1. By making workers unnecessary (largely hypothetical right now?)
2. By companies spending big on AI, but it didn't pay off yet so they need to cut back on something else.
3. AI is a good excuse for layoffs they want to do anyway.
Also - the investors would rather hear "AI" than "oops we are in trouble so we need to do layoffs". For example, if you spent a lot of billions on a 2nd life clone with fewer players than developers ...
All of these tech companies (with perhaps the notable exception of Apple) massively overhired during the pandemic, and that overhiring was on top of a decade+ of the ZIRP era. So there are 2 main drivers of these layoffs:
1. Correcting pandemic overhiring
2. In the ~2010-2022 timeframe, tech companies poured all this money into speculative bets that never went anywhere, at least from a profit perspective (think Amazon's Alexa devices division, Google Stadia, and perhaps most famously the Metaverse itself). All those diversions are now toast, and they employed a ton of people. The only speculative bet that is now "allowed" is AI, which is one reason why I giggle whenever I hear people trying to defend their companies or projects by adding "AI" somewhere in the name.
So perhaps my second point is similar to your #2, but I think the important difference is that the end of the ZIRP era would have caused companies to kill these inherently unprofitable projects even if AI never came on the scene.
> 1. Correcting pandemic overhiring
> 2. In the ~2010-2022 timeframe, tech companies poured all this money into speculative bets
Any data/sources on which this might be based? The pandemic was 6 years ago; do these "Agile" (the tech term) companies really carry many unproductive lines-of-business for so long?
> speculative bets that never went anywhere ... think Amazon's Alexa devices division, Google Stadia, and perhaps most famously the Metaverse itself
Organizations make speculative bets all the time. Is there an accounting of the profitability of Alexa/Nest etc.?
> end of the ZIRP era would have caused companies to kill these inherently unprofitable projects
if you plug in the years 2020-2026 in the Fed Rate - Unemployment chart here at [1], it shows that from 2020 - 2022, rates were near zero while unemployment spiked during Covid and then fell. From 2022 through 2023, rates rose sharply while unemployment stayed relatively low. 2024-2025 the labor market softened. You can add the Federal Funds Effective Rate and the Unemployment Rate easily through the menu.
Unemployment stayed low through the rise in rates for almost two years prior to 2024. Given that companies operate on a quarterly reporting basis and program/project decisions are at least on that cadence, I don't think that the line you're suggesting that Rates-Go-Up -> Projects-Get-Killed -> Layoffs-Increase quite lines up with the economy-wide data in this exceptional case of 2022-2023.
We may have to look elsewhere for the reasons behind the current labor market weakness ... cough..*economy*..*trade walls*..cough...*structural re-alignment* [2]...cough...
[1] https://fred.stlouisfed.org/graph/?g=1duFv
[2] 6% employment decline in 22-25 year old workers https://digitaleconomy.stanford.edu/app/uploads/2025/11/Cana...
These were supposed to be farsighted geniuses who took credit for their stock performance. But they made the most obvious blunder: opening a Pandora’s box that they can’t close.
In 2022, Meta laid off 11,000 (13%)
In 2023, Meta did two layoffs, one of 10,000 (another 13.6%), one of 600
In 2024, Meta did a layoff (layoffs.fyi doesn't list the number)
In 2025, Meta did three layoffs, one of 3,600 (5%), 100 and 600 people
In 2026, Meta has done two layoffs, one of 200, and one of 8,000 (10%).
If you can't correct overhiring in 4 years as a CEO, you've failed. If you repeatedly have to do >10% layoffs every year for 4 years you've failed (if it's to correct for a one time hiring spree).
Meta's employment didn't skyrocket during the pandemic making this argument even more bullshit. Between 2012 and 2018 Meta averaged an employee count growth of 40% YoY. In 2020, they grew by 31%, 2021 22% and 20% in 2022. Meta slowed their growth during the pandemic.
This "pandemic overhiring" is bullshit.
In 2018, 35,500.
From 2019 to 2022, 44k to 86k.
2026, as of now, 70k, 26k higher than 2019.
Using relative comparison is not appropriate for a software company, which is incredibly scalable by nature. Yoy is not the appropriate metric here. These are not workers on the line in a widget factory.
I do agree with your point about overhiring, its been way to long and this continues to be an excuse without any real evidence to back it up.
