The Social Llama

Meta launches a social AI app and Llama API, expanding its AI strategy with open models and developer tools. But can its momentum match its ambitions? Image credit: Ken Yeung/Adobe Firefly
, , , , ,
"The AI Economy," a newsletter exploring AI's impact on business, work, society and tech.
Welcome to "The AI Economy," a weekly newsletter by Ken Yeung on how AI is influencing business, work, society, and technology. Subscribe now to stay ahead with expert insights and curated updates—delivered straight to your inbox.

IN THIS ISSUE: Meta hosts its first-ever event around its Llama model, launching a standalone app to take on Microsoft’s Copilot and ChatGPT. The company also plans to soon open its LLM up to developers via an API. But can Meta’s momentum match its ambition?

The Prompt

At Meta’s inaugural LlamaCon developer conference, the company introduced two AI advancements that signal its intent to better compete with industry leaders like Microsoft and OpenAI. Central to these announcements was the introduction of a standalone Meta AI app, designed to function both as a general-purpose chatbot and as a replacement for the “Meta View” app used with Meta Ray-Ban glasses. This app distinguishes itself by integrating a social component—a public feed showcasing user interactions with the AI, allowing users to explore and remix each other’s prompts.

Complementing the app launch, Meta previewed its Llama API, designed to simplify the integration of its Llama models into third-party products. This move positions Meta to attract AI developers by offering an open-weight model that supports modular, specialized applications. It follows a similar playbook to the one the company formerly known as Facebook followed years ago, which helped its social network outperform its competitors.

LlamaCon was designed to highlight Meta’s commitment to open-source AI, a space where the company has long positioned itself as a champion. It was apparent during the livestreamed sessions as Meta repeatedly emphasized the vital role the technology plays in its AI strategy, framing it as a key differentiator in its ongoing rivalry with OpenAI and other tech giants.

Open-source AI is also doing well for Meta. The company revealed earlier this week that its newest large language model (LLM), Llama 4, has been downloaded more than 1.2 billion times. And it found this success despite the benchmarking controversy surrounding the model. And since it rolled out Meta AI into its product suite, more than a billion people have used it each month across the company’s apps and on its Ray-Ban glasses.

Meta AI Levels Up

Until this week, the main way to access Meta AI was through integration with other Facebook, Facebook Messenger, Instagram, WhatsApp, or Meta-powered devices—an approach that guaranteed steady engagement by tapping into established user bases. But accessing them felt disconnected and unnatural. Releasing a standalone app gives Meta greater freedom to experiment with how its AI functions, free from the constraints and limitations of its existing platforms. In addition, it permits Meta to “focus on pushing the limits and [offer] a fresh take on how people could use AI,” according to company Chief Product Officer Chris Cox.

As he explains, “We were very focused on the voice experience—the most natural possible interface. So, we focused a lot on low latency, highly expressive voice. The experience is personalized, so you can connect your Facebook and Instagram accounts, and the assistant will have a rough idea of your interests based on your interaction history. It will also remember things you tell it, like your kid’s names, your wife’s birthday, and all the other things you want ot make sure your assistant doesn’t forget.”

In a fashion, Meta has done precisely what Microsoft did in April: Turn its AI into a personal assistant. However, there are some differences between Meta AI and Microsoft’s Copilot. Meta AI offers basic but practical tools, such as problem-solving assistance, contextual understanding, and duplex speech capabilities. In contrast, although Microsoft does offer those capabilities with Copilot, it has also promised more advanced features, including Memory, the ability to read webpages, and interpret the world through a camera, custom podcast generation, and a forthcoming option to personalize your AI avatar.

The question is: Will these features be compelling enough to drive users to adopt one or the other?

One distinct advantage that Meta has over its competitors is its access to social media data. As model providers are starving for more information to feed models, Meta has a seemingly endless supply. It can extract rich behavioral data, trends in real-time language use, emotional and contextual nuance, and social graph insights. It’s a distinction that Meta also holds over OpenAI, which has reportedly been considering building a social network of its own. This could prove advantageous if you’re looking to create a consumer-friendly AI assistant as opposed to one focused mainly on productivity.

That being said, a peculiar feature in the new Meta AI app is the ability to share prompts with the community. Cox states that many people often have no idea what to do with this technology, and “it’s not until they see the way that others are using it, doing stuff like them, that they get inspired.” On one hand, it could be a great way to help show value. On the other hand, it’s creating a potential privacy issue as people may accidentally share personal prompts not meant for public consumption.

Opening Llama to Developers

The second announcement from Meta was that an API is now being tested. With it, developers can seamlessly integrate their applications with the LLM using a single line of code. They would have access to the litany of Llama models, including those from the Llama 4 family. Meta says that it is also providing Python and Typescript SDKs. The company is also offering faster inference speeds through its Llama API, thanks to support from Cerebras and Groq. 

“Our goal from the beginning…has been, it should be the fastest and easiest way to build with Llama. Not only that,…it should be the best way to customize these models. And going one step further, you should be able to take these custom models with you wherever you want, no lