this post was submitted on 17 Jan 2026
90 points (100.0% liked)

technology

24173 readers
405 users here now

On the road to fully automated luxury gay space communism.

Spreading Linux propaganda since 2020

Rules:

founded 5 years ago
MODERATORS
 
you are viewing a single comment's thread
view the rest of the comments
[–] Wheaties@hexbear.net 21 points 1 day ago (1 children)

It is clear to us that a lot of people want to use a lot of AI and don't want to pay, so we are hopeful a business model like this can work

hm, do you think they'll do the smart thing and have the chatbots reprint advertising copy and premade images? Or are they gonna burn even more tokens generating ✨unique ads nobody wants or reads?

[–] plinky@hexbear.net 21 points 1 day ago* (last edited 1 day ago) (1 children)

the smart thing is actually organically generate the tokens for ads, and influence consumer behavior, because that's what would print them money. i once again remind that people are getting divorced and going on "planned" trips with that damned thing, of course naturally suggestible users will go for a bit of fresh snack™ when their trusted friend produces tokens that they need a crunch of brandname™

(not to mention supplements and pills)

[–] Wheaties@hexbear.net 12 points 1 day ago (2 children)

Would it be effective advertising? Yes. Would it be a profitable business model for OpenAI? Ehhhh... kinda doubt it. Advertising (and tracking) may subsidize practically all search engines, but I don't think there's anywhere near enough marketing departments with the kind of spending budget Sam Altman needs to dig himself out of the hole.

[–] hector@lemmy.today 4 points 1 day ago

Yeah from what I hear each query costs several pennies to run responses, magnitudes more than a search on an engine.

The sad truth is when it does crash western governments will bail them out, in exchange for deals, private kickbacks and preferential treatment/ pervertion of results.

Central banks like the fed will bail all big players out too, this so called monetary easing, buying bad corporate bonds, is just disguised bailouts.

[–] plinky@hexbear.net 6 points 1 day ago (2 children)

It’s enough to finance google and meta, no? 5% of that will make him a new zuck

[–] fox@hexbear.net 17 points 1 day ago (2 children)

The difference is that what they're doing is cheap on a per-user-operation level. If Google makes $0.00001 per search, but they spend a tenth of that serving it, they come out ahead. OpenAI and LLMs generally are so massively unprofitable that they cannot recover the cost of a query through ad spend at any price an advertiser would be willing to pay.

[–] thethirdgracchi@hexbear.net 5 points 1 day ago (1 children)

The expensive part of LLMs is the training though. Actual token output is rather efficent and quite cheap. For example, for Deepseek to generate a 200 token paragraph of text it costs about $0.000084. Image generation is also rather expensive, but most of the data centers and cost are around training models, not serving LLM output. It still might be more expensive than advertisers are willing to pay, but not crazy expensive.

[–] fox@hexbear.net 6 points 1 day ago* (last edited 1 day ago) (1 children)

You'd think, but efficiency gains are erased by the LLMs having bigger context windows and self-referencing "thinking" or "agent" modes that massively extend token burn. There's public data out there showing how training costs are an enormous fixed point, but then inference costs very quickly catch up and exceed the training cost.

A model that's token-efficient is a model that's pretty useless and a model that's useable for anything is so inefficient as to have massively negative profit margins. If there was even one model out there that was cost effective for the number of tokens burned, the provider would never shut up about it to buyers

[–] thethirdgracchi@hexbear.net 1 points 1 day ago

Wow, really? I guess context windows have been going up but did not realise they were so ruinously expensive. Where can I read more about this?

[–] plinky@hexbear.net 3 points 1 day ago

Depends if they are more effective, innit. You are reaching demographic which likely not seeing ads, aspiring rich tech bros and bored professionals.

[–] Wheaties@hexbear.net 5 points 1 day ago (2 children)

Facebook was never burning billions of venture capital just as operating expenses.

[–] plinky@hexbear.net 7 points 1 day ago

uber was, but in any case, ad market has what, 1 trillion of surplus value stolen to grift? It’s plausible to find 20 billion there. Again, data center build out is nutso, those they cannot fund, but inference for college student with first credit card? Maybe

[–] JoeByeThen@hexbear.net 5 points 1 day ago

No, that was Uber. Lol.