Do big tech companies like FB and Google even pretend to be "agile" anymore? I think they mostly sell themselves on institutional stability and monopolist market positions rather than speed of execution
Excepting for trivial-size, freshly formed startups, companies cannot be "Agile", because finance and legal and HR and even marketing have constrains setting the tempo - you cannot just drive them with a sprint as if it was a clock signal.
> Excepting for trivial-size, freshly formed startups, companies cannot be "Agile", because finance and legal and HR and even marketing have constrains setting the tempo - you cannot just drive them with a sprint as if it was a clock signal.
Implementations of Agile at different companies can be an issue, yes. But that is to be expected in any large organization, simply because of scale. It doesn't change the fact that the on-the-ground teams at agile orgs work to a different cadence and approach than historically traditionally structured companies.
There are a few different ways to manage interfacing with parts of the org that need to march to a different beat. That always creates friction, and has to be managed properly. Any large org can suffer from hubris, middling management skills and capacity, wasted effort. Problems of scale, I guess.
“Move fast with stable infrastructure” as they say.
Folks from those companies will have to speak up, but my understanding is that yes, internally these large tech orgs use the Agile Methodology, as opposed to the 'traditional' 'Waterfall' development methods.
I think it is hard to overstate the effect that Waymo will have.
2. If it captures most of the transportation market. That's debatable because self-driving tech works even better for trains, trams, subways, buses, etc. And that's before we go into other self-driving car companies.
3. If it becomes truly production ready within the next 10 years or so. By "truly production ready" I mean a Uber/taxi competitor in at least 20 alpha global cities outside of the US. Otherwise yeah, it's going to be a great tech but with a 30+ years ROI.
But waymo created the category by showing it is possible.
But also, while there have been layoffs in engineering teams, Ive seen a lot of "support staff" get absolutely obliterated. Things like "agile coaches", "technical project managers", UX testers, marketing roles, etc. etc. While I've seen most of my laid off soft engineer friends find new jobs relatively quickly, I've seen lots of folks in these other roles suffer long bouts of unemployment, and often leave tech entirely. It's these folks I feel the most for. A lot of them were making low 6 figures 10-15 years ago, and now many of them have no hope of making that much in their careers again because companies have vastly reduced the number of those roles.
Lot of these companies are bloated from having way too many Engineers anyway. Once you have mature software that brings in bagfuls of money, you don’t need that many people to keep the ship steady. I have seen this first hand at MSFT, we started a new team back in 2019 and it probably had ~40 people full time across US and India. By 2024 when I left, we had about 20 people in India who could easily run the service, the US team was dissolved and they moved to other teams in MSFT. The fact was that new features were few and the team was in KTLO mode. I have seen the reverse happen too, the team I was working on was dissolved and we were moved to different teams and everything moved to the US last year, managers were converted to ICs and a few folks were probably fired but it was a ~10 year old service that didn’t need that many people to run, even more so after AI tools became big last year.
"Agile" can go and die in a hellfire for all I care.
But good technical project managers aka "bridges between the higher-up beancounters and the workers" are worth their weight in gold.
Yeah but it's not easy do distinguish those from the snake oil salesmen who are just good at smooth talking during the interviews.
Beyond that though, there's the probation period. If they can't do the job, they're supposed to be let go before they become permanent.
Trouble I see from most interviewers is a tendency of asking questions with a "right" answer. Those tend to be a lot easier to game. They then fallback on sorting applicants by pedigree - the old, "no one ever got fired for choosing IBM" method.
Then, they come back and rant about how PM's are trash, and Agile is trash, etc. etc.
The remaining piece is to speak with some personal references to verify they did some real work.
Not a thing where I am. You can BS your way in these jobs and so many people did.
There's no LEETCODE for management positions, just bullshitting smooth-talk and using your connections (nepotism).
There are real differences in the knowledge and work behavior of great PMs and the bullshitters, and it's usually not that hard to tease out the bullshit in an interview if you know what you're doing.
I was talking about a country, not a workplace, and franky, so what? I can't change that. What I can do, is tell you how it is, and how the system gets exploited.
If you live in some magic utopia where things are different(the US maybe?), good for you, but this information doesn't change anything for me where I live.
Best I can do is adapt and exploit the system in my favor as well if I can the same the rest do, otherwise I get left behind by the unscrupulous clueless scammers who do.
I am skeptical of Doctorow's theory because it looks like LLMs will continue to improve enough over the near term to be able to handle issues caused by AI-written code from the past few years.
