Meta's AI research and development

[wjfox]

We have threads for Google and OpenAI, but another major player is Meta. So let's have a dedicated thread for this, the parent company of Facebook.

Meta has been quietly progressing in AI research. It was notably excluded from a recent high-profile White House summit on AI. But in fact Zuckerberg has been increasingly vocal about Meta's AI plans (he mentioned "AI" at least 27 times during the last earnings call). He has detailed plans on how to integrate the technology more deeply into Facebook, WhatsApp, Instagram, and other apps. AI has already been used for years to recommend posts in users' feeds, moderate content, target ads, etc. and he now wants to expand its abilities to generative AI, as well as chatbots.

There's been talk of a "semiautomated social network" where AI can create its own posts, or even entire accounts​. However, this will obviously raise concerns around the spread of misinformation, bias, polarising content, and so on.

Meta has been slow to secure the kind of chips needed to scale AI projects, putting it at a competitive disadvantage. But the company has made some exciting announcements in recent weeks that could make it a rival to Google and OpenAI, perhaps sooner than we think. So in this thread, we can follow all the latest news and developments.

[wjfox]

[wjfox]

Meta says its new speech-generating AI model is too dangerous for public release

Jun 17

Meta announced a new AI model called Voicebox yesterday, one it says is the most versatile yet for speech generation, but it’s not releasing it yet:

There are many exciting use cases for generative speech models, but because of the potential risks of misuse, we are not making the Voicebox model or code publicly available at this time.

The model is still only a research project, but Meta says can generate speech in six languages from samples as short as two seconds and could be used for “natural, authentic” translation in the future, among other things.

https://www.theverge.com/2023/6/17/2376 ... ic-release

[spryfusion]

[wjfox]

More details here:

https://ai.meta.com/llama/

[spryfusion]

[spryfusion]

[spryfusion]

Meta Is Developing a New, More Powerful AI System as Technology Race Escalates
Source: WSJ

Meta Platforms is setting its sights on OpenAI.

The parent of Facebook and Instagram is working on a new artificial-intelligence system intended to be as powerful as the most advanced model offered by OpenAI, the Microsoft-backed startup that created ChatGPT, according to people familiar with the matter. Meta aims for its new AI model, which it hopes to be ready next year, to be several times more powerful than the one it released just two months ago, dubbed Llama 2.

The planned system, details of which could still change, would help other companies to build services that produce sophisticated text, analysis and other output. It is the work of a group formed early this year by Meta Chief Executive Mark Zuckerberg to accelerate development of so-called generative AI tools that can produce humanlike expressions. Meta expects to start training the new AI system, known as a large language model, in early 2024, some of the people said.

Plans for the new model, which haven’t previously been reported, are part of Zuckerberg’s effort to assert Meta as a major force in the AI world after it fell behind rivals. Competition in the area has sharply intensified this year, spawning divergent views on everything from which business models are best to how the technology should be regulated.

The company is currently building up the data centers necessary for the job and acquiring more H100s, the most advanced of the Nvidia
chips used for such AI training. While Meta joined with Microsoft to make Llama 2 available on Microsoft’s cloud-computing platform Azure, it plans to train the new model on its own infrastructure, some of the people said.

Zuckerberg is pushing for the new model, like Meta’s earlier AI offerings, to be open-sourced and therefore available free for companies to build AI-powered tools.

Zuckerberg will be among a group of top tech executives attending a summit organized by Senate Majority Leader Chuck Schumer (D., N.Y.) on Wednesday to discuss how to handle AI. Sam Altman, OpenAI’s CEO, and Sundar Pichai, Google’s CEO, will also be attending.

The model under development may not close the gap with Meta’s competitors.

Meta hopes it will be roughly as capable as GPT-4, which OpenAI launched in March. GPT-4 underpins OpenAI’s moneymaking initiatives such as the recently launched ChatGPT for Business tool, and the company has been courting others to build on top of the technology as it tries to cover the enormous costs for advanced AI models. Meta’s new model also likely would come out after the expected debut of Gemini, an advanced large language model being built by Google.

(Archive)

[wjfox]

Meta announces AI chatbots with 'personality'

21 minutes ago

Meta has announced a series of new chatbots to be used in its Messenger service.

The chatbots will have "personality" and specialise in certain subjects, like holidays or cooking advice.

It is the latest salvo in a chatbot arms race between tech companies desperate to produce more accurate and personalised artificial intelligence.

The chatbots are still a work in progress with "limitations", said boss Mark Zuckerberg.

In California, during Meta's first in-person event since before the pandemic, Mr Zuckerberg said that it had been an "amazing year for AI".

https://www.bbc.co.uk/news/technology-66941337

[wjfox]

Meta’s new AI image generator was trained on 1.1 billion Instagram and Facebook photos

Benj Edwards - 12/6/2023, 9:52 PM

On Wednesday, Meta released a free standalone AI image-generator website, "Imagine with Meta AI," based on its Emu image-synthesis model. Meta used 1.1 billion publicly visible Facebook and Instagram images to train the AI model, which can render a novel image from a written prompt. Previously, Meta's version of this technology—using the same data—was only available in messaging and social networking apps such as Instagram.

https://arstechnica.com/information-tec ... ok-photos/

[wjfox]

[spryfusion]

Blog post: https://ai.meta.com/blog/v-jepa-yann-le ... hitecture/

[weatheriscool]

[weatheriscool]

[wjfox]

[weatheriscool]

[Powers]

https://nitter.poast.org/rowancheung/st ... 2846641335

[spryfusion]

Blog post: https://ai.meta.com/blog/meta-llama-3-1/

[spryfusion]

[wjfox]

Sigh.

Really wish we could go back to late 2000s-era Facebook.

-----

Mark Zuckerberg says a lot more AI generated content is coming to fill up your Facebook and Instagram feeds

October 30, 2024 at 11:39 PM GMT

First we had friends. Then we had influencers. And if Mark Zuckerberg is correct, the next big thing in our social media feeds will be AI generated content. Lots of it.

Zuckerberg described our future feeds during Facebook-parent company Meta’s third quarter earnings conference call on Wednesday, describing it as a natural evolution.

“I think were going to add a whole new category of content which is AI generated or AI summarized content, or existing content pulled together by AI in some way,” the Meta CEO said. “And I think that that’s gonna be very exciting for Facebook and Instagram and maybe Threads, or other kinds of feed experiences over time.”

Zuckerberg touted the company’s Llama large language model and the success of products it powers, such as the Meta AI chatbot that is now used by more than 500 million users every month. But Llama will increasingly play a role across Meta’s business, Zuckerberg said, including tools for business customers and advertisers.

As AI tools become more widespread, AI content will proliferate within social media feeds. Such feeds are actively being worked on inside Meta, Zuckerberg noted. “It’s something we’re starting to test different things around.”

https://fortune.com/2024/10/30/mark-zuc ... dia-feeds/