Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The CEO quiet part out loud was very clearly: "salaries".

This will scale back when AI replacement attempts slow down as expectations temper (Salesforce, Klarna, etc).





The weird thing is that the AI companies themselves are hiring like there's no tomorrow, doing talent aquisitions etc. Why would you do that if the purpose of your product is to reduce necessary workforce?

Why isn't that the first question that comes to mind for a journalist covering the latest acquisition? It's like an open secret that nobody really talks about.


To answer your questions (I don't think it's what you wanted, but people will scratch their heads after reading them):

On reality, they are hiring because they have a lot of (investment) money. They need a lot of hardware, but they also need people to manage the hardware.

On an alternative reality where their products do what they claim, they would also hire, because people working there would be able to replace lots of people working in other jobs, and so their workers would be way more valuable than the average one, and everybody would want to buy what they create.

Journalists don't care about it because whatever they choose to believe or being paid to "believe", it's the natural way things happen.


Just to clarify: Most AI companies don't own their hardware, with a select few exceptions. That's why a handful of hyperscaler stock has rallied recently on letters of intent on large orders from AI companies. Which technically is a handful of shell companies under complete control of their parent companies, which can then take on credit without it being visible on the parent company balance sheet.

But addressing the specific question: It is still a valid. If the product sold is a 10x developer force multiplier, you'd expect to see the company fully utilizing it. Productivity would be expected to increase, rapidly, as the product matures, and independently of any acquisitions made at the same time.


Everyone is trying to be the shovel sales person for AI, not the gold diggers buying shovels.

I'm not sure if even the LLM companies themselves are selling shovels yet. I think everyone is racing to find what the shovel of LLMs are.


It was collectively decided some time ago that this particular shovel is called nVIDIA.

And memory. What I'm surprised by is that memory production isn't being scaled up since it's basically a universal part of any computing device.

That was decided when crypto mining became too expensive, I guess.

It's always about cutting the OpEx spend. Companies are nothing more than giant piles of money seeking to grow themselves in any way possible.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: