Revenue growth without employment growth? But Amazon spent $78 billion on capex in 2024 and more in 2025. So companies that get this cash hire more people than Amazon sheds?
By Wolf Richter for WOLF STREET.
Amazon announced in a blog post that it would reduce its global corporate workforce by 14,000 roles, “reducing in some areas and hiring in others,” it said, “to get even stronger by further reducing bureaucracy, removing layers, and shifting resources to ensure we’re investing in our biggest bets and what matters most to our customers’ current and future needs.”
It will offer most affected employees “90 days to look for a new role internally (the timing will vary some based on local laws), and our recruiting teams will prioritize internal candidates to help as many people as possible find new roles within Amazon,” it said.
But the net reduction would be 14,000 roles, or roughly 0.9% of its global headcount of 1.556 million full- and part-time employees as of December 31 (not counting contractors), or about 4% of its global corporate workforce of about 350,000.
Amazon is a global company with a global workforce, and it did not specify how many of these to-be-eliminated corporate roles are in the US, but it may have made reference to other locations with the phrase, “the timing will vary some based on local laws.”
Why? AI & Corporate Efficiency – amid strong demand.
AI: “This generation of AI is the most transformative technology we’ve seen since the Internet, and it’s enabling companies to innovate much faster than ever before (in existing market segments and altogether new ones),” said the blog post, signed by Beth Galetti, Senior VP of People Experience and Technology.
Corporate efficiency: “We’re convicted that we need to be organized more leanly, with fewer layers and more ownership, to move as quickly as possible for our customers and business.”
Amazon had been a hiring machine: It had gone in five years from 341,000 employees at the end of 2016 to 1.608 million employees at the end of 2021. Over the two years of 2020 and 2021 alone, it had added 810,000 employees, more than doubling its headcount in just two years! This type of all-out hiring not only by Amazon but lots of other big companies was in part responsible for the “labor shortages” in that sector.
Since then, Amazon has announced a series of global work force reductions to undo some of the over-hiring, and it has made an all-out push to use AI and automation across the company to do more with fewer employees. So the headcount dropped in 2022 and again in 2023. But in 2024, Amazon once again increased its headcount and ended the year at 1.556 million. By December 2025, the headcount may have dipped again.
Revenue growth without employment growth?
At the headcount peak in 2021, Amazon booked $470 billion in revenues. By 2024, with 52,000 fewer employees, revenues had soared by 36% to $638 billion. That is the result of the corporate drive for efficiency – including automation and AI.
Amazon, after the over-hiring binge, is doing more with fewer employees through process improvements, automation, and AI. It could continue the net headcount reduction even as revenues increase at a substantial clip.
Amazon has been throwing tens of billions of dollars at capital expenditures every year: In 2023, $48 billion; and in 2024, $78 billion. It expects capex in 2025 to be higher still. These investments went “primarily” into technology infrastructure (AWS data centers and the like) and “additional capacity to support our fulfillment network,” according to its 10-K filing.
These billions of dollars of capex are revenues for other companies, such as construction companies that build AI data centers, companies selling data center equipment, server makers, chipmakers, companies manufacturing the robotic equipment at Amazon’s fulfillment centers, software providers, etc. etc.
The hope is that these other companies will hire more people in the US with the Amazon cash than Amazon lays off in the US. The hope is that it would all work out, sort of (though a laid-off Amazon corporate worker may not be the best fit for a construction company).
In the past, technological innovation has led to continued employment growth, but many people who lost their jobs to technological innovation had a hard time finding equivalent new jobs that they could do.
But that theory that AI will increase overall employment, based on past observations about technological innovation, has yet to be proven.
Every company is different and makes different decisions. Over-hiring during the pandemic by portions of corporate America, especially tech, led to serial layoffs starting in late 2022, amid AI implementation and improved automation.
Walmart last month said it wants to freeze overall employment at about 2.1 million for three years, but get more efficient, thereby cut costs, while pushing up revenues by incorporating AI tools and agents into nearly all its roles, with its white-collar jobs the first to be affected. It committed to train its employees in using AI tools and agents, and those who cannot make the shift will be replaced. “It’s very clear that AI is going to change literally every job,” CEO Doug McMillon said in an interview with the WSJ in September.
