• Moderators, please send me a PM if you are unable to access mod permissions. Thanks, Habsy.

Tech Thread

ruh roh

HXslSxiX0AA3Lhg
 
Galloway predicts that Meta wins the AI race because they have the stickiest product on earth and massive revenues coming in while they spend insane sums to build things out.
 
But, what does a "win" even look like? What, they're going to spend 10's-100's of billions developing an LLM so that they can sell the same ads they were selling before and earn the same amount of ad revenue prior to AI?

We already know who the winner is, it's open weight models and the GPU manufacturers.
 
I mean, a win is revenue growth and expense reduction, which AI is supposed to be delivering. So, even better targeted content and ads, costing less to deliver once the buildout essentially concludes and ramps down.
 
I mean, a win is revenue growth and expense reduction, which AI is supposed to be delivering. So, even better targeted content and ads, costing less to deliver once the buildout essentially concludes and ramps down.

Capex is a thing though. Meta spent 145 Billion on AI in 2026 alone and they're seeing increases of single digit billions in net operating revenue this year, most of which is entirely independent of their AI usage. When all is said and done they're going to end up spending maybe half of a fucking trillion to juice revenue a few billion a year?

That's the winner?
 
Holy fuck, I wrote out a reply but when I posted it, it disappeared thanks to the malfunctioning site.

In short, I just said that I'm no expert on the subject, but we're in the very early stages of the AI buildout, and whatever the early gains might be, I'd expect the later gains, once we're further down the road, to be substantially more. I don't think they're so dumb to put so much capital behind it if it wouldn't translate to huge gains.
 
I don't think they're so dumb to put so much capital behind it if it wouldn't translate to huge gains.

iu

In short, I just said that I'm no expert on the subject, but we're in the very early stages of the AI buildout, and whatever the early gains might be, I'd expect the later gains, once we're further down the road, to be substantially more.

But they're already cancelling project and pulling back on the expenditures. There's minimal revenue from selling AI (as in, inference costs are significantly higher than can be charged for said inference).

We're not in the early stages of the AI build out at all, well over a trillion has already been spent with revenue in the very low 10's of billions to show for it.
 
Back
Top