyone else. They consider other people to be interchangeable objects to be used. Not surprising, since top managers are so laser focused on data and money today. And they think they are accountable to no one except the march of technology and insatiable Wall Street. They think they are creating a perpetual growth machine.
But the people need a society fit to live in.
Human thriving is not inherent in "optimization" processes. It also got neglected by the "modern" idea in general. This is why people are unhappy with the way the world is today.
One of Drucker's key ideas was that if managers fail to make their organizations serve society the people might eventually become desperate, and turn to people like Hitler as their last chance.
Yes, I think there is an over-optimization problem. Wall street and (even moreso) private equity has encouraged management to expose everyone to the same slightly broken set of incentives, and so while the incentives are only slightly broken - its generally true everyone should try to make money for their employer up to a point - the issues with them get magnified further and further.
Just how freaking out of touch can our ruling class get? Do they all live along the beach in Marin County and get chauffeured to their Gulfstreams so they can dine out in Barcelona? Why don't they just buy California and stay home for dinner? Do a leveraged buyout.
Count this as a victory for optimization: all the SUVs now look exactly the same.
All the tech bros are becoming more and more alike.
Ai is like a new military technology. You only know how good it is when it is used in a peer to peer engagement. Let the battle begin. For those of you still needing a job, dig a deep fox hole.
Every company I've ever worked for - and I've worked for a lot of them - definitely didn't follow Drucker's prescriptions. It was always been, and remains, top down control.
"He views action and responsibility as going together, stressing both that business has a broader social role (profit is no more than a necessary condition to stay in operation, and business must pay attention to the common good)..." Then along came Milton Friedman...
True top down control is very hard to do in any professional organization. Whether its tech or accountancy, you are always delegating huge amounts of responsibilty down ot the most junior levels. Its very dangerous to miss this. What happens when execs don't think seriously about delegation and whether the people being delegated too can do thier jobs is not true top down control but micromanagement.
Have you ever worked in organizations that weren't "professional" like tech or accounting? The situation might be different in those cases. There are a lot of young people graduating high school and college this year who might be heading straight to Starbucks, or sports gambling. These are the people who are really getting the shortest end of the stick right now. A family member who is a software engineer works for a company where the sociopathic CEO was boasting about how many people he was going to be able to get rid of using AI. Later he had to backtrack, but the sheer entitlement of this crop of executives, after decades of having things go their way, is breathtaking to behold. Not that the knives are not out in the headquarters building, where the losers get "transitioned" out of the company, as if it were an act of God doing it. Children playing with knives.
Yes, hence my qualifier. Its hard to get alignment from low paid staff so managers resort to top down control and tolerate high turnover. I much prefer Drucker's vision where hotel maids or Amazon warehouse employees could be largely self-managing, but all of society has to accept higher costs to make that happen.
I don't think AI gets us to the point of professional staff being in quite the same spot. Anyone can tell whether a hotel room was cleaned or a widget was sent to the right address in the "right" amount of time. No-one can tell whether your accounts are audited correctly or your software works to spec without replicating almost the same work needed in the first place. This is the fundamental problem of software engineering management that Fred Brooks identified fifty years ago. AI can produce much more software much faster, but only a trained professional can tell it what to create or check that it created the right thing. Just as ATMs initially increased the demand for bank tellers AI is going to initially increase the demand for engineers. Especially hardware engineers, interestingly.
Which is my long winded way of saying, fuck those tech CEOs, they're going to suffer for their hubris. Fortunately they're not all like that.
You can imagine after a while the demand for human professional labor is going to fall. That the AIs will be well trained enough that one can write the spec, another can sign off on it, a third can write the code and a fourth can test it, with humans only involved in a high level project management perspective. It will be turtles all the way down at that point. But I don't think that's a short term prospect. They're just not reliable enough, but more importantly the pent up demand for software is massive. Every unfixed bug, every little thing you'd like an app for, its all now possible. The most interesting statistic in all of this is that GitHub shows project backlogs clearly shrinking - that's a real change, project backlogs never shrink.
Maybe I'm using these terms in an idiosyncratic way, but when I hear top down control, I think "command and control", which was kind of default management strategy in the Taylorist era, before Drucker et al. Data is reported up the management chain. Commands come down. The only delegated responsibility is to implement the commands. If senior management misunderstands the data and the commands sent down are stupid, no-one can do anything about that. This is what the millitary and then (much later) business worlds tried to move away from in the 20th century, because it produces very bad results, like invading Russia in the winter, the Ford Edsel, and the Ground Nut fiasco.
Micromanagement is when managers step in and try to do the jobs of people lower in the hierarchy for them either directly or indirectly by issuing such detailed instructions the employee is just a robot. It tends not to happen in command and control organizations, because generally there's a strong expectation that managers only direct work that they control, never do it. Rather, micromanagement happens when managers believe their high level commands won't be obeyed, often because command and control has been abandoned and no usable framework has been put in its place. This is the case in many tech companies - everyone knows you're not meant to do command and control, but managers are often very bad at aligning their staff with their goals so they end up getting far too deeply involved in some topics and ignoiring others.
A lot of micromanagement is done by people who used to do the job, maybe did it well, then were turned into managers and couldn't restrain themselves from just jumping in and dictating every detail of the job to get it done the way they would do it. And it is very easy to think there is only one way to do the job: my way.
Been there. Sometimes as the one telling the newbie how it should be done!
There's a big difference between mentoring and micromanaging. Once the trainee knows how it's done, a mentor steps back and let's them go at it. A micromanager keeps putting a stick the their necks.
I've worked for such managers. They don't actually >do< the work, if they do any real work at all. They just stand over your shoulders, breath down your neck and bark commands of exactly what to do and exactly how to do it - even if you know how to do it better than they do.
My brother used to work for such a company, in a call center. Quite a comedown for a former Learjet captain, but that's how things go sometimes. He said it was more stressful than flying the Lear, which if mishandled can turn into a fireball on the runway when the tip tanks hit, or fall out of the sky at 51,000 feet if you go a little bit slow.
Apparently they didn't bother to train their managers, who seemed to think something was wrong if the customers were not equally happy every single day. Half the people in the building had letters in their files putting them on "performance improvement plans," the first step toward firing. Turnover was high, despite good pay.
The managers never learned literally the FIRST LESSON Deming taught in his quality courses, which was that statistical variation was normal, and can't be improved by haranguing employees. The managers thought there should be no variation, even though all of the customers and their problems were unique. They kept asking: "What's wrong?" They wanted the calls to end as fast as possible, and they wanted reps to upsell. They did NOT put the customers first. That's a nice slogan.
Naturally, the managers who were judging and coaching the employees couldn't do the work as well themselves.
TOP DOWN is still the default management approach almost everywhere.
It makes sense when you realize the very top are psychopaths, and they promote people who are like themselves, that is, the psychopaths. They're easy to identify because they're the biggest back stabbers.
What the management class believes in is domination.
Great piece. The “ lurid science fiction” line had me wondering about the relationship between “The Practice…” and “Player Piano” and whether anyone read Player Piano and said “sounds good to me…” — is it torment nexuses all the way down?
So many newsletters I get, and now I have to get this one too, which was recommended by a co-worker regarding a book we're working on.
What I most appreciate about this piece is that it also signals an extension of another metaphor. People like Musk describe people who aren't like Musk as "non-player characters." NPCs, for those who haven't played Dungeons and Dragons for nearly 50 years, are individuals in a roleplaying game who aren't directed by the players of the game. They're the barkeep, the server in the bar with the rumor that starts an adventure, the local who hires a party to do some job. Set aside that Musk sees real life as a roleplaying game, what Musk means is, these people aren't real so they don't matter. While they may have personalities and skills, they can be abused, ignored, laid off, fired, killed whatever. Which is, in real life terms, sociopathic. It also suggests he treats his assistants like garbage.
But "Matthew, Prince of Cloudfare" (genius) takes things a step further. He doesn't see his employees as NPCs. He sees them as summoned monsters. Here's the difference. If your player character defeated an NPC, such as a wizard with a pet bear, your PC would get experience points (which level up a character) for both the wizard and the bear because they are individual opponents, even if they fight together. But if your PC defeated the wizard, despite their casting a spell to summon a bear out of thin air, then your PC would only get experience points for defeating the wizard because those points would account for the wizard's ability to summon the bear. The bear in this case does not exist separate from the wizard. The bear is inherent in the wizard. And that is how this idiot prince sees his employees: as inherent in himself. He can summon them, then dismiss them at will. They must do his bidding. They have no agency or will of their own. What skills they have are merely extensions of his own skills, that is, these people are less even than slaves; they are simply tools. Which is, in real life terms, delusional and deranged.
I find it interesting that so many heads of tech corporations especially, but most larger corporations as well, never learned the lessons of totally centralized management from the fate of the Soviet Union. Of course, if the only goal is short-term wealth extraction, then the long-term survival of the corporation, and the fates of its employees, customers, and other investors, is of little or no concern.
