Are you willing to pay $100k a year per developer on AI?
- Reference: 1755255607
- News link: https://www.theregister.co.uk/2025/08/15/are_you_willing_to_pay/
- Source link:
It's a win-win if your title starts with a C or you're a stockholder.
Companies deny they're doing this, of course. Take Microsoft, for example. CEO Satya Nadella claims AI tools like GitHub Copilot now [1]write up to 30 percent of Microsoft’s software code . Simultaneously, [2]Microsoft has laid off over 15,000 people , nearly 7% of its workforce. Coincidence? I think not.
[3]
The funds Microsoft is saving from all its ex-staffers are helping to pay for Microsoft's spending $75-80 billion on its AI CapEx this year.
[4]
[5]
There's only one little problem with all this. This presumes that 1) AI can successfully get work done and 2) AI will keep being cheaper.
As for the first, of course, AI can replace some workers. Call center help staff? Maybe. That may not save as much money as company executives think. We've been shipping call center jobs offshore for decades. This is another step in a long-established work pattern. There's nothing new here.
[6]
The real savings, though, is to get rid of developers, engineers, designers, you know, people like you and me. Once you've separated the truth from the chaff of AI hype-spam, the evidence that AI can really deliver value becomes much less clear.
I find it telling that, according to the [7]2025 Stack Overflow Developer Survey , 84 percent of programmers [8]now use or plan to use AI tools in their workflows . 46 percent of AI-using developers don’t trust their results. And, even more interestingly as AI developer tools have been "improving," programmers are trusting them less than ever.
Why? Because instead of writing code, they’re spending - wasting? - a ton of time fixing AI coding blunders. This is not a productive use of mid-level, never mind senior, programmers.
[9]
Take AI's latest and greatest release: GPT-5. OpenAI’s CEO Sam Altman calls GPT-5 “ [10]the best model in the world .” Funny. GPT-5 will confidently tell you that [11]Willian H. Brusen is a former US president . For those of you not from the States, there's no such person, never mind a former president. Serious GPT users were so thrilled with GPT-5 that they demanded, and got, [12]OpenAI to bring back the older, but more reliable, GPT4o model .
The verdict of some users is in. They hate GPT-5.
Let's make this even scarier. Suppose that the current state of AI is as good as it gets? That's not just me being cynical. Both a recent [13]Apple and an [14]Arizona State University study indicate that our current ways of improving LLMs have gone as far as they can go. Sure, you can increase the tokens and throw more GPUs at training, but to quote from the Arizona State paper, "Our results reveal that CoT [Chain of Thought] reasoning is a brittle mirage that vanishes when it is pushed beyond training distributions."
Still, AI is good enough to economically replace knowledge workers, isn't it? Isn't it? Nope.
First, despite all the AI leaders' marketing-slop, as the Economist recently pointed out, [15]only 10 percent of firms are using AI in a meaningful way . AI, in short, is not as big a deal as the insane stock market would have you believe.
Moreover, and this is the real killer, customers are not paying anything like AI's real cost. Today, every AI company is selling you AI at loss leader pricing. As writer Ewa Szyszka for the [16]Kilo Code blog observed , people have been assuming that since "the raw inference costs were coming down fast, the applications' inference costs would come down fast as well but this assumption was wrong."
What that means is that today's more advanced "models can require over 100x compute for challenging queries compared to traditional single-pass inference." Compute isn't cheap. So AI-enabled code editors, such as Cursor and Claude Code, are replacing their introductory $20 a month plans with $200 a month plans. All this as [17]vibe coding's reputation continues to circle the drain .
Oh, and those plans with the tempting low-prices? They usually come with token limitations that make them far less powerful than those you get at a higher tier.
Of course, Sam Altman, OpenAI's CEO, can predict "The cost to use a given level of AI falls about 10x every 12 months," but I believe that just as much as I would an Old West snake oil huckster who'd guaranteed me his miracle elixir would grow my hair back.
