I'd bet revenue per employee gets gamed just as fast as headcount did. It's an easy number to juice: push work out to contractors, agencies, or your supply chain and it disappears from the denominator. Nvidia runs north of $5M per employee, and part of the trick is that the people physically making the chips work for TSMC. New scoreboard, same old gaming.
It's the AI hype/craze. The market is voting/betting that this is the way to victory, which says a lot about the current state of affairs. And those X posts are mere theater to take advantage of that, instead of looking bad due to mass layoffs (which used to reflect badly and drive stocks down).
What's worse is that many investors probably know this, but they buy the stocks anyway and fuel the fire because they expect others to do so (and then they can sell for a profit once the price has gone up). It's a bit disconnected -- or rather, it's all based on vibes.
CEOs are basically salesmen, and what they must sell is the appearance of success (or anything which drives up stock prices for their shareholders). It doesn't matter what's going on inside the ship, as long as the hull and the deck look good. But of course, this can only go on for so long.
As data people, we know this. Most of us have been asked to pull up the best looking numbers at one point or another.
yeah, and in some cases, I don't think it's necessarily all cynical. Like, there are definitely some tech companies (and many others, I'm sure) that have a lot more employees than they necessarily need. But, as you say, the message from the CEO will be the same regardless.
as someone starting their career in 2010s, it's riveting to know there used to be something before this that maybe didn't seem so horrid. means maybe the future can again look differently.
i love that 1 person businesses are exploding, but will there be enough real problems to solve or will everything become a mass of oversold slop that doesn't last more than 3 weeks.
it does feel like a a bit of the shein era of silicon valley, where in a few years, there are going to be a bunch of single wear companies out there that sold some product for a minute, and then never really went anywhere.
Which, for the people working there, seems bad. For the ecosystem though (and the people not in silicon valley), that might be better? Or at least, lots of people experimenting on lots of ideas, and most of them getting thrown away but a few really good ones winning, is at least arguably good? (Although, then, maybe the most apt comparison is social media, where the most popular performers aren't selected from a handful of professional actors but from millions of people posting on tiktok, and I'm not sure that's really made things better.)
I want to believe that most people strive for quality, but sometimes you need to put in the reps for visibility before you can put the best quality forward, otherwise it’s great stuff for an empty room
Yeah, same, I'm not so cynical to believe that many people start wanting to create something lousy. But, when push comes to shove (in big ways and lots of small ones), I'm sure there are a bunch of gradual sacrifices that people make that drift away from that. And the more the ecosystem rewards it (a la social media), the more the temptations, I'm sure.
"an awfully low threshold from which to draw any meaningful conclusions...."
this is the winning quote for me on this one. I have also looked this at the leadership status heuristic point - recruiters use size of org as a proxy for leadership level. This has some merit as motivating and moving large organizations is often a different skill than doing that in smaller numbers - I see this as another two sided issue - one of the reasons people need to have smaller organizations is that they experience the Peter principle in their ability to lead when the org gets big. No matter how you look at this there are some serious evolutionary changes happening to the world of work and you are right to start with the lens of status/ego since that tends to be the unexamined motivator behind so many behaviors and decisions.
In political news, there's always been this kind of odd refusal to every question if people are just lying. It's like, if a headline says "politician raises concerns about X thing," the rest of the story has to talk about the potential origins of their sincere and honest concerns, and can never say "we all know they don't care about X, they are only pretending, this is obvious to everyone." And in the rare cases where someone does say that the politician inevitably gets indignant about it, and how unacceptable it is to possibly question their sincerity.
It feels a bit like that sort of thing here, where of course people make decisions for all sorts of reasons, often because they want to be liked or whatever, but the polite thing to do is to pretend none of that is real.
I would say "“jobs are bad” narrative has some level of truth in it, if you look at it from a macroeconomics POV: development of tech --> increase productivity, decrease cost & threshold.
