> Instead, they should focus on using AI to drive revenue.
There is a complete disconnect between wages of employees and company's revenue => Why aren't employees working towards revenue? What a mystery. Children, let's help Elmo solve this mystery.
And then random mass layoffs to make numbers for shareholders look great in quarterly reports. Surely this motivates people work to their fullest potential and to care for company's revenue.
This is almost entirely on Anthropic and the stupid C suite people trying to push TokenMaxxing. GPT5.5 is much more token efficient, other models are much cheaper, and if used in moderation rather than than trying to get everyone to OpenClaw 24/7 with token leaderboards, it's much more economical.
Also ironically, a lot of GenZ and young Millenials who were already bitter at their employers have used the tokenmaxxing push to sabotage the AI rollouts by burning tokens on stupid shit. It seems to be working.
Just a week ago, Anthropic barely breaking even was hailed as AI companies being close to profitability much earlier than forecast.
In fact it is all smoke and mirrors, pure mania from C-level executives out of their depth trying to one-up each other with company money, and they aren't even close.
CFO: “Um, this is costing a ton and we’re not seeing savings or efficiency materialize.”
CEO: “Are we getting any value out of this?”
COO: “Not really, and frankly I’m getting annoyed at all the AI slop turning up all over.”
CEO: “OK, well, let’s do a big layoff and then I’ll just say it was because of AI. Hopefully folks won’t blame me for the mess and I’ll just talk about how amazing AI is.”
"An AI consultant tells Axios one of their clients recently spent half a billion dollars in a single month after failing to put usage limits on Claude licenses for employees."
Like physically, how could this even happen?
Just wait until companies are dependent on on it. When their employees can't think without it. When their AI generated codebase is such a mess they'd need a rewrite to understand it without AI. When they've got AI embedded in all their internal processes and tools. Then massive price hikes will come because they've been bent over a barrel and they'll have no alternative that isn't at least as painful in the short-term as letting the AI company fuck them. The long term won't matter then because any company capable of seeing past the short term wouldn't let themselves get into that position in the first place.
Overheard recently: "Thanks to AI we're producing more code and more MRs, faster than ever, but the milestones aren't getting hit any sooner. Actually the opposite, if anything."
I wonder how widespread that phenomenon is. Perhaps it's no wonder the prominent actors are trying to rush to IPO...
Well, if AI has a massive sticker shock attributed, so we should target the high value roles and should save money, right?
So im looking at CEO, CTO, CFO, and all the chief-something-officer. If LLMs are that totally amazing at thinking, then we should be targeting upper management, not the workers.
That would save a LOT of money for the shareholders! /snark
>Corporate leaders are starting to question whether soaring AI spending is delivering meaningful returns.
This isn't surprising. Ive recently run into quite a few rabbit holes where AI is bad enough that its much more efficient to do it myself. I wanted to refactor some code, gave it a design pattern to go towards, some specific classes and methods, etc. making it a well described problem. AI just couldn't do it satisfactorily. The code was ugly, overly verbose, and after multiple tries with multiple prompts saying to keep things simple. They still would introduce new classes, useless fields, etc.
When tokens get correctly priced, all of the insane over-investment in capital will need to draw back: buying data centers, semiconductors, and politicians.
Even then, it won't be right-priced with regard to actual costs. The environmental impact should have been priced in from the beginning. There seems to be a parallel with subsidizing fossil fuels, under pricing them which encourages over dependence, ignoring the real costs society will pay later.
We use GH Copilot at work and this week sat for a presentation by GH about optimizing token usage and maximizing ROI on tokens used. Anyone else get this presentation? They didn’t have time for questions because they had to run and give it to the next big enterprise on their list…
They basically said that everything is too expensive, you have to watch it like a hawk. It was as if they poured a bucket of cold water on the room. People were wondering how they could do anything faster with all these strategies. And then “sorry no questions. Bye!”
I really can't tell what is going on with AI these days. I hear AI labs claiming theyre profitable or close to it. I hear companies say they're dubious the juice is worth the squeeze. I've seen anecdotal claims of a measurable increase in productivity of 2x in PRs created and merged coming in at cost 20% of engineering employee budget. Others say they're still getting no value (which I doubt). Simon Willison's recent post went into debunking the AI sticker shock claims somewhat. Either way this seesawing between new golden era and the greatest VC money furnace is becoming exhausting.
I'd like to see real numbers at this point, and this article is just a few bullet points that link to other articles. Talk is far cheaper than tokens and I'd like to have a workflow that I can rely on being there in six months.
I wish I could do the same thing. Coworkers would be allowed to ask me X number of questions per month, and once they hit that limit I get the rest of the month off.
This is because the current AI approach relies on AI to be a glorified search engine – know everything about everything requiring enormous, ever growing models, and demanding search-engine like near instant responses requiring bigger more complex chips and sprawling data centers to run them in. This leads to a loop demanding ever bigger models, updated at a more and more expensive cost, and chipsets that become much more expensive to deploy.
If you move those things to software and utilize tools that are cheap at scale (databases, web search etc.) the hardware arms race ends and the price becomes sustainable. With the right tools preparing dynamic context for a conversation, models are used for their reasoning and not for their knowledge. And waiting even a minute or two for a model to prepare a response, evaluate it, and iterate to improve quality makes a huge difference.