Last year, Intuit made similar headlines when it said that it would lay off approximately 1,800 employees in non-AI related roles (over 1,000 of those due to performance), while planning to hire at least that many new folks for AI-related roles, with essentially no change in employment.
A word about the rumors-based headlines yesterday evening…
Amazon’s 14,000 reduction of roles globally, and not necessarily in the US, dispels the clickbait headlines yesterday evening, all based on a Reuters article citing anonymous sources, “Exclusive: Amazon targets as many as 30,000 corporate job cuts, sources say,” which generated a flood of similar stuff across the internet, such as: “Amazon may lay off 30,000 employees this week,” or “Amazon to cut 30,000 corporate jobs in largest layoffs since,” or the WSJ headline, “Amazon to Lay Off Up to 30,000 Corporate Workers,” which the WSJ tuned down this morning after Amazon released the blog post, to “Amazon Lays Off 14,000 Corporate Workers.”
Enjoy reading WOLF STREET and want to support it? You can donate. I appreciate it immensely. Click on the mug to find out how:
Did you mean to insert a chart at “*** chart”? Thanks for the article!
The chart was late, got hung up in rush hour traffic because AI has taken over traffic lights. It made it now.
Looks like the tide has turned and companies are no longer nay-saying AI.
Once agents have real power to spend money (authorize returns, change reservations, adjust staffing), then AI use will take-off.
People can call Chase and there will be no options. They just say what they need and the bot says no or it says yes and takes care of it. Done.
FYI- you have a chart placeholder ***Chart.
Lol. No one needed AI to make customer service crappier and less responsive, but it sure is a convenient excuse!
I’m surprised that Wallyworld has a much higher employment than Amazone! The breadth of one with more limited human resources is amazing.
Wallyworld has a lot of real estate to manage, which takes cashiers, stockers, etc… So, they have all the same warehouse/distribution costs as Amazon but with the added burden of retail stores (of which Amazon has only a handful).
And all of the C-suite execs and shareholders can enjoy delicious hyperbubble profits, while workers lose their jobs and fall into abject poverty.
Of course the money saved by greater corporate efiicicency will be spent elsewhere thus creating jobs.
Improved efficiency has not led to mass poverty during the last 200 years of capitalism. That is unlikely to change now.
You are aware of the staggering amount of debt that has been issued since 2009+/- GFC and specifically since 2020…
(Debt out the Wazoo per Wolfs charts)
We are in the middle of the greatest monetary experiment known to man… to presume the last 15+/- years shared any resemblance to historical economics (last 200 years) is a bit disingenuous imo…
Time will tell, always does..
I watched a podcast the other night that AI is already coming for the execs and managers. Hamza Mudassir on AI outperforming CEOs and creating exec digital twins.
Although undoubtedly AI and automation has something to do with it, I suspect this move has more to do with Amazon overhiring during the pandemic (they went from a headcount of 800K in 2019 to 1.6M in 2021) and now they are trying to right-size the ship.
Note that they are not unique in this respect. Many other tech-focused companies also went through a massive hiring boom during the pandemic.
The “self-checkout” lane is even more insulting than any efficiency brought about by AI. Not only did the store get to fire someone, but now they are getting you to work for free…
No surprise Amazon the the grocery delivery apps are doing so well. It should be clear to everyone, with the shutdown in day 24, that the real corporate owners want everyone to work for free.
Self-check-out lanes have nothing to do with this. They’re just a shift of the cost of labor from the store to the customer. Baggers decades ago blazed the trail with this cost-shift to customers, who’ve been bagging their own stuff without demur ever since.
I believe that trend started in healthcare. Send the patient home faster and have the family take over post-procedure care.
It is many times in the patient’s best interest to leave the hospital as soon as reasonably possible.
It potentially keeps the costs lower for the patient, so they are less likely to incur debilitating med debt. Shorter stay equals less cost.
It potentially opens hospital space for patients that are more ill than the released patient.
And most important….The quicker the patient is out of the hospital, the less likely they are to contract other serious problems such as MRSA. The longer a patient stays, the longer the exposure to other serious illnesses that will complicate recovery.
And before I am destroyed here….