But the bigger truth now is many bad leadership just use that to mask what they are truly doing - bad management. And I am pretty sure a few them realized their own mistake and quietly rehiring the same ppl hey laid off back...
yeah, there was a lot of skepticism about this early, especially with some of the big companies. At a minimum, the companies said stuff like "we need to be prepared to finance the transition to using tons of AI" or something kind of similarly vague that made it plausible for them to say it was about AI without pointing to the AI it was about. But now, at least with some of the big tech companies doing it, I don't think it's all that. Some, maybe, but it also feels like a lot of CEOs are just true believers.
And those CEOs are usually too far from the frontline and lack an in-depth understanding of how exactly that work is being done and how value is being created... I would argue that if the CEOs (or any leaders at all) actually have a deep understanding of a certain domain, they will also tend towards empowering the existing team with productivity boost by using AI instead of 100% replacement.
This is nothing new. AI is just the newest form of outsourcing (and much easier to work with than a foreign team). So we can expect the same push (or even stronger).
There was some post somewhere (aaron levie maybe? who is ironically a ceo) that I remember seeing where someone said that CEOs are often the most delusional about AI's capabilities because they're the farthest from seeing all the ugly edges that make it hard. They hear the big strategic transformation pitches, but never see the realities in the trenches.
The questions that come to my mind reading your great post are the following:
If all of this is true and hiring the fewest people is the new metric, why are VC firms giving tech start-ups all that money? To fulfill what need is that money required?
Also, possible small typo here, with "And" instead of "An" ?
"And engineer who works at Anthropic built a literal scoreboard."
Certainly, some of it's hiring, though I suspect a lot of it is to spend money on AI costs?Eg, if you went to raise $100m ten years ago, they'd assume something like 80% of that was on headcount. Now, if you raise $100, I bet you could get away with saying it's 40% headcount, 40% AI, and 20% other. Though, there also seem to some companies that still can't resist hiring.
My inner Nicholas Nassim Taleb is now jumping up and down and screaming about headlines like "X company lays off people because of AI" in the same way that the real Nicholas Nassim Taleb used to jump up and down and scream about headlines like "Stock market drops today because of semiconductor rout". Or, at least, he had very opinionated statements, which I agreed with, in books like "Fooled by Randomness."
My point being that he talked a lot about how these statements by the press are non sequiturs, or, at least, nobody had reliably demonstrated a true causal link. Media people had to write headlines because that was their job, and headlines that just said "X happened" didn't get as much attention as "X happened because Y." So they would just make up a plausible sounding "because Y" to append to the already-known "X happened." And then readers would read the whole headline and have a sort of philosophical confusion because they wouldn't properly disentangle the hypothetical second part of the statement from the factual first part.
I know that "because AI" isn't quite in the same category as "because of semiconductor rout" in the sense that "because AI?" is a genuine question on everyone's mind and it's less random than just picking something out of a hat like "semiconductor rout." I just still can't help but feel frustrated at the human habit to always need to immediately attach a cause to everything.
I'd bet revenue per employee gets gamed just as fast as headcount did. It's an easy number to juice: push work out to contractors, agencies, or your supply chain and it disappears from the denominator. Nvidia runs north of $5M per employee, and part of the trick is that the people physically making the chips work for TSMC. New scoreboard, same old gaming.
Profit (EBITDA) per employee fixes that, no?
(Revenue - COGS - G&A) / Employees
yes but don't forget to adjust it for the community https://www.asimplemodel.com/bips/bip-feed/community-adjusted-ebitda
Ok so ((Revenue - COGS - G&A) + COGS + G&A) / Employees!
Wait...
So he was just ahead of his time?
organic chemistry is so inefficient
thats why i wrote this comment without capitals or punctuation
so much more efficient i saved a whole 9 seconds
thats minutes per day
hours per week
days per month
weeks per year
if i keep this up i can buy a beach house for my yearly day off
https://www.tiktok.com/@mendelfogelman/video/7482876384214338859?lang=en
(ht to angela, like and subscribe: https://human.substack.com/p/figure-out-what-makes-you-tick-and)
My blood sugar spiked just looking at that cake. I have negative 200 employees. I hire people and lay them off the same day
that's the hero we need. hire people, solely for the sake of showily firing them. a true artist.