Yes, I read this announcement and thought first that its backwards, and second that its probably bullshit.
Tech companies that are not hitting their profitability targets always lay off product management, marketing, QA, HR, documentation. Everything except "build and sell". It has nothing to do with AI, but with how easily explicable and directly motivated the roles are. A bunch of sales people with commission and developers with stock will fill in the other roles because they'll be forced to. But you lose cultural capital and important balancing processes when you do this. So like all these stories about how tech companies have laid people off "because of AI" this is bullshit. They laid people off to save money because that's what their investors want and they laid off the people who are always laid off. At best they vaguely hope AI will make up for it.
And they have this exactly backwards. AI can build very fast. Your builders can be more effective, so build more stuff with fewer people. What AI can't do is tell what to build or determine whether the thing it built is any good. Anyone who ever used AI tools knows the main problem is they have no judgement. Its only humans who know the purpose to which the AI is actually being put and so only humans who can tell whether its right. The bottleneck is not building, its measuring.
I read a book recently called “free to obey” by Jonathan Chapoutot. It’s about how much our modern management thought and practice comes from the Nazis. Prince is a prime example. His conception of the use of AI sounds like a better way to “work toward the führer.” I prefer real Drucker—not Prince’s Nazi version.
Investors and speculators alike are happily rewarding companies that slash headcount citing 'AI productivity' even as the core businesses approach a cliff ironically placed there by AI itself.
This is a great piece by Henry.
Here we go again. The greatest shortage today, as before, is not in our analysis, it is in our collective humanity.
The Nazis collected and kept plenty of data on their subjects, especially their Jewish ones. And they drove all the waste out of their killing machines. They were as tech-forward as anyone.
Drucker saw this in person. He was present at his university the day of the Nazi takeover. His account is chilling to read, here in The Atlantic, or in his autobiography "Adventures of a Bystander," which is all about people, his primary source:
https://www.theatlantic.com/magazine/archive/1978/12/the-monster-and-the-lamb/662832/
If we don't get ourselves right, how can we expect our works to go right?
May I also bring to your attention Peter's soulmate Francis Hesselbein? She didn't come from Stanford or Wall Street, she came from the Girl Scouts:
https://www.druckerforum.org/blog/leadership-icons-frances-r-hesselbein-and-peter-f-drucker-a-legacy-of-shared-leadership-by-elizabeth-haas-edersheim/
They both knew how to show up on time. And they both agreed with Kahlil Gibran:
"Work is love made visible."
https://www.poetry.com/poem/54030/on-work
https://archive.nytimes.com/www.nytimes.com/books/98/12/13/specials/gibran-secrets.html
Shouldn't we begin with the most important things, not put them aside until later, when it may be too late?
Thanks for this rescue of Drucker, who along with John Kenneth Galbraith, seems to have been flushed from collective memory just when their ideas might help think through the problems of ensuring this new social technology is used to manage organizations rather than control or replace workers.
How strange for Prince to choose Drucker for this exercise! I wonder what prompt led him or his pet writer there.
Not only Drucker and Galbraith, but W. Edwards Deming as well, who probably understood the use of measurement better than nearly anyone writing about management.
There is more than a whiff of "liberating" AI from oversight in this chap Prince's view, rather than the apparent premise of "liberating" businesses from "superfluous" middle management in the interest of "productivity".
Definitely both.
The more I read about AI the more it freaks me out.
As well it should.
So it turns out a tech AI evangelist can’t read. Not surprising, he probably read the AI summary rather than Drucker’s actual book. Great essay.
He can also be using "Drucker" to justify what he already wanted to do, rather than what Drucker actually wanted him to do. Same as you can always grab a phrase out of the Bible or Orwell to "justify" stuff. At least he acknowledged having to think about it. But it is very hard to get a long existant idea out of one's head.
Well, that would be one of the points of reading, to challenge one’s preconceived ideas. He clearly didn’t read closely enough
Liked "That would be one of the points of reading."
Ideas are like the other addictive substances: the person has to want to change. A lot of people went to the meetings only because the court ordered them to.
https://www.youtube.com/watch?v=g7f6HiQ2LuU
You may love the idea, but that doesn't mean it loves you back.
This is a wonderfully precise distinction: measurement can support judgment, but it cannot replace it. The danger is not that AI measures too little, but that leaders start believing measurement is the same thing as management.