Emphasis on “reasonably possible”
Sandy,
That’s a good thing. Hospitals are depressing and deadly. The rule is: Go in, get the job done, get the hell out and recover at home. Much better outcomes. A problem arises when there is no support at home, for example, an elderly patient lives alone and doesn’t get help at home, including things like buying the right food.
Costco must be benefitting from that, assuming people are being honest during checkout. I have noticed As a check, Costco does match receipts to what’s in the cart before a customer leaves.
If they pass some of that benefit to the customer via lower prices, I’m not bothered by it.
If you consider scanning your groceries and putting them in a bag to be “work,” I don’t know what to tell you.
scanning groceries is work, it’s job, and people get paid for it, and feed their family with this work. You can still see them working over at the other checkout lanes.
I’m talking about from the perspective of the customer using self-checkout. I value my time more than I value the idea that doing someone else’s “work” is beneath me. If bagging one’s own groceries at a self-checkout is work, then so is taking items off the shelf to put in the cart.
I’d much rather bag the things myself at a self-checkout and be in and out in 30 seconds than wait on a line for a cashier to do it for me just so that I can gloat that I don’t “do unpaid work.”
It’s odd to me that people would rather waste time than do something as simple and efficient as check out themselves, especially with only a few items.
Even more insulting is when you don’t use the self-checkout but the checker doesn’t even move a muscle, and makes you bag your own.
UPS in on the race:
“Meanwhile, UPS said in its third quarter earnings results on Tuesday that it had cut its “operational workforce by approximately 34,000 positions” in the first nine months of the year as it looked to be more efficient, while about 14,000 positions, primarily in management, had also been eliminated.”
Except UPS has been losing revenue year over year, and its stock price is back to 2014 levels.
Someday Amazon will go bankrupt and be sold for scrap the way A&P, Sears Roebuck, Kmart and many other once dominant retailers have been. They need to do everything possible to maximize efficiency and financial strength now to delay thier inevitable end. New technology only exists if it makes something faster, better or cheaper, and AI is for reducing costs and cutting expenses including people. Same with Walmart. It will be gone someday too when another company does it better.
“Someday” we will all be dead.
Lord Acton…
There seem to be more “hiring down” than “hiring up” headlines, due to AI, this year. Klarna has halved its headcount, due to AI, for example. The direction of travel for human recruitment is starting to feel negative, not positive.
Maybe if companies weren’t dumping billions at a time on the AI fire for almost no returns they wouldn’t be “hiring down”. They claim the need for trillions of spend on power and data centers while bringing in a tiny fraction in revenue. AI is proving to be little more than a more thorough search engine and all of its suggestions and results have to be double checked for accuracy by someone competent. When models start to ingest garbage input from other AI models, look out below.
Amazon is making lots of money off AI via its AWS division. Nvidia is making huge amounts of money off AI, lots of companies are making lots of money off AI. Microsoft, Google, Intuit, etc. are shoving AI down their customers’ throats by charging them more for their services that now include AI, whether they want them or not. Google raised its price per small corporate email account by $1 per account, from $6 to $7, per month, so my two accounts combined went from $12 to $14 per month. That’s a 16.7% increase in revenues for Google due to AI. Microsoft has done the same thing with 365. Intuit has done the same thing with its Mailchimp email services, which it acquired. Businesses are paying for AI. And AI services providers are raking in the bucks.
Who is not paying for AI, or at least not directly, and not yet, are retail customers.
CEOs at many companies follow the trend of what other execs do. Many CEOs have little idea of what different people do or what issues are being addressed. But the current CEO action is AI and employee cuts. These decisions are often the actions by leaders who read about current trends in exec magazines. And the media stories don’t clarify what is really happening or why.
CEOs are CEOs because they surround themselves with good people who have good teams who spend years working through these issues of efficiencies, step by step. CEOs of companies like Amazon don’t get to be CEOs of Amazon by “reading about current trends in exec magazines” 🤣
I don’t know a lot of CEOs personally, but I have seen a lot of bad managers, across multiple industries.
It’s not at all uncommon for people to be hired for a job based on personality/viewpoint or relationships with the people around them.
Merit can have something to do with it, but it often seems to be a surprisingly mixed bag.