I have this theory that the bloat will follow to the full extent possible. OpenClaw has this, the OpenEye or whatever that comes on another day, with better models, will have 3 million lines of code. All of the possibilities that you mention will not come to fruition the way you'd like to, because speed is preferred over building better things, and to hell with maintainability.
Eventually these things will become a ton of black boxes, and the only option will be to write them from scratch with another next gen LLM. Lots of costly busywork, and it will all take time.
I moved to the Seattle area during the dotcom boom.
Within 18 months I was unemployed.
There was DEFINITELY a feeling, like the whole “internet” thing might have been a bubble. I helped a friend move to Pleasanton CA and there were so many empty office buildings, it looked like a zombie movie.
But it all came back, and more.
replacing SRE-4s with AI is the point
It can be argued that the demand for graduates in other industries might have stagnated or even dropped, and it's "spread over" many different industries, so it's not really that seriously felt.
But if you were hiring in the software industry during the covid peak years, you would seriously feel the shortage. I used to interview candidates in a FAANG, and at some point it was more likely than not that a candidate that we liked and prepared to make an offer would tell us they already accepted an offer from another FAANG...
Post-2015 and accelerated during Covid, many east coast techies I knew that didn’t go to the right school, get a 4.0 and take the “right path” suddenly started popping up on my LinkedIn ad having joined a mag7.
And then everyone else lower on the food chain got even less picky.
And then we had the whole PM / DS / etc new job path fads that pulled in non-CS and sometimes non-STEM people who did a bootcamp.
If an engineering graduate has a chance to make $0.8X at a US company that makes hobby drones, $0.9X at a US company that develops 3D printers, $1X at a US carmaker that's struggling to develop a good EV, or $1.5X at a US adtech company - you can imagine where they end up.
Alexa devices had so much potential, if they'd only practised customer obsession like Amazon used to (still?) preach, and not stuff their products with ads.
That means that a large number of high ranking people in these companies projected they would need these people in the coming years, and then some.
I think it may be darker than that, and the overhiring was a tentative measure to build up a charge, like electrons in a capacitor, to release a shock to the market that would achieve two aims:
1) drive salaries down through fear and contest
2) reduce the bargaining power of employees (software engineers were starting to look to unionizing in the past 5 years especially and that has gone deathly quiet even though working conditions have worsened and demands have increased on 'ICs')
Any objection as to whether these companies executive teams were collaborating should be seen as very empty by now, given we know they are all deeply circularly invested in each other and thus are bound to each other's success or failure.
On the long term, prolonged stress kills innovation and engagement.
There also no financial reason to lay off people. The revenue and profit for employee is already way better than per 2020 levels.
I suspect it really is to pour more money into AI capex.
I don’t think Apple is an exception. I think they have also over hired but they are also scaling, albeit slower than they used to. The scaling elsewhere is not happening, especially meta where they are trying to extract money from every corner they can find out of desperation, and so the books need to become lighter.
For Apple, hiring more than they need can be soaked into the books because their sales and profits keep increasing, though the rate of growth has slowed. However, if it’s an expense that can be avoided, then it’s an expense that should be avoided.
I'm not one to invoke "incompetence" willy-nilly, but I'm having a hard time explaining these policies apart from it.
Anyway, I bring this up because it didn't quite fit into your 3.5 rubric.
Good people are leaving, people who can't, don't. It's hard not too see how it dilutes talent.
One of the consequences of having bean counters run tech companies.
The core business is still meta ads, but Zuck had decided they needed big investment into a new business for future-proofing, growth or whatnot.
That business was initially the meta stuff. Now it is Ai. That's a pivot.
You were soooooo close.
Meta is fundamentally a targeted advertising company, not a media company.
The amount of subscription driven media today is relatively a very small part of the industry.
From the customer perspective, it works. The main problem is price competition for access to audiences. That's why fb revenues are what they are. The ads work, largely because ofntracmjng/targeting.
Most language translations and asset creations for CMSs are now AI driven.
In big corps delivery teams were already being reduced by relying in LEGO building with SaaS, iPaaS and serverless/microservices (aka MACH architecture), now with agents, the integrations teams get further reduced into writing the tools/skills modules instead.
If 1 employee can do the work of 3 now but Meta's TAM can't grow 300%, then they can cut some employees.
In other words, worker productivity might be higher than what the ad business can grow into, so Meta can safely cut cost and still hit their growth targets.