I hope you're posting about those layoffs on X and proclaiming AI productivity improvements so that your stock price goes up.
That always gets me. These companies announce they are laying off thousands ,yet their stock price goes up. Why?
Lower overhead costs = more net profit
It's the AI hype/craze. The market is voting/betting that this is the way to victory, which says a lot about the current state of affairs. And those X posts are mere theater to take advantage of that, instead of looking bad due to mass layoffs (which used to reflect badly and drive stocks down).
What's worse is that many investors probably know this, but they buy the stocks anyway and fuel the fire because they expect others to do so (and then they can sell for a profit once the price has gone up). It's a bit disconnected -- or rather, it's all based on vibes.
CEOs are basically salesmen, and what they must sell is the appearance of success (or anything which drives up stock prices for their shareholders). It doesn't matter what's going on inside the ship, as long as the hull and the deck look good. But of course, this can only go on for so long.
As data people, we know this. Most of us have been asked to pull up the best looking numbers at one point or another.
yeah, and in some cases, I don't think it's necessarily all cynical. Like, there are definitely some tech companies (and many others, I'm sure) that have a lot more employees than they necessarily need. But, as you say, the message from the CEO will be the same regardless.
as someone starting their career in 2010s, it's riveting to know there used to be something before this that maybe didn't seem so horrid. means maybe the future can again look differently.
i love that 1 person businesses are exploding, but will there be enough real problems to solve or will everything become a mass of oversold slop that doesn't last more than 3 weeks.
it does feel like a a bit of the shein era of silicon valley, where in a few years, there are going to be a bunch of single wear companies out there that sold some product for a minute, and then never really went anywhere.
Which, for the people working there, seems bad. For the ecosystem though (and the people not in silicon valley), that might be better? Or at least, lots of people experimenting on lots of ideas, and most of them getting thrown away but a few really good ones winning, is at least arguably good? (Although, then, maybe the most apt comparison is social media, where the most popular performers aren't selected from a handful of professional actors but from millions of people posting on tiktok, and I'm not sure that's really made things better.)
I want to believe that most people strive for quality, but sometimes you need to put in the reps for visibility before you can put the best quality forward, otherwise it’s great stuff for an empty room
Yeah, same, I'm not so cynical to believe that many people start wanting to create something lousy. But, when push comes to shove (in big ways and lots of small ones), I'm sure there are a bunch of gradual sacrifices that people make that drift away from that. And the more the ecosystem rewards it (a la social media), the more the temptations, I'm sure.
yeah agreed, sigh. but we derail from the original post 😄 thanks for entertaining me with this chat
"an awfully low threshold from which to draw any meaningful conclusions...."
this is the winning quote for me on this one. I have also looked this at the leadership status heuristic point - recruiters use size of org as a proxy for leadership level. This has some merit as motivating and moving large organizations is often a different skill than doing that in smaller numbers - I see this as another two sided issue - one of the reasons people need to have smaller organizations is that they experience the Peter principle in their ability to lead when the org gets big. No matter how you look at this there are some serious evolutionary changes happening to the world of work and you are right to start with the lens of status/ego since that tends to be the unexamined motivator behind so many behaviors and decisions.
In political news, there's always been this kind of odd refusal to every question if people are just lying. It's like, if a headline says "politician raises concerns about X thing," the rest of the story has to talk about the potential origins of their sincere and honest concerns, and can never say "we all know they don't care about X, they are only pretending, this is obvious to everyone." And in the rare cases where someone does say that the politician inevitably gets indignant about it, and how unacceptable it is to possibly question their sincerity.
It feels a bit like that sort of thing here, where of course people make decisions for all sorts of reasons, often because they want to be liked or whatever, but the polite thing to do is to pretend none of that is real.