Fortunately most of the bad managers were from pretty far back, but it was certainly an eye-opening experience.
Sure there are bad managers, everyone knows them, or thinks they know them, but they’re unlikely to become CEO of Amazon.
Of course, if you’re only talking specifically about companies like Amazon, then that’s a whole other kettle of fish.
Amazon will eventually replace its warehouse workers with robots, considerably reducing even more of its workforce. All it will need are a few people to run the warehouses, people to deliver the goods (many of which are already contracted out), and a few nerds to manage everything.
Walmart has a lot of brick and mortar stores which need people to run them, e.g., cashiers, store-managers, and people to put stuff on shelves. Unlike Amazon, Walmart has to deal with retail theft.
Long term winner is Amazon, unless Walmart can get rid of its brick and mortar albatrosses. But then it would be just like Amazon.
People are cheap, robots are not. Robots break, and need more skilled technicians and engineers to repair/replace.
A car assembly line is one thing (same task a couple hundred thousand times per year). Using robots to address one-offs is inefficient.
Note I’m not saying they won’t push more automation, but we’re closer to the break even point of labor vs automation than most people think. So I’m not sure penetration will get that much deeper
Robots will be almost indestructible in a few years and much cheaper to make than now. I hope their software has a backdoor kill switch.
Robots are generally custom made for the application at hand. I had several in my plant in the 1980’s in manufacturing and as long as they performed their designed single task (not one that requires going from location to location and doing different things), they were fine. But I had a tech and an engineer in the wings of these machines. And, they ARE machines, and will continue to be machines.
If you look at a typical automotive assembly line , most robotic machines are used for spot welding, paint spraying and other very repetitive tasks. They are good for that, but require backup.
1. All manufacturing tasks are by definition repetitive.
2. I think you need to step into a modern assembly plant, something that came online over the past couple of years.
The robot supervisor will have the kill switch but he’ll have to clear it with the robot manager.
Walmart is a HUGE ecommerce retailer. That’s the end of the business that’s growing in leaps and bounds. At its stores in the US, groceries do well, the rest of the stuff in its stores is on a flat to declining trajectory.
But they still have nice sections which is good for one stop shopping.
Health, drugs, beauty and hygiene.
Auto, hardware, crafts, lawn and garden,
Clothes, shoes, household, kitchen, bedroom furnishings.
School, office supplies, camping, sports, outdoor stuff.
Computers, tvs, music, video, phone.
I remember the Walmart in Laredo Texas. It was beautiful and it was packed with customers. They must of had an army working overnight to keep it shipshape.
I’m just stopped at a Walmart and Sam’s Club in Texas and Oklahoma after twenty years living west of the Rockies. Walmart has upped it’s game especially their deli and bakery. Even the international aisle was impressive. Sam’s also had a fresh sushi stand with 5x the offerings of others.
Amazon got slammed (by consumers) during Covid. They hired like crazy in response to that. I am comforted by your charts indicating they may have over hired. On the other hand, I’m pretty darned sure that one entry level employee using AI can EASILY do the work of at least 5 entry level employees. I’m pretty darned sure this is going on across the retail sector. We are going to know in the next 24 months.
No, they can’t. Certainly not “easily”. The Gen AI can surely help with a lot of things and make some tasks easier, like all forms of automation before it, but in some ways it’s worse than previous forms of AI because by Gen AI’s basic nature it hallucinates and makes more errors. This isn’t “fixable” it’s basic to how it works and trained, like any other tech it has pros and cons and there truly are a lot of pros to it, esp being more flexible but the very thing that makes it more flexible on tasks also makes more prone to errors. And the worst kinds of errors, the subtle ones hard to catch. That’s why it’s often a serious problem in sensitive tasks have to be done with precision. It does add some value at first and if used carefully with human expert supervision, it indeed makes some things easier, but errors and hallucinations quietly start to build up, especially when models change and the AI starts to intake a lot of AI slop from previous AI’s. Inevitably some managers get deluded in thinking they can get brownie points for “look, I reduced head-count and replaced with trendy AI” and they start laying off people who can catch subtle errors and check the AI’s output.