[18]Some signs of AI model collapse begin to reveal themselves
[19]It's fun making Studio Ghibli-style images with ChatGPT – but intellectual property is no laughing matter
[20]AI running out of juice despite Microsoft's hard squeezing
[21]Hey programmers – is AI making us dumber?
The AI developer analysis company DX has pointed out that "The real cost of implementing AI tools across engineering organizations often runs [22]double or triple the initial estimates , and sometimes more."
As Laura Tacho, CTO of DX, puts it: “We were just having a conversation about how many tools each of us personally are using on a daily basis, those are all like 20 euros a month or 20 bucks a month. When you scale that across an organization, this is not cheap. It’s not cheap at all.” For example, Justin Reock, DX Deputy CTO, said.” A single engineer might use GitHub Copilot for code completion, ChatGPT for brainstorming, and Claude for documentation, resulting in overlapping costs without centralized visibility." It all adds up.
So, what happens on the day that OpenAI, with a burn rate of $8 billion a year, and Anthropic, with a $3 billion burn rate, must make an honest-to-goodness profit? Good question. Smarter financial people than me, which wouldn't take much, call OpenAI's path to profitability " [23]an open question ." Anthropic faces the same doubts.
Other companies, like Microsoft and Google, are sneaking AI charges into your existing software-as-a-service (SaaS) bills. As Saas management company Zylo pointed out, "AI tools embedded in existing SaaS platforms can be [24]deceptively expensive . For example, Microsoft Copilot now adds up to $30 per user per month to Microsoft 365 subscriptions, while Google has increased Workspace prices but bundled … That makes it difficult to compare options side by side — and even harder to calculate the total cost of ownership."
My best guesstimate is that, at a minimum, you can expect to pay ten to fifteen times more for real AI work you're having done today to achieve the same results in 2026. Switching over to AI suddenly doesn't look like that much of a bargain, does it? ®
Get our [25]Tech Resources
[1] https://www.theregister.com/2025/04/30/microsoft_meta_autocoding/
[2] https://www.theregister.com/2025/07/25/microsoft_ceo_job_cuts/
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2aJ9ZmtVLpITvPuNhV1AgsAAAAEk&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44aJ9ZmtVLpITvPuNhV1AgsAAAAEk&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33aJ9ZmtVLpITvPuNhV1AgsAAAAEk&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[6] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44aJ9ZmtVLpITvPuNhV1AgsAAAAEk&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[7] https://www.theregister.com/2025/07/29/coders_are_using_ai_tools/
[8] https://survey.stackoverflow.co/2025
[9] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33aJ9ZmtVLpITvPuNhV1AgsAAAAEk&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[10] https://www.theregister.com/2025/08/07/openai_gpt_5/
[11] https://www.theregister.com/2025/08/08/gpt-5-fake-presidents-states/
[12] https://www.theregister.com/2025/08/13/gpt5_updated_again/
[13] https://www.theregister.com/2025/06/09/apple_ai_boffins_puncture_agi_hype/
[14] https://arxiv.org/pdf/2508.01191
[15] https://www.economist.com/finance-and-economics/2025/07/17/why-is-ai-so-slow-to-spread-economics-can-explain
[16] https://blog.kilocode.ai/p/future-ai-spend-100k-per-dev
[17] https://www.theregister.com/2025/07/25/opinion_column_vibe_coding/
[18] https://www.theregister.com/2025/05/27/opinion_column_ai_model_collapse/
[19] https://www.theregister.com/2025/04/14/miyazaki_ai_and_intellectual_property/
[20] https://www.theregister.com/2025/03/14/ai_running_out_of_juice/
[21] https://www.theregister.com/2025/02/21/opinion_ai_dumber/
[22] https://getdx.com/blog/ai-coding-tools-implementation-cost/
[23] https://lipperalpha.refinitiv.com/2025/04/breakingviews-openais-profit-trajectory-is-an-open-question/
[24] https://zylo.com/blog/ai-cost/
[25] https://whitepapers.theregister.com/
Re: Blind spot
elsergiovolador you have not disappointed with today's reply.
