Nvidia executive: The cost of AI tools is 'far beyond' the cost of human workers Fortune

[usonian]

https://fortune.com/2026/04/28/nvidia-executive-cost-of-ai-is-greater-than-cost-of-employees/

Bryan Catanzaro, vice president of applied deep learning at Nvidia, recently told Axios.

Meta announced last week in a memo that it plans to lay off 10% of its workforce, about 8,000 employees, as well as scrap plans to hire for 6,000 open positions. It’s part of an effort to “run the company more efficiently and to allow us to offset the other investments we’re making,” according to the memo. Microsoft has offered thousands of its own employees a voluntary buyout, the largest the company has ever offered.

Other tech headers, however, suggest that right now, AI isn’t saving companies money on labor; it’s actually costing them more than the humans they currently employ.

snip

An MIT study from 2024 backs up Catanzaro’s experience. Analyzing the technical requirements of AI models needed to perform jobs at a human level, researchers found that AI automation would be economically viable in only 23% of roles where vision is a primary part of the work. In the remaining 77% of the time, it was cheaper for humans to continue their work.

In other instances, AI has proved to be fallible, with one engineer saying an AI agent destroyed his database and network as a result of what he called “overuse.”

Paper? https://arxiv.org/abs/2603.29121

One suspects that an Nvidia exec is hinting that the balance will tip when Nvidia gets lots more money. Like Sam.




The Fortune nag screen can be beaten with a browser "reader" function, and possibly using archive.ph or turning off javascript. I used the first one.

[questionseverything]

1. Ai costs more, over heats the planet and uses all our fresh water 💦

What could possibly go wrong?

[haele]

2. Lol. Most economists have been pointing that out for almost a decade.

Especially since it was well understood that most people don't like dealing with a computer when it comes to anything more complex than an ATM or cash register; they almost inevitably ask to talk to a real person if there's any issue they are concerned about.
And people hate, hate Hate telemarketing and phone menus.
Even though certain personal electronic assistants like the Echo are considered "okay" - it's only because you control it and tell it what to do, it doesn't act in lieu of a gatekeeper or soliciter for some outside organization.
So, why would anyone think that AI could be accepted by most of the public in anything other than a personal toy or tool like Alexa or Photoshop? Oh, CEOs and Beancounter types love the hype, the pretense that AI could replace a workforce, especially a creative or research based workforce, but it just isn't "there".
It's still a LLM, with programmed responses based on ones and zeros, and a search filter/auto-complete system.

It ain't the Star Trek Computer, or Azimov's Robots in Foundation. It might eventually become something like that, but that's not going to happen until we can figure out how brains and consciousness works. Not just how brains react when they're healthy or injured. We need to figure out how to create a bio-computing based system.
And we aren't even that close to figuring worm brains out.

[hunter]

3. May it all come crashing down on Trump's watch.

The sooner the better.

I'm wary of companies that say they are laying people off in favor of AI. It seems more likely that their futures are not looking bright but they don't want to spook investors.

When the AI industry stops selling their imitation intelligence products at a loss things could get ugly quick.

[usonian]

4. One article I read, in the form of a future tale, said 2027.

https://ksaweryskowron.substack.com/p/ai-without-the-bs-part-4

I personally think that the ability to run models on desktop systems (You can't buy a mac mini or studio) will eat up all the low lying, and maybe stepladder, fruit, like Excel (actually VisiCalc) ate up mainframe Cobol apps in large measure.

Lots of new markets give away freebies trying to corner said new market, until one is left standing and can demand monopoly pricing.,

To me, these games are bets against technology, that it will filter down when more efficient hardware and algorithms show up. And who wins bets against technology smartening up?

[tinrobot]

5. All so they can flood our feeds with pointless AI slop.

[usonian]

6. Do the tech exec's care how their product/process is used?

Probably they do, to some extent, since it's all interactive (and deterministic? I just don't know)

And they have real control over how it works.

It might be the least neutral technology that I've seen.

Of course, slop is what they feed on, and slop is largely what they kick out.

Ouroboros? No doubt some critical uses have to migrate off the giant funnel, and the internet has already been deliberately tainted with propaganda and other threats.

-------
I am just now reviewing this goodie, unless you want to beat me to it. OUCH

Steve Blank Anthropic Mythos – We’ve Opened Pandora’s Box
Frightening stuff, and for really real.
https://steveblank.com/2026/04/28/anthropic-mythos-weve-opened-pandoras-box/