Then suddenly things start to break and the company realizes their software or data files have been contaminated with junk and errors that got hallucinated in and built on because they foolishly thought the AI could do things alone with human proofreading and intimate knowledge of task needs. The AI is a sophisticated probability machine by design, it doesn’t know what’s right or wrong and subtle problems esp creep in unpredictably. The companies then have to spend millions to clean up the mess, some go bankrupt or get permanently damaged. (And often funny enough, re-hire staff they let go, with the most permanent layoffs being for any over-eager managers who thought they could just replace people with the hallucinating AI) And even if the AI didn’t have these problems, the ROI would be questionable to backwards at best, due to all the costs of the data centers that often passed onto communities seeing spikes in their power and water pipe bills. That’s why I strongly doubt Amazon’s layoffs really have much to do with the modern AI bubble and all the hype around it–automation sure, this has been going on for awhile, but this press release sounds more like Amazon trying to ride the AI bandwagon and sound hip to the hype with FOMO investors. While, the real reasons are more about offshoring and mainly, just still dealing with over-hiring from pandemic and doing cost-cutting, and tariffs probably also a factor.
It makes sense but also seems like AI is just used as a reason to eliminate talking to a real person in most circumstances. Honestly I actually prefer that but 90% of the time the AI can’t resolve an issue and finding your way to a human is really hard. Guessing AI will be the great downsizing excuse for the next 5 years, where in some cases it may be accurate, and in other simply a deflection.
Two things.
1) Shout this from the rooftops whenever a simpleton says technology creates more jobs than it destroys: “many people who lost their jobs to technological innovation had a hard time finding equivalent new jobs that they could do.”
2) AI agents will hit HR quite hard, and similarly affect other white collar roles whose principal work is managing information inside an organization. From a Harvard Business Review article about Hitachi automating HR with AI agents:
“An employee query to Skye triggers the agentic AI system, which sends the employee’s specific request to an intent classifier agent. Its job is to route the query to the appropriate AI agent.”
“For example, the intent classifier agent sends a simple policy question like “What are allowed expenses for traveling overseas?” or “Does this holiday count in paid time off?” to a file search and respond agent, which provides immediate answers by examining the right knowledge base given the employee’s position and organization. A document generation agent can create employee verification letters (which verify individuals’ employment status) in seconds, with an option for human approval. When an employee files a request for vacation, the leave management agent uses the appropriate HR management system based on its understanding of the user’s identity, completes the necessary forms, waits for the approval of the employee’s manager, and reports back to the employee.”
“Questions that require information outside of HR are also addressable by this agentic AI system. An employee looking to get IT approval for downloading a piece of software is routed to an IT help desk agent, which files a ticket in the correct IT service-management-software system, waits for the response, and then responds to the employee. A payroll question is managed by the payroll agent. The orchestration across these agents turns what was once a maze of systems into a seamless experience for employees.”
In both the situation of waiting for a manager’s approval for leave, and IT’s approval for a software download, AI can do the job itself. There is a decision tree of issues managers go through to determine if leave should be granted to a person on a date, as well as whether a piece of software is okay to download. AI can be trained on both and make the decisions immediately. This is why AI will crush white collar jobs.
Robots don’t need vacations or take sick leave or complain or have 401k’s or sex issues or harass other workers or get paid a wage. They need some minimal maintenance from time to time and software upgrades. If your job can be replaced by a robot, now or eventually, it is time to learn a new profession.
I think you’re overstating AI’s supposed benefits.
It would be nice if AI could get rid of the attorneys too.
That’s already been happening for years. In big law firms, a lot of the grunt work is already done by AI instead of hires newly out of law school. Briefs are being written by AI (with sometimes hilarious results), etc. Law firms are all over AI.
Confirming what Wolf said. We have a kid studying to be a paralegal and their classes are all about how to use AI integrated into LexisNexis so the field stays relevant.
People getting hallucinated results are because idiots don’t learn how to prompt correctly to avoid AI making stuff up. GIGO
This is bringing back memories of my X — when her new boss informed her she was going to train the new bosses pick to replace her. She quit that week — losing tons of cash, but it was fine.
This also brings up the issue of flattening the hierarchy and demoralizing workers, to a point, where inefficiency makes people expendable — and exchangeable — replaceable by software.
The idea of upskilling people is absolutely ridiculous — because, it’s just a step towards eliminating people, who won’t be able to keep up the pace.
Except technology does produce more jobs and it produces jobs that pay more and are less back breaking.
The very well-documented wealth inequality would strongly disagree with this.
Except when the robots get good at managing, programming, and repairing other robots. It should at least be considered whether this Robotic revolution is indeed different and will likely reduce the need for human labor. That is what all these businesses are betting on,
I wonder how in the larger context a growing economy without a growing workforce impacts everything. The prior assumptions were mostly that a growing economy meant growing employment and this consumer spending, payroll taxes, income taxes and so forth would keep pace. Feels like that ship has sailed.
Sounds like the 14,000 might just be the first wave.
From the internal Amazon communication that has been published in the news:
“ While this will include reducing in some areas and hiring in others, it will mean an overall reduction in our corporate workforce of approximately 14,000 roles…
…Looking ahead to 2026, as Andy talked about earlier this year, we expect to continue hiring in key strategic areas while also finding additional places we can remove layers, increase ownership, and realize efficiency gains…”
It’s that last sentence that keeps the door open for future rounds of layoffs…
This is a long-term project. They have been doing it, and they’re continuing to do it.
Southern California Edison will be starting layoffs after January 2026.
And it’s certainly not because electricity demand is down. I hope they’re letting their H-1b employees (that replaced American workers) go first. They got into a massive scandal over this a few years ago.
I received an email that rates will now be increasing 13% to help provide a more robust infrastructure. Never ending increases of cost of living in all directions in CA.
Increases in the cost of living do not occur only in CA. Think of the US post office. One of the better investments in recent years in limited amounts would have been “Forever stamps”. Thanks to the 1978 Proposition 13, my property taxes in CA are a real bargain.
The classic retort to these white collar layoffs is to just learn a trade. If we get a sudden influx of a bunch of blue collar workers doesn’t that just lead to the same low wage constant layoffs cycles that white collar work are experiencing with the automation of cognitive labor? I’m not sure what happens to the glut of people. Super sad. Unsure what to do about my own career at this point… feels like everyone’s waiting for something to break and offering no solutions
Even the trades are turning to technology. Auto techs have been using computers for many decades and AI more recently because diagnosing a problem correctly is the most difficult part of the job. Replacing the sensor is easy.
What career advice would you give a young person right now wolf? Something I’ve heard is to learn AI itself. Is it just a race to the bottom to escape an ever growing permanent under class as AI continues to eat away at industries?
I would still research trades. Plumbing, electrician, roofer, etc, etc, etc.
AI and technology will help these jobs diagnose and research, but for now, a human still needs to complete the task.
Agricultural automation and robotics has lots of greenfield opportunities (pun intended) and will actually benefit from AI. For example https://spectrum.ieee.org/lely-dairy-robots
People need to be very good at using the AI tools that apply to their roles. It’s not easy to use AI tools correctly. People need to be very smart about it, and not just surrender their work to AI.
In terms of career advice for the trades: Go for it if it appeals to you. But there is a lot of learning to do, steep hills to climb, and a lot of dues to pay. Good auto techs are always in short supply, and they make a lot of money. But it takes years to get to that level. In general, the money can be excellent for highly skilled and fast workers. But make sure you reach manager level or owner level before your body wears out. The trades can be hard on a human body.
In terms of AI, this is where I really feel my age: I refuse to use AI for this site. It will always be by a human for humans. AI will never be allowed to post articles on this site. That’s just how I want it. And I will go down swinging.
But I constantly have contact with AI when I interact with other companies, because they’re all using it, from my medical care provider to the software companies the produce the server software that Wolf Street runs on, and they’re mostly very good. Good AI tech support is far superior to human tech support, in my experience. I’ve run into a few things where improvements are necessary, one of which was a trial, and the company then discontinued it, likely went back to the drawing board, and will bring it back in new and improved form in the future. No one is sitting still here. Everyone is working on this stuff. Except me, it seems, with my quaint by a human for humans motto.
My family deleted the Amazon Prime account that we all shared. Our consensus is to eliminate shopping at Amazon as much as we can even if it costs more. We hate Bezos, Zuck et al.
Why?