"“It’s a real dollar investment,” says Tan, speaking for Claudeholics who go full blast. “You actually have to spend six to seven figures on tokens—I’m on a run rate to do seven figures this year.”"
The AI fever pitch has done a great job at exposing which companies were run with a degree of sanity, versus who bought blindly into the hype train narrative of worker replacement and went all-in.
Look, LLMs thrive when they’re given structured data that’s well annotated, clear direction, and treated as the probabilistic machines they are. Not one of those meshes with the AI narrative of “works on existing stuff, requires minimal guidance, and can behave deterministically.”
I said as much in 2024 when my employer at the time was grading folks on AI usage while my role was entirely deterministic in nature. It didn’t resonate with specific leadership then, it doesn’t seem to be doing so in the larger market now, and unfortunately not one of these dolts will suffer any consequences for their organizational myopia.
If I understood correctly, a few months ago Anthropic and OpenAI both started charging per-token billing at API pricing for Enterprise customers? i.e. representing roughly a 10x price increase? That's kinda nuts.
42 comments
[ 4.2 ms ] story [ 53.7 ms ] threadThere is a complete disconnect between wages of employees and company's revenue => Why aren't employees working towards revenue? What a mystery. Children, let's help Elmo solve this mystery.
And then random mass layoffs to make numbers for shareholders look great in quarterly reports. Surely this motivates people work to their fullest potential and to care for company's revenue.
Also ironically, a lot of GenZ and young Millenials who were already bitter at their employers have used the tokenmaxxing push to sabotage the AI rollouts by burning tokens on stupid shit. It seems to be working.
We should start to question whether soaring CEO salary spending is delivering meaningful results.
In fact it is all smoke and mirrors, pure mania from C-level executives out of their depth trying to one-up each other with company money, and they aren't even close.
CEOs: “Get me some of that GenAI”
CTO: “OK, we have all the GenAI”
CEOs: “Employees, it’s AI or bust”
Employees: Tokenmax
CFO: “Um, this is costing a ton and we’re not seeing savings or efficiency materialize.”
CEO: “Are we getting any value out of this?”
COO: “Not really, and frankly I’m getting annoyed at all the AI slop turning up all over.”
CEO: “OK, well, let’s do a big layoff and then I’ll just say it was because of AI. Hopefully folks won’t blame me for the mess and I’ll just talk about how amazing AI is.”
I wonder how widespread that phenomenon is. Perhaps it's no wonder the prominent actors are trying to rush to IPO...
So im looking at CEO, CTO, CFO, and all the chief-something-officer. If LLMs are that totally amazing at thinking, then we should be targeting upper management, not the workers.
That would save a LOT of money for the shareholders! /snark
We all know why they wont.
This isn't surprising. Ive recently run into quite a few rabbit holes where AI is bad enough that its much more efficient to do it myself. I wanted to refactor some code, gave it a design pattern to go towards, some specific classes and methods, etc. making it a well described problem. AI just couldn't do it satisfactorily. The code was ugly, overly verbose, and after multiple tries with multiple prompts saying to keep things simple. They still would introduce new classes, useless fields, etc.
When tokens get correctly priced, all of the insane over-investment in capital will need to draw back: buying data centers, semiconductors, and politicians.
Even then, it won't be right-priced with regard to actual costs. The environmental impact should have been priced in from the beginning. There seems to be a parallel with subsidizing fossil fuels, under pricing them which encourages over dependence, ignoring the real costs society will pay later.
They basically said that everything is too expensive, you have to watch it like a hawk. It was as if they poured a bucket of cold water on the room. People were wondering how they could do anything faster with all these strategies. And then “sorry no questions. Bye!”
I'd like to see real numbers at this point, and this article is just a few bullet points that link to other articles. Talk is far cheaper than tokens and I'd like to have a workflow that I can rely on being there in six months.
https://simonwillison.net/2026/May/27/product-market-fit/#th...
The mood has gone quickly from “this is cool” to “screw AI and any business that wants to use it”
This is particularly clear among the taste making class
If you move those things to software and utilize tools that are cheap at scale (databases, web search etc.) the hardware arms race ends and the price becomes sustainable. With the right tools preparing dynamic context for a conversation, models are used for their reasoning and not for their knowledge. And waiting even a minute or two for a model to prepare a response, evaluate it, and iterate to improve quality makes a huge difference.
https://www.wired.com/story/how-ai-agents-plunged-tech-world...
Look, LLMs thrive when they’re given structured data that’s well annotated, clear direction, and treated as the probabilistic machines they are. Not one of those meshes with the AI narrative of “works on existing stuff, requires minimal guidance, and can behave deterministically.”
I said as much in 2024 when my employer at the time was grading folks on AI usage while my role was entirely deterministic in nature. It didn’t resonate with specific leadership then, it doesn’t seem to be doing so in the larger market now, and unfortunately not one of these dolts will suffer any consequences for their organizational myopia.
https://archive.ph/crTG8
If I understood correctly, a few months ago Anthropic and OpenAI both started charging per-token billing at API pricing for Enterprise customers? i.e. representing roughly a 10x price increase? That's kinda nuts.
Similar discussion yesterday:
https://news.ycombinator.com/item?id=48296794