Edit: I should be clear that I think #1 has been achieved for software development.
1 employee doing the work of 3 is I think is a stretch
but 1 employee doing the work of 1.1 employees from a year ago I think is almost certainly true - at least, me and everyone i work with is _at least_ 10% more productive, and using AI extensively
In my 20 year career I’ve rarely been on a team with more than 3-5 people on a team or within region on a team.
So at that scale it’s not really reducing a team member on a given team still. But you get more productive which is notoriously hard to measure in SWE, so yeah. It’s possible that translates to iterating faster or closing tickets further down the backlog which is useful but not per-se staff reducing.
Maybe in mag7 where you have massive engineering orgs the 10% can impact a given team more..
and how many of them are totally wrong, or right about it!
[1] and how it might be changing with new generations of models
For all the hype about the 1X vs 10X distinction the real stumbling block is how many 0Xes there are out there and how frequently they tend to make it through hiring.
If you go by the measure of LoC per employee, then your number is probably even higher, somewhere between 10-20x per employee. The problem being, producing 10.000 lines of AI-slop per day is not a good productivity measure - all it does is create more technical debt and issues that now nobody is reviewing because a) people get fatigued and at some point just wave the AI-slop through b) there is not enough manpower because people got laid off because of "AI" c) People are generally feeling irritated by being asked to review and correct AI slop. There is a societal pushback brewing and it won't be nice for the so-called AI in the end. Think about the fact that most people who are exhilirated by the "AI" are either incompetent or incompetent and old. Most of the young folks, even those not in the technical domains, firmly reject AI. When did you ever hear of a revolutionary new tech that was actively hated on by the young people?
> Edit: I should be clear that I think #1 has been achieved for software development.
Maybe in the world of WP-plugins/typo3 and other simple work, though even those are fairly complex in their own ways which the retard-LLMs will trip on fair amount of times. Not if you are doing anything remotely complex. The retard-LLMs will still either put your secrets in plain text, suggest the laziest f*ing implementation of a problem etc. It's just a shitshow nowadays, compounded by the LLM companies trying to keep the costs low (and therefore keep the "users" hooked), which they currently accomplish by shortchanging you and dumbing the LLMs down - because otherwise they'd have to charge for true cost - upwards of tens of thousands of dollars per seat - which would render their initial value proposition completely useless. Something has to give.
Now not all the extra code is necessarily useless (you can imagine some refactoring or perf improvements) but clearly a lot of it is.
We still aren’t very good at knowing how to use AI judiciously.
This was a reasonable position to hold 9 months ago but it’s absurd now. I’m not going to convince you - but you really should give it a try.
That's cute, I hope you enjoy that high, it's really impressive at first - fyi -I've been using GH Copilot since early days (invited to early access) AND paying it for my entire company ever since MS published the first commercial plan. All the way to the latest entshittification drama with Opus 4.6 being pulled away and Opus 4.7 taxed at 7.5x rate. Yeah, its great for quick tryouts or similar. But using it in wider scope, with complex reqs and dynamic environment? Complete shitshow.
> This was a reasonable position to hold 9 months ago
Browse my comment history. There were people just like you, 6, 9 or 12 months ago telling me exactly the same. Some also threatening that I would "be the first to go away". Like I said, cute actually :) You know in January Dario Amodei announced again, AI would write ALL code in 6 months. Do you see it happening?
What kind of smug asshole says something like this?
If you think early days of GH Copilot are remotely relevant to what is happening now, we do not live in the same world.
Let's learn some English grammar, I had said: "I've been using GH Copilot since early days" . Translated from plain English: "have been using..." = *an action or state starting in the past AND continuing UP TO the present time*. I did not say "I used it only during the early days and never again". Plus my team. Honestly - and it's only a hypothesis at this point in time I cannot prove, but I meet more and more people whose usage of LLMs seems to be seriously impacting their reading comprehension and writing skills. Do some LLM-detox, I mean this without any kind of disrespect.
> What kind of smug asshole says something like this?
I've also noticed over the years, the folks who get this offended are usually those who are uncertain about their opinion. I simply found your statement cute, that's all. You are obviously in the early stages of using it and I am assuming using it to alleviate your own little, relatively simple tasks in your job which otherwise would have been overburdening for you. It's only human to try and do so, but it hardly describes "productivity" the way you seem to think it does.