Amazon just paid for the destruction of the East Wing of the White House. Bezos paid $40M to Melanie for a documentary of her life. The next highest bidder was Disney at $12M. Bezos’ Washington Post got censored and lost its soul.
As for me, shopping at Amazon is just so easy I can’t give it up. Welcome to the corporate fascist present and future.
I cancelled my Prime account years ago. I am very pleased that Amazon is paying out billions of dollars in penalties to the FTC and to its customers for its deceptive trade practices related to its Prime business. But I still continue to do business with them when their prices are competitive.
I’m biased as a tech employee but I got a feeling that the current generation of big tech companies is about to be outcompeted by a new wave of AI-using startups. I give it 3-5 years.
Big part of it is I’m seeing first hand how much more efficient an engineer can be with modern AI tools in a small team vs in a big bureaucratic company.
Another part is big companies destroying their employees’ good will with these layoffs.
Success in tech is all about having smart people with right motivation working hard IMO.
I work in engineering consulting and so work with many different large and small companies and I’ve yet to see AI gains. We have concrete deliverables and have to detail each hour we spend on a project. We had an ‘AI’ initiative to focus on using AI and the AI fan boys gave crazy gains estimations. But when the projects were complete they didn’t complete their deliverables 30% faster.
I’ve really only seen entry level engineers really using AI anymore and I’m not impressed with the speed or quality of their work. We recently just lost a project to one of these AI start ups, only for the customer to come back and say they had so low quality that the project had to be scrapped and had us do it.
Maybe other industries are different. But AI has had no impact on my industry. But finance bros keep telling me that actually they know my industry better than I do. As a side note, the only 2 people I’ve seen fired from customer projects(IE customer employees) have been extensively using AI and worked about twice as slow for a worse work output.
Again, maybe other roles see better gains. But I’m yet to speak to a human in person who claims they work 30% faster with AI. It’s always someone else who we ‘know’ can have their work done with AI. I think in reality often those people are just unneeded roles.
Agreed. I don’t see AI as anything revolutionary. It’s merely a much improved search engine that can aggregate results. It’s the next logical step in what has been improving for decades, not something different, in my opinion.
You’re just thinking of AI as the stuff Google dishes you up. But that’s just an example of what AI does. That’s just the surface in human-readable form. Think more broadly, think of the term machine learning. Look the Waymos driving around all over the place. That’s machine learning. When one Waymo vehicle learns something, all Waymo vehicles learn the same thing. They’re learning as one unit. And they’re amazingly good, far better than human drivers. They see out at all sides simultaneously. They see in the dark and bad weather. They can make complex multiple decisions at the same time in milliseconds. They’re never tired or distracted or angry. And they’ll do it better next time. That’s AI right in front of you. AI comes in many forms. And it’s already everywhere.
It really depends on the problem space in my experience. Old large systems utilizing obscure in house technologies are not very AI friendly, but try building a new product using well documented popular technologies and you’ll see a large boost from gen AI tools.
That is the pattern with major technology disruptions. The big companies have a hard time changing their business models (see Clayton Christensen’s The Innovator’s Dilemma) and they get their lunches eaten by newcomers who are built ground up on the new technology. I’ve mentioned this multiple times to my employer, we’re going to get smoked by a smaller AI first provider somewhere in the next 3-5 years. Advertising companies are already reporting that their clients are coming back and wanting to renegotiate contracts downward due to efficiencies.
If they don’t get their head out of the sand soon, I’ll jump ship to that smaller AI first provider and see them at the finish line.
and Fed rate cuts will not bring the jobs back….
but the Fed will point to layoffs as reason for cuts, though the numbers may not be large enough to justify such a move.
Is there not a glaring disconnect between interest rates and the promotion of employment? But it seems as long as the interpretation that lower rates will somehow reverse layoffs such as noted here, the game will go on. Because low rates pump assets
In theory, lower rates could spur capital investment, which could lead to job gains.
The problem is that companies have to have an idea of capital investment (that would create jobs) and that makes sense if they can borrow money at X% but not if they have to borrow money at Y%. I’m not overly convinced that it makes a huge difference, except at the margins.
Even if you can borrow money for free, you still need that idea.
SNAP in tranches might deflate AMZN and WMT. UPS 50K down.