It's remarkable that this guy absolutely swears that he's put a lot of thought and reading into this decision, and then makes these huge sweeping generalizations that belie a gaping incuriousness.
He obviously has spent no time with his middle managers, understanding what they actually do or how they make the business succeed.
The first person we have to manage is ourself. Peter Drucker was someone who knew and did that. If you read his autobiography it is mostly about the interesting people he met. And he seemed to be a magnet for interesting people.
There is no getting around the need to find out what is really going on. In war the Generals have to go to the front lines to find out, because the chain of command will not get the true picture to them. Same problem in companies.
You can think the problem to death, and read all the books, and look at all the numbers on your screen, but you still have to go and find out.
Be where it happens.
Otherwise, to borrow a line from the great comedian, Warren Buffett, when the tide goes out we get to find out who's been swimming naked.
This is where many senior managers in tech companies get stuck, though. They think that because they have to go to the front to find out "what's really going on" that the middle management structure must be broken. That's not the case. The middle managers know about and are balancing many factors that the senior managers don't see. If their priorities are misaligned that's usually the senior managers fault, probably because they are spending too much time micromanaging their most important problem.
It absolutely is true that sometimes senior managers have to focus on some specific problem that's in the span of control of someone it would normally have been delegated to. But that's because shorter lines of communication help with alignment. The problem I've seen again and again is senior management does not even try to get wider alignment and so thinks they need to micromanage everything. The actual military has much better processes, has done for a very long time, and would not fall into this trap.
No argument from me about the danger of micromanagement.
From my experience, the top managers who ignore the middle managers also ignore everyone else. They consider other people to be interchangeable objects to be used. Not surprising, since top managers are so laser focused on data and money today. And they think they are accountable to no one except the march of technology and insatiable Wall Street. They think they are creating a perpetual growth machine.
But the people need a society fit to live in.
Human thriving is not inherent in "optimization" processes. It also got neglected by the "modern" idea in general. This is why people are unhappy with the way the world is today.
One of Drucker's key ideas was that if managers fail to make their organizations serve society the people might eventually become desperate, and turn to people like Hitler as their last chance.
Yes, I think there is an over-optimization problem. Wall street and (even moreso) private equity has encouraged management to expose everyone to the same slightly broken set of incentives, and so while the incentives are only slightly broken - its generally true everyone should try to make money for their employer up to a point - the issues with them get magnified further and further.
Key question is WHO have we optimized things FOR?
Our world:
A new $600,000 electric Ferrari designed by the famous Jony Ive that looks like a clone of a Nissan Leaf that costs ONE/TWENTIETH as much.
WHO are these people that have that kind of money to throw away, and why do they have it?
Sign of sanity: the people have savaged this car online.
https://cybernews.com/tech/ferrari-electric-car-ev/
Just how freaking out of touch can our ruling class get? Do they all live along the beach in Marin County and get chauffeured to their Gulfstreams so they can dine out in Barcelona? Why don't they just buy California and stay home for dinner? Do a leveraged buyout.
Count this as a victory for optimization: all the SUVs now look exactly the same.
All the tech bros are becoming more and more alike.
Conformity MAXXING.
Let's call it a day.
Ai is like a new military technology. You only know how good it is when it is used in a peer to peer engagement. Let the battle begin. For those of you still needing a job, dig a deep fox hole.
Every company I've ever worked for - and I've worked for a lot of them - definitely didn't follow Drucker's prescriptions. It was always been, and remains, top down control.
"He views action and responsibility as going together, stressing both that business has a broader social role (profit is no more than a necessary condition to stay in operation, and business must pay attention to the common good)..." Then along came Milton Friedman...
True top down control is very hard to do in any professional organization. Whether its tech or accountancy, you are always delegating huge amounts of responsibilty down ot the most junior levels. Its very dangerous to miss this. What happens when execs don't think seriously about delegation and whether the people being delegated too can do thier jobs is not true top down control but micromanagement.
Have you ever worked in organizations that weren't "professional" like tech or accounting? The situation might be different in those cases. There are a lot of young people graduating high school and college this year who might be heading straight to Starbucks, or sports gambling. These are the people who are really getting the shortest end of the stick right now. A family member who is a software engineer works for a company where the sociopathic CEO was boasting about how many people he was going to be able to get rid of using AI. Later he had to backtrack, but the sheer entitlement of this crop of executives, after decades of having things go their way, is breathtaking to behold. Not that the knives are not out in the headquarters building, where the losers get "transitioned" out of the company, as if it were an act of God doing it. Children playing with knives.
