Hacker News new | ask | show | jobs
by kamranjon 3 hours ago
Does anyone understand how Meta is able to spend so much money on AI with basically no AI product to speak of? Especially after sinking billions of dollars into a failed VR product? I just don't really understand why they are investing in data centers, I don't know of any actual product they offer that anyone is seriously considering using in the space.
11 comments

> with basically no AI product to speak of?

I don't use Meta products regularly, but they've crowbarred the "Meta AI" feature into everything - FB Messenger, Whatsapp, Instagram... There's even a standalone website to chat with the Meta AI model similar to chatgpt:

> https://www.meta.ai/

I'm certainly prepared to agree that they don't have a huge amount of traction or marketshare, but they absolutely have already built AI products and they exist.

Outside of their own consumer products, there's the family of llama models too etc.

There are many uses for AI other than selling API/chat access. For Meta it can be for example use internally as a software tool, in the same way that they have their own datacenters instead of running on AWS. They can also use them to power recommendation algorithms to increase time on platform. Or they can use them to better target adverts and thus increase the revenue from ads. They can also use them to help people make ads on their platforms etc....
The market forgives misadventures cause Meta is still solvent and they make money YoY. Additionally, they are developing heavily in the AI space with making Llama available to the public and all the AI integrations into their products.
Because Zuckerberg is the king and has complete control, but Meta is so far behind in this so-called AI model race it will be canceled quietly. It will just be but a footnote in about 2 to 3 years.
They're not using their money, they're borrowing it and using Special Purpose Vehicles to book the debt so it doesn't hit their own balance sheet.
Meta uses its ai products internally. I can't actually go read the reporting for you where they've talked about, but they talk about it regularly.
They are using the same infinite money glitch as Google - ads revenue.
Internal use to watch everything and control everything
They will sell the capacity to others. And building data centers let's them leverage local tax advantages/incentives.
Imagine buying ad space on their platforms, but instead of writing copy and providing images, you simply give it your website.

Then they generate unique copy and images for each user (or hyper-targeted bucket of users), tailored to what would make them click. All continuously A/B tested.

I'm imagining it, and it'd at best be the same as if you gave your website to an ad sales agent without any instructions as to what your actual product or target audience is. At worst it'd be writing copy that is fully fabricated, doesn't match your brand language, and opens you up for false advertising claims.
Who cares about them, at least maybe there might be another viable nuclear reactor after they throw a lot of money at it. They lost money on dumb stuff: the metaverse, VR. By all means, throw money at a few nuclear reactors, at least those are useful to keep the lights on.

They probably want to better target ads and generate outrage slop with AI. That was always the plan: spy on people, make them unhappy, sell ads and maybe also provide the means to scoundrels to scam the elderly in the process.