Hypocrisy ftw
+1 for funfetti
Clearly the answer is to measure success by total WAR across your employees.
yes, though of course, the most hardcore companies use average WAR
I would say "“jobs are bad” narrative has some level of truth in it, if you look at it from a macroeconomics POV: development of tech --> increase productivity, decrease cost & threshold.
But the bigger truth now is many bad leadership just use that to mask what they are truly doing - bad management. And I am pretty sure a few them realized their own mistake and quietly rehiring the same ppl hey laid off back...
I also wrote a new article elaborating my POV: https://jimmypang.substack.com/p/no-ai-is-not-the-point?r=2i7jb6&utm_campaign=post-expanded-share&utm_medium=post%20viewer
Would great to have your thoughts on this piece of mine as well.
yeah, there was a lot of skepticism about this early, especially with some of the big companies. At a minimum, the companies said stuff like "we need to be prepared to finance the transition to using tons of AI" or something kind of similarly vague that made it plausible for them to say it was about AI without pointing to the AI it was about. But now, at least with some of the big tech companies doing it, I don't think it's all that. Some, maybe, but it also feels like a lot of CEOs are just true believers.
And those CEOs are usually too far from the frontline and lack an in-depth understanding of how exactly that work is being done and how value is being created... I would argue that if the CEOs (or any leaders at all) actually have a deep understanding of a certain domain, they will also tend towards empowering the existing team with productivity boost by using AI instead of 100% replacement.
This is nothing new. AI is just the newest form of outsourcing (and much easier to work with than a foreign team). So we can expect the same push (or even stronger).
"A country of geniuses in a data center".
There was some post somewhere (aaron levie maybe? who is ironically a ceo) that I remember seeing where someone said that CEOs are often the most delusional about AI's capabilities because they're the farthest from seeing all the ugly edges that make it hard. They hear the big strategic transformation pitches, but never see the realities in the trenches.
They only hear the sales pitches ("I no longer use my IDE!"), and confuse those sales pitches for actual technical guidance.
But it's no different than every other sales pitch I've ever heard: grossly oversold.
And we should know this, because "Big Data" was similarly oversold not too long ago. Same with every other new thing. It's the hype cycle.
The questions that come to my mind reading your great post are the following:
If all of this is true and hiring the fewest people is the new metric, why are VC firms giving tech start-ups all that money? To fulfill what need is that money required?
Also, possible small typo here, with "And" instead of "An" ?
"And engineer who works at Anthropic built a literal scoreboard."
Certainly, some of it's hiring, though I suspect a lot of it is to spend money on AI costs?Eg, if you went to raise $100m ten years ago, they'd assume something like 80% of that was on headcount. Now, if you raise $100, I bet you could get away with saying it's 40% headcount, 40% AI, and 20% other. Though, there also seem to some companies that still can't resist hiring.
My inner Nicholas Nassim Taleb is now jumping up and down and screaming about headlines like "X company lays off people because of AI" in the same way that the real Nicholas Nassim Taleb used to jump up and down and scream about headlines like "Stock market drops today because of semiconductor rout". Or, at least, he had very opinionated statements, which I agreed with, in books like "Fooled by Randomness."
My point being that he talked a lot about how these statements by the press are non sequiturs, or, at least, nobody had reliably demonstrated a true causal link. Media people had to write headlines because that was their job, and headlines that just said "X happened" didn't get as much attention as "X happened because Y." So they would just make up a plausible sounding "because Y" to append to the already-known "X happened." And then readers would read the whole headline and have a sort of philosophical confusion because they wouldn't properly disentangle the hypothetical second part of the statement from the factual first part.
I know that "because AI" isn't quite in the same category as "because of semiconductor rout" in the sense that "because AI?" is a genuine question on everyone's mind and it's less random than just picking something out of a hat like "semiconductor rout." I just still can't help but feel frustrated at the human habit to always need to immediately attach a cause to everything.