Yes, hence my qualifier. Its hard to get alignment from low paid staff so managers resort to top down control and tolerate high turnover. I much prefer Drucker's vision where hotel maids or Amazon warehouse employees could be largely self-managing, but all of society has to accept higher costs to make that happen.
I don't think AI gets us to the point of professional staff being in quite the same spot. Anyone can tell whether a hotel room was cleaned or a widget was sent to the right address in the "right" amount of time. No-one can tell whether your accounts are audited correctly or your software works to spec without replicating almost the same work needed in the first place. This is the fundamental problem of software engineering management that Fred Brooks identified fifty years ago. AI can produce much more software much faster, but only a trained professional can tell it what to create or check that it created the right thing. Just as ATMs initially increased the demand for bank tellers AI is going to initially increase the demand for engineers. Especially hardware engineers, interestingly.
Which is my long winded way of saying, fuck those tech CEOs, they're going to suffer for their hubris. Fortunately they're not all like that.
You can imagine after a while the demand for human professional labor is going to fall. That the AIs will be well trained enough that one can write the spec, another can sign off on it, a third can write the code and a fourth can test it, with humans only involved in a high level project management perspective. It will be turtles all the way down at that point. But I don't think that's a short term prospect. They're just not reliable enough, but more importantly the pent up demand for software is massive. Every unfixed bug, every little thing you'd like an app for, its all now possible. The most interesting statistic in all of this is that GitHub shows project backlogs clearly shrinking - that's a real change, project backlogs never shrink.
You do live in my world.
Have a look this:
https://theconversation.com/home-care-the-dutch-model-that-challenges-bureaucracy-272381
Wish somebody here were familiar with this and could tell us how it is going.
Actually, top down control very much is micromanagement. In order to delegate huge amounts of responsibility, the top has to let go of control.
Maybe I'm using these terms in an idiosyncratic way, but when I hear top down control, I think "command and control", which was kind of default management strategy in the Taylorist era, before Drucker et al. Data is reported up the management chain. Commands come down. The only delegated responsibility is to implement the commands. If senior management misunderstands the data and the commands sent down are stupid, no-one can do anything about that. This is what the millitary and then (much later) business worlds tried to move away from in the 20th century, because it produces very bad results, like invading Russia in the winter, the Ford Edsel, and the Ground Nut fiasco.
Micromanagement is when managers step in and try to do the jobs of people lower in the hierarchy for them either directly or indirectly by issuing such detailed instructions the employee is just a robot. It tends not to happen in command and control organizations, because generally there's a strong expectation that managers only direct work that they control, never do it. Rather, micromanagement happens when managers believe their high level commands won't be obeyed, often because command and control has been abandoned and no usable framework has been put in its place. This is the case in many tech companies - everyone knows you're not meant to do command and control, but managers are often very bad at aligning their staff with their goals so they end up getting far too deeply involved in some topics and ignoiring others.
Thank you, that was really well put.
A lot of micromanagement is done by people who used to do the job, maybe did it well, then were turned into managers and couldn't restrain themselves from just jumping in and dictating every detail of the job to get it done the way they would do it. And it is very easy to think there is only one way to do the job: my way.
Been there. Sometimes as the one telling the newbie how it should be done!
There's a big difference between mentoring and micromanaging. Once the trainee knows how it's done, a mentor steps back and let's them go at it. A micromanager keeps putting a stick the their necks.
I think the term "control freak" describes this sort of person.
Maybe even sadist in bad cases.
For fun, read the chapter about "Chainsaw" Al Dunlap in Jon Ronson's book "The Psychopath Test."
I've worked for such managers. They don't actually >do< the work, if they do any real work at all. They just stand over your shoulders, breath down your neck and bark commands of exactly what to do and exactly how to do it - even if you know how to do it better than they do.
My brother used to work for such a company, in a call center. Quite a comedown for a former Learjet captain, but that's how things go sometimes. He said it was more stressful than flying the Lear, which if mishandled can turn into a fireball on the runway when the tip tanks hit, or fall out of the sky at 51,000 feet if you go a little bit slow.
Apparently they didn't bother to train their managers, who seemed to think something was wrong if the customers were not equally happy every single day. Half the people in the building had letters in their files putting them on "performance improvement plans," the first step toward firing. Turnover was high, despite good pay.