I am not. You do not have any knowledge about what I've done or for how long. The arrogance is truly breathtaking. You should work on it.
I've also used GH Copilot since early days, ChatGPT since launch. I did not start using agentic systems (mostly Cursor) until about a year ago, and found them to be of little help in generating code in our large, long-lived and sprawling code base. Something fundamentally changed with Opus 4.5 but I didn't notice it for quite awhile because I had already constrained my usage patterns to what I thought I knew to be their limits.
> little, relatively simple tasks in your job which otherwise would have been overburdening for you
I have done that, and many other things. You do not know anything at all about what I have done with AI. You are projecting your ignorance in a way that makes sense to you, but it is still nothing but ignorance.
If you are actually interested in my recent observations I posted this in another thread recently.
The sentiment is that the latter is more expensive and less flexible than the former... which seems backwards to me. (AI is expensive, robots are mostly not that adaptable to new tasks, automations only really help the most repetitive tasks and they must be rewritten/updated when the task changes)
At least in the software development front, i really cannot see that happening.. until now we were all understaffed. Now it is the first time that actually with our team we can handle the workload properly.
The days of meta having network effects to defend its position are long gone, and I suspect we'll see the products die when an AI-first UX comes to mobile.
There are really just a handful of people from my extended social circles still actively posting to FB/Instagram and that traffic is mostly drowned out by slop content.
Cross-checking against actual expenditure, Meta spent $118B total last year, with the second largest component of total spending being stock comp at $42B, of which vast slabs went to the top leadership that's presumably also not getting fired.
You calculate the cutoffs as savings for this years while imagining that the future payments are payments only for this year. At the same time the commitments are for 5-20 years ahead and the laid off people would be off the payroll for the same multiple years ahead.
They're a suicidal bet, because they assume cloud LLMs are efficient and inevitable.
Neither of those is true, or even likely, and we're going to see the consequences by the end of the decade.
4B over 5 years is 20B, which is significant.
There's a whole lot of circular funding being passed among the same dozen or so companies right now with very little actual construction or assets to show for it and at some point someone will be holding the bag when actual money is called for, and nobody wants it to be them. The parallels with both 2008 and the '90s S&L crisis are troubling.
It's not like Meta has nothing to show for the money it spend, but it seems like they could have spend that money on improving Facebook or Instagram, not that I think Zuckerberg really cares about those product anymore.
Mark is just looking for the next fad.
Finance is destroying the real economy in search for "optimization" because business value doesn't neatly fit into an excel sheet cell. All these layoffs have been done because of AI washing right from the start.
That said, AI isn't helping.
Promises/hopes of what AI can do, and also execs being misinformed about what their own companies are doing/achieving with AI. I know of one very well known large company where the CEO is in the press preaching about the need to restructure/layoff because of AI, yet in the trenches there is close to zero AI adoption - only contractors claiming on their JIRA close-outs to be using GIT copilot because they have been told to say so.
And a bunch of yes-men down the lower layers of management funneling these ideas.
In a meeting at my last job, one of the execs was bragging about how a chatbot was reading Jira customer service tickets and calling tools/APIs to solve those tickets, and it "only costs 1.5USD per ticket. How much would a human cost, huh?"
Little did the exec know, but my team was already using a ~600 lines python script to solve the problem with a higher rate of precision. The chatbot-automation thing was largely pushed by my manager when I was out on vacation, just so he could earn his good-boy points with higher ups. Worst manager I've had in my 14 years of career btw.
1. Cut costs
There are multiple reasons companies may need to cut costs. Generally, historically, they are not positive
One of them is that, internally, the business is not doing well or the company believes there are problems on the horizon
Perhaps a company that is healthy and optimistic about the future could at the same time be conducting successive rounds of layoffs
But are there many examples of healthy, shrinking companies from history
Usually company growth is taken as a sign of prosperity. Multiple rounds of layoffs are not. More like survival mode
But of course this time it will be different and the remaining employees after the layoffs have nothing to worry about
No, it's absolutely happening that way.
I work in tech in a senior IC role. I'm now expected to single-handedly do the work that I would previously have led a team of 5-10 people to do. The junior ICs, we aren't really hiring for. This is not specific to my company, it's like this across the board.
This was in 2022... So I cannot imagine how much worse the "automoderation" will probably become.
Given the unironic use of 'woke mind virus' by certain billionaires, there's something to be said about the sudden obsession of building data centres, no expense spared.