Meta Is Delaying the Rollout of Its Flagship AI Model
Source: WSJ
Meta is delaying the rollout of a flagship AI model, prompting internal concerns about the direction of its multibillion-dollar AI investments, people familiar with the matter said.
Company engineers are struggling to significantly improve the capabilities of its Behemoth large-language model, leading to staff questions about whether improvements over prior versions are significant enough to justify public release, the people said.
Early in its development, Behemoth was internally slated for an April release to coincide with Metas inaugural AI conference for developers. Meta put out two smaller models in its Llama AI model family ahead of the event, but later pushed an internal target for the larger Behemoths release to June. Now, its been delayed to fall or later.
-snip-
Metas recent challenges mirror stumbles or delays at other top AI companies that are trying to release their next big state-of-the-art models. Some researchers see the pattern as evidence that future advances in AI models could come at a far slower pace than in the past, and at tremendous cost.
-snip-
Read more: https://www.wsj.com/tech/ai/meta-is-delaying-the-rollout-of-its-flagship-ai-model-f4b105f7
Another sign that AI companies are getting diminishing returns on investment. Most of Meta's $72 billion in capital expenditures this year will be on AI.
Generative AI IS a bubble, which an unbelievable amount of money has been wasted on. February thread:
Ed Zitron: There Is No AI Revolution (DAMNING analysis of a house of cards built by con men & venture capitalists)
https://www.democraticunderground.com/100220066913

Crowman2009
(3,083 posts)It reminds me back in the 80's when the main selling point for home computers was printing out recipes. Which turned out to be absolutely useless.
highplainsdem
(56,054 posts)politics and technology in 1985.
I don't recall anyone ever suggesting printing out recipes as a selling point for home computers, though.
Generative AI, which is what this article is about, is badly flawed tech with new models hallucinating more and not less than older models, and with stolen intellectual property as training data. It's expensive to operate, damages the environment, and dumbs down and deskills users. Its primary use is fraud of various types, whether students cheating or people pretending to have abilities they don't have.
It might be the dumbest as well as the most harmful non-weapon tech ever.
And it's probably the most incredibly hyped.
Skittles
(164,345 posts)
highplainsdem
(56,054 posts)highplainsdem
(56,054 posts)recipes - and neither can Skittles - some other people can.
Found a discussion of that here:
https://boards.straightdope.com/t/why-did-so-many-early-computer-ads-mention-usage-for-storing-recipes/778569
Using a home computer for recipes doesn't seem to have been a main selling point, though. Apparently it was a sexist addition to some ads that some clueless marketer thought would be necessary to appeal to women. I had nothing against cooking, but as a woman with a shelf of cookbooks, nothing would've appealed to me less than typing up recipes very carefully for computer files. And I think I would have had a pretty angry feminist reaction to a sexist ad for PCs. So I'm still sure I never saw any of those ads.
SheltieLover
(68,266 posts)
jmowreader
(52,297 posts)Aren't all these Large Language Model AI systems really just search engines for people who don't know how to set up a search query, and Generative AI just automatic plagiarism?