Amazon growth has stalled. Most of the increase in revenue from 2020 until now can be attributed to inflation.
This is the insidious nature of inflation. It distorts widely used valuation metrics.
These cuts have little to do with AI and more to do with cutting overhead, in my opinion.
Nonsense.
1. Over the period that I mentioned (2022, 2023, 2024) Amazon’s revenues grew by 36%, and CPI rose by 12%.
2. Consumer price inflation is dominated by services that Amazon doesn’t sell at all, or sells very little of: housing, insurance, healthcare, etc. It sells some services, such as streaming, and a little bit of healthcare, but that’s only a small part of inflation. Amazon didn’t sell new and used vehicles and gasoline during the high-inflation years, but new and used vehicles and gasoline were the largest forces in the inflation of goods during the high-inflation years.
The big growth at Amazon comes from its services sales (data center and cloud). Last quarter, services sales jumped 13% yoy to $191 billion; product sales rose 7% to $132 billion. Amazon doesn’t sell data center and cloud services to consumers and consumer price inflation doesn’t apply to them at all.
3. You cannot adjust everything to CPI. That’s just BS. For example, GDP is adjusted to inflation via the GDP deflator that tracks all the transactions that go into GDP. GPD is NOT adjusted to inflation via CPI (which is just consumer price inflation).
While not directly related to Amazon gutting workers — as they invest in AI, in the background of society, it’s been shown in various studies, that prolonged AI exposure causes brains to atrophy — an especially interesting phenomena in younger people, in high school or college, that are being educated by hallucinations and low quality output that bombards them by the minute.
However, the adults in the room, who have jobs, and are being pressured to adopt and adapt to AI — their prolonged exposure has a different outcome, in the form of overload anxiety and burnout.
If your problem solving time is mandated to speed up, working your brain at higher frequency, is not unlike a GPU that overheats and has its lifespan shortened; it’s almost like addiction.
I think we’re in the infancy of disposable labor, with far faster turnover rates — but it’s fascinating that the younger, fresh replacements are already out to lunch. Good times ahead!
The founder of the company has recently reassured Earth:
“Bezos also batted away any doom and gloom about AI, adding that it can only enhance our lives.”
-What could POSSIBLY go wrong?
The “enhancements,” OH, the enhancements that we already experience!
I’m not generally a luddite, but this one is starting to turn me. I’m SURE there’s Great and Wonderful applications for the tech, I’m NOT so sure those will be primarily sought after.
there will be a huge crash when this AI bubble bursts, just like there was a huge crash when the internet bubble burst and investors lost huge amounts of money and lots of them got wiped out. But the internet thrived, it’s everywhere, bigger than ever, and has moved into every part of business and life, and people use it all the time and just take it all for granted. That’s the future of AI.
From the Yahoo Finance page:
“Nvidia CEO downplays AI bubble fears as he enlists new partners.”
Nvidia valuation closed just below $5 Trillion today.
That is just ONE TECH COMPANY.
Not only is this a bubble, it is 1999 on steroids.
You can search “Raymond Chen microspeak convicted” for a blog post about this word being used at Microsoft. Apparently manager types think “convicted” feels stronger than “convinced”, so they use the wrong word, and then it becomes part of the lexicon.
I could believe there’s a similar culture at Amazon.
It’s akin to Christian/ Biblical lexicon. As in “I felt convicted by the Spirit,” meaning there’s an Assurance, beyond oneself.
It’s a whole new world: my daughter is learning about irrational numbers. The AI response from Giggle said they are a “nonrepeating, transcendent decimal.”
The god of the Tech is whatever’s behind the creation?
Easy rule of thumb – all transcendental numbers are irrational, but not all irrational numbers are transcendental.
Piece of cake
Listen, we are living in AI first gen rollout. I was recently stuck in an endless Verizon AI tree making it impossible to speak to a human. But there is a trick to a fix, select new devices or billing and, surprise !, you get a human. Then merely explain your situation and they transfer you to human, they KNOW the problem and will tell you such.
thanks for the insight!
Wolf, is it a typo? Para headlined ‘ Corporate Efficiency ‘. Is it convinced (not convicted)
It’s how she phrased it. Some people here in the comments pointed out that this now common corporate speak.