The managers never learned literally the FIRST LESSON Deming taught in his quality courses, which was that statistical variation was normal, and can't be improved by haranguing employees. The managers thought there should be no variation, even though all of the customers and their problems were unique. They kept asking: "What's wrong?" They wanted the calls to end as fast as possible, and they wanted reps to upsell. They did NOT put the customers first. That's a nice slogan.
Naturally, the managers who were judging and coaching the employees couldn't do the work as well themselves.
TOP DOWN is still the default management approach almost everywhere.
LYING is ubiquitous in corporations.
What people DO is what they actually BELIEVE.
It makes sense when you realize the very top are psychopaths, and they promote people who are like themselves, that is, the psychopaths. They're easy to identify because they're the biggest back stabbers.
What the management class believes in is domination.
excellent
---
who programmed Prince's AI tool and who feeds it tue required data across all the different technical boundaries?
Unless, he had a fully integrated system (ERP or similar) beforehand, I cannot imagine AI giving him all the required KPI data points just like that.
Also, KPI are very tricky / misleading if used out of context of the underlying 'story'.
....
Great piece. The “ lurid science fiction” line had me wondering about the relationship between “The Practice…” and “Player Piano” and whether anyone read Player Piano and said “sounds good to me…” — is it torment nexuses all the way down?
So many newsletters I get, and now I have to get this one too, which was recommended by a co-worker regarding a book we're working on.
What I most appreciate about this piece is that it also signals an extension of another metaphor. People like Musk describe people who aren't like Musk as "non-player characters." NPCs, for those who haven't played Dungeons and Dragons for nearly 50 years, are individuals in a roleplaying game who aren't directed by the players of the game. They're the barkeep, the server in the bar with the rumor that starts an adventure, the local who hires a party to do some job. Set aside that Musk sees real life as a roleplaying game, what Musk means is, these people aren't real so they don't matter. While they may have personalities and skills, they can be abused, ignored, laid off, fired, killed whatever. Which is, in real life terms, sociopathic. It also suggests he treats his assistants like garbage.
But "Matthew, Prince of Cloudfare" (genius) takes things a step further. He doesn't see his employees as NPCs. He sees them as summoned monsters. Here's the difference. If your player character defeated an NPC, such as a wizard with a pet bear, your PC would get experience points (which level up a character) for both the wizard and the bear because they are individual opponents, even if they fight together. But if your PC defeated the wizard, despite their casting a spell to summon a bear out of thin air, then your PC would only get experience points for defeating the wizard because those points would account for the wizard's ability to summon the bear. The bear in this case does not exist separate from the wizard. The bear is inherent in the wizard. And that is how this idiot prince sees his employees: as inherent in himself. He can summon them, then dismiss them at will. They must do his bidding. They have no agency or will of their own. What skills they have are merely extensions of his own skills, that is, these people are less even than slaves; they are simply tools. Which is, in real life terms, delusional and deranged.
I find it interesting that so many heads of tech corporations especially, but most larger corporations as well, never learned the lessons of totally centralized management from the fate of the Soviet Union. Of course, if the only goal is short-term wealth extraction, then the long-term survival of the corporation, and the fates of its employees, customers, and other investors, is of little or no concern.
Yes, I read this announcement and thought first that its backwards, and second that its probably bullshit.
Tech companies that are not hitting their profitability targets always lay off product management, marketing, QA, HR, documentation. Everything except "build and sell". It has nothing to do with AI, but with how easily explicable and directly motivated the roles are. A bunch of sales people with commission and developers with stock will fill in the other roles because they'll be forced to. But you lose cultural capital and important balancing processes when you do this. So like all these stories about how tech companies have laid people off "because of AI" this is bullshit. They laid people off to save money because that's what their investors want and they laid off the people who are always laid off. At best they vaguely hope AI will make up for it.
And they have this exactly backwards. AI can build very fast. Your builders can be more effective, so build more stuff with fewer people. What AI can't do is tell what to build or determine whether the thing it built is any good. Anyone who ever used AI tools knows the main problem is they have no judgement. Its only humans who know the purpose to which the AI is actually being put and so only humans who can tell whether its right. The bottleneck is not building, its measuring.
I read a book recently called “free to obey” by Jonathan Chapoutot. It’s about how much our modern management thought and practice comes from the Nazis. Prince is a prime example. His conception of the use of AI sounds like a better way to “work toward the führer.” I prefer real Drucker—not Prince’s Nazi version.
Investors and speculators alike are happily rewarding companies that slash headcount citing 'AI productivity' even as the core businesses approach a cliff ironically placed there by AI itself.