You (AI or human) are really unhappy with corporations' people, politics and profits in general. Just above paranoia levels, I feel.
Enjoyable and educational as always.
Replied from someone just above the scum in the pool of the corporate cess pit.
Re: Blind spot
That's just tosh I'm afraid:
"Hoard top talent" - twas ever thus, for the purpose of outcompeting others rather than paying millions just so others can't have them.
"Overpay them just enough" - not to enslave, but to retain. The employment market is a free one, so employees can move to competitors for 'just enough' more.
"Leave with a fat CV" - not if they've stagnated in the same role for 2 decades!
Your thinking is way more paranoid than it needs to be, it's purely about money!
Re: Blind spot
The idea that it’s “purely about money” and the market is “free” doesn’t survive contact with how these systems actually work.
“Hoard top talent” isn’t just paying for the best - it’s about denying that talent to competitors. Warehousing people in roles that don’t build portable skills is deliberate. If someone spends a decade inside one company’s bespoke systems, they’re far less valuable anywhere else. Stagnation becomes a retention tool.
“Overpay them just enough” also isn’t the free negotiation it sounds like. “Market rate” is a corporate euphemism for informal wage coordination - kept in check through industry salary bands, recruiter collusion, and the same benchmarking surveys used across competitors. You don’t need a cartel meeting when HR departments all drink from the same data well.
On top of that, in many sectors the “free market” exit ramp - starting your own business - has been narrowed or closed. In the UK, IR35 and similar rules have fenced off much of the services market, making it almost impossible for individuals to compete with large consultancies on equal terms. When employers aren’t competing for your skills, you can’t just create your own job - the ladder’s already been pulled up.
And then there’s the psychology. Workers tied to mortgages, school catchment areas, and family obligations will often stick with the devil they know rather than uproot for a 10-15% bump - especially when taxes, moving costs, and uncertainty eat the gain. Employers know it and design retention strategies around it.
None of this is new in human history. Control of labour has always been the foundation of power - from outright slavery to bonded servitude to today’s “captive” labour markets. The language has changed, the suits are sharper, but the underlying dynamic is the same: a C-suite that sees itself as owning “its” workforce in the same way plantation owners once saw theirs.
So yes, it’s about money - but it’s also about control: over the talent pipeline, over wage ceilings, over who can participate in the market, and over the psychological calculus that keeps people in place.
Gilded cage
> "hoard top talent so competitors can’t get it, overpay them just enough to keep them comfortable, then let them stagnate inside corporate process hell"
Microsoft has *long* been known for hiring some of the most talented researchers at very well-paid rates and sending them to work in their labs where they come out with some very interesting stuff.... which is almost always completely sidelined in favour of MS's vested interest in the status quo and churning out more of the same.
It's long been assumed that the motivation was the one you describe- to keep them away from rivals where their ideas may pose a risk to MS's market dominance.
I don't think that goes so far as intentionally letting their skills atrophy, that's most likely just a convenient side effect of something they'd be doing anyway.
Re: Blind spot
...then let them stagnate inside corporate process hell until their skills atrophy. After a few years, they’re released back into the market with a fat CV but out-of-date skills...
Yes!
It's just like in hospitals: top-level surgeons are being hired to cut out moles, so the competition can't get them.
They are being kept there, tied to a mountain of cut moles and peeled bunions with a rope woven from dollar bills, bitterly crying over their vanishing skills - while the competing hospitals watch from the outside, sobbing, their warm tears plonking two drops at the time on the tip of their alligator leather boots.
What an absolute pile of hot, steaming bull biofuel aggregate. Someone is attacking the bottom, drill in hand.
"Funny. GPT-5 will confidently tell you that Willian H. Brusen is a former US president. For those of you not from the States, there's no such person, never mind a former president."
I'm sure in no time at all search engines will all agree that there was such a person and that he was US president.
If Prokofiev were still around he'd write an orchestral suite about him.
All your presidents are belong to us.
Most likely ChatGPT will tell you that Prokofiev *is* still around and link to the [1]suite he wrote about Brusen .
[1] https://m.youtube.com/watch?v=To7wHvDFu2M
The bubble is a-bursting
https://www.reuters.com/world/europe/powerful-new-ai-models-knock-wind-out-european-adopter-stocks-2025-08-15/
The markets are becoming hawkish about companies who have "bet the farm" on AI..
But I suspect Rotters has the wrong analysis here. It's not so much because the models are improving rapidly (so apparently early-adopters will be left behind) but because it is NOT improving - if anything the more recent models are getting worse (and becoming "ever more powerful" only in terms of how much energy they waste) and everyone is starting to move out of the hype-fog and realise what the "fundamental limitations" of bullshit generation are.
If the markets are punishing AI customers now, then they will go for AI suppliers next.
Call it crack and get it overwith
Yeah, that's some good s**t man, do you want some more? I got a little more, but it will cost ya a bit extra...
reminds me of a Robin Williams skit a the NY Met, he talks about shoe dealers giving away basketball shoes. but only one shoe, the other costs you! Try that one on, do you like it, the other will cost you...
Hook'm/pluck'm - part 2
My [1]cynical comment seems to be much closer to the truth than one might think.
Soon these AI companies will make sure you must use their "service" by law...
[1] https://forums.theregister.com/forum/all/2025/06/20/github_begins_enforcing_premium_request/#c_5094479
It's a win-win if your title starts with a C
That would be the C that rhymes with Hunt.
The Artificial Ignorance bubble was always nothing more than a hype-bubble. "AI" has failed miserably at almost everything I tried to do with it, requiring multiple models and many days of prompting just to figure out how to set up multiple JPA repositories in a Spring application. That's not "efficiency" - I could probably have done the same with online searches in about the same amount of time, but I'd never really tried to push any of the "AI" systems to actually create complex code, and I wanted to see how it would do.
It was very misleading, producing erroneous and bug-ridden code styles along the way that completely refactored what the previous LLM had come up with, but not actually fixing anything in most cases.
"AI" is statistics and averages gone mad. It produces plausible output for a prompt, not correct output, and no wonder - it's not actually "intelligent" in the slightest, with no understanding model of what you are actually trying to accomplish, just the "knowledge" that "90% of the time you ask for foo , you want code that looks like bar ."
I could probably have done the same with online searches in about the same amount of time
Have you tried to search for anything lately?
I’m willing
to charge them $100,000 per dev for my data.
Putting the AI in FAIL
Ed Zitron's series of articles about the AI hype bubble, the colossal burn rate of AI companies with little revenue to show for it, and the gradual enshittification of AI are worth a read.
https://www.wheresyoured.at/
Deja vu
I vividly remember the "new economy" during the 2nd half of the 1990s. Venture capitalists would sink million$ into companies that had never even designed, much less made, a single product, and tiny start-ups would go IPO and be valued astronomically overnight.
None of that made sense to me, but Those In The Know told me at the time that that was because I lacked vision and business acumen. This was the New Economy, and this was how it worked. Silly me.
And then the wheels came off.
Looking at the current AI playing field, I get a hauntingly familiar feeling. Value and investments are once again purely based on hype rather than actual performance or commercial viability. I suspect it will end the same way as it did the last time.
the AI "cost-savings" will disappear
At the same time as there are no developers left who actually now how to cut code.
I look forward to the argument that the 'AI' must physically start coming into the office.
Any 'AI' that wants to work from home is a slacker!
Blind spot
The author has a massive blind spot here - layoffs aren’t always about AI at all. Big corporations have been running a well-worn playbook for years: hoard top talent so competitors can’t get it, overpay them just enough to keep them comfortable, then let them stagnate inside corporate process hell until their skills atrophy. After a few years, they’re released back into the market with a fat CV but out-of-date skills - useless to rivals.
AI hype is just the latest convenient cover story for this cycle. It lets execs frame the headcount shuffle as “efficiency” rather than a deliberate competitive chokehold. The real game isn’t replacing humans with machines - it’s making sure no one else can get the humans they need to compete.