upvote
3 and a half ways AI takes jobs:

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 ...

reply
It's #3 - it's always #3.

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.

reply
> here 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

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...

reply
How many hall passes do CEOs get for pandemic over hiring? Its been four years, and four years of layoffs and hiring to go back to layoffs. Pandemic over hiring might explain the first layoff or two a year or so later, but not the tenth or so layoff nearly half a decade later.
reply
> 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?

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

reply
Did they ever pretend, and anyone ever believed that? "Agile" organization is even more of a bullshit concept than "Agile" in the team.

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.

reply
> "Agile" organization is even more of a bullshit concept than "Agile" in the team.

> 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.

reply
Yes, they do. In fact, they have several competing committees to market the concepts internally through armies of PMs and TPMs and KPIs and efficiency metrics tracking. Also your AI token use of course.
reply
> Do big tech companies like FB and Google even pretend to be "agile" anymore?

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.

reply
agreed. zirp has become a meme, it does not really explain any current dynamics.
reply
> In the ~2010-2022 timeframe, tech companies poured all this money into speculative bets that never went anywhere

I think it is hard to overstate the effect that Waymo will have.

reply
1. If it scales worldwide. There's a decent chance that considering current global geopolitics, China will never allow them in, India and the EU will probably do the same or favor local competitors, Africa/Latin America will probably not have the money to tier 1 markets, which kind of leaves the US + Anglosphere.

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.

reply
I didn’t say that Waymo will be a commercial success or that there won’t be copycats more successful.

But waymo created the category by showing it is possible.

reply
Stadia was a minor footnote compared to Android, Pixel, and the other large organizations at Google. But there was plenty of hiring there during the pandemic, so your broader point is not wrong.
reply
the only thing I miss about stadia was the controller. Man, that was a nice controller. Also, playing games on a nest hub was a neat party trick
reply
Re: pandemic overhiring - haven't most of the big tech companies already corrected for that? Starting in 2022 we've seen very large rounds of layoffs multiple times a year. That's 3-4 years of layoffs now. Meta alone has laid off over 30k people since 2022, not to mention closing many open roles.
reply
Where did all the extra employees come from? Did this result in shortages in other industries that the employees supposedly transitioned from?
reply
IIRC, CS majors in universities grew a lot over the past 2 decades in particularly past couple years.

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...

reply
These West Coast tech companies had some of the highest salaries on the planet, so yes, they took employees from other locales/areas.

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.

reply
My experience working at Big tech companies is that people with roles like “agile coaches", "technical project managers", UX testers add questionable value. And the QA is usually outsourced to service companies like MindTree, TCS etc anyway.

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.

reply
What do you think about Cory Doctorow's theory that the AI produced code is going to come back to bite companies due to tech debt / unmaintainability?

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.

reply
I've heard OpenClaw got over 600k lines of code vibed over 80 days.

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.

reply
Claude Code is similar. It's fairly clean for AI coding standards but it's also most likely much, much bigger than what it should be for what it does.
reply
the near term is not an issue, because most ai code is still reviewed by experienced engineers with experience. the problem comes in the future where junior engineers who never acquired enough experience to handle engineering problems
reply
In the mature service I worked on, adding new code was “templatized”, you had to add feature flags, logs etc which didn’t vary much no matter which feature it was. The business logic was also not that complex, I can see AI tools one-shotting that and it indeed is a productivity boost. You would be surprised to know that most work was exactly this, writing rather mundane code. Majority of the time was spent coordinating with “stakeholders” (actually more like gate keepers) and testing code (our testing infrastructure was laborious). This was at MSFT. There are lot of teams that are innovating at the frontier (mine wasn’t, at-least not technically), I don’t know how AI tools work in those situations.
reply
> My experience working at Big tech companies is that people with roles like “agile coaches", "technical project managers", UX testers add questionable value.

"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.

reply
At MSFT, Product Managers were Technical Program Managers. Yes, a good PM is a joy to work with.
reply
>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.

reply
> 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.

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.

reply
The number of tech employees worldwide in 2026 is probably 2x or more compared to 2001. Maybe even 3x. Things are not the same.
reply
My observation is mag7 got less picky in hiring.

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.

reply
Pulled from other technical fields, leading to them having an ageing workforce and struggling against far-eastern competitors.

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.

reply
Other industries and other countries. Until recently, other industries were struggling very hard to find developers for example.
reply
> notable exception of Apple

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.

reply
Meta has no real product moats at this point. Yes, many of us still use instagram and facebook on occassion - but I'm not giving it any data beyond "what short form slop content will I accept".

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.

reply
1 is certainly happening in agency delivery work.

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.

reply
Meta basically "pivoted."

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.

reply
Meta is fundamentally a media company and they're using AI towards that end. They don't seem to be productizing their tech beyond that in the medium term.
reply
Why is #1 hypothetical?

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.

reply
Because I think "1 employee can do the work of 3 now" still hasn't actually been demonstrated
reply
> Because I think "1 employee can do the work of 3 now" still hasn't actually been demonstrated

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

reply
Right I think orgs are unclear how to wield this yet though

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..

reply
I wonder what proportion[1] of knowledge workers believe they have at least one colleague who the business would be better off replacing with software

and how many of them are totally wrong, or right about it!

[1] and how it might be changing with new generations of models

reply
I think the bigger issue is employees who can largely just not be replaced.

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.

reply
Pretty sure it has for coding
reply
Personally, no, haven't seen it in professional settings. Colleagues using AI heavily did not show any superhuman development speed, but definitely increased review burden compared to writing code by hand.
reply
> f 1 employee can do the work of 3 now but Meta's TAM can't grow 300%

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.

reply
> Maybe in the world of WP-plugins/typo3 and other simple work, though even those are fairly complex

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.

reply
> 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?

reply
> 1. By making workers unnecessary (largely hypothetical right now?)

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.

reply
4. When the whole thing implodes, because at that point you need to keep the balance sheet as green as possible.
reply
They're not really in trouble, they're still printing money. And I didn't know why they would need an excuse besides "this will help us print slightly more money" which everybody already knows.
reply
Yeah, this is a justification, but still -- they save single digit billions doing this, while AI capex is $150B (same timeframe) and RL spend is $16B. Feels like you could make the same cut from AI capex and barely notice a difference.
reply
My gut feel was that you can't be right, but it looks like you are: cutting 8000 employees * $500k/year total cost to company (rough but useful ballpark figure) is "only" $4B.

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.

reply
Though not all of that capex is cash; there's a whole phantom wampum AI economy where the big players are trading promissory notes for compute that doesn't exist yet (and may never exist) some time in the future and booking it as revenue.
reply
Maybe you're thinking of AWS/Azure/Oracle? Meta isn't selling their compute.
reply
Meta plays that game too; they're on the hook to buy compute that CoreWeave has yet to build (and may never be able to build) which counts as "revenue" for CoreWeave and an "asset" for Meta even though no actual money or compute has changed hands.
reply
Meta promised to buy dc capacity for ai workloads. If I remember it correctly, it created a common company with an investment fund as well that took on debt to build capacity.

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.

reply
These games are far more dangerous to the industry than "AI now = badcode future".

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.

reply
The figure I've seen floated is $6B, but yeah, same idea.
reply
It’s slightly more nuanced than layoffs = capex. You’re right, they don’t. That said, they do create free cash flow, which the market uses as one important input into the value of a given stock. Moving FCF positively when capex spending is moving it the other way is the real financial accounting move that is happening here.
reply
salaries are opex, data centers are capex, you can't compare them in the same timeframe.

4B over 5 years is 20B, which is significant.

reply
4B over 5 years is still 4B. It's 0.8B/yr
reply
If I read it right, the 4B ballpark figure is based on total annual per employee cost of 500k * 8000 employees, so the figure is actually 4B/year. 20B over five years.
reply
I think any company that is seen to reduce capex right now is going to be the Bear Sterns of this cycle
reply
Is this because you think the market will short sell them, or because capex is so worthwhile right now that a company which doesn't invest will fall fatally behind?
reply
The former; I think there's simply not the capacity to build the infrastructure that would be required to significantly improve the current tech as much as is expected (and there may never be the capacity to build that much infrastructure).

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.

reply
Doing slightly less than 150b looks bad to investors. Or at least it looks small.
reply
Imagine the productivity gains if they just spent $150B on booze and cake for employees!
reply
That is one of my huge complaints about the current levels of AI investments. You can do pretty much anything you want if you got $150B to spend, and then you go burn it on being uncompetitive in the AI space. Then you have the $80B Metas Reality Labs spend on a failed virtual reality and a pair of Ray-Bans.

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.

reply
And coke. Not the cold drink.
reply
you get another taste of coke every time you clear 100 tickets. not the cold drink.
reply
In other words: Nothing has changed.

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.

reply
I think in several ways the promises of AI to leadership is taking jobs not what AI is actually doing.
reply
Absolutely.

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.

reply
I mean AI was auto-warning me for a Facebook comment I made, which was a somewhat widely known internet meme. Long story short, video of a young girl (maybe 8 or 9 idr) with a GIANT spider on her hand. I wrote the words "Put that spider down, we have to burn down the house" and Facebook's AI flagged it, I tried to appeal, a human allegedly reviewed it and denied it, because they didn't understand a subtle meme. That's when I knew, I had to stop logging on to Facebook altogether.

This was in 2022... So I cannot imagine how much worse the "automoderation" will probably become.

reply
The WWeWork model of being in real estate and not tech, I guess.

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.

reply
Funny how AI took all the jobs, but not from automation
reply
Yup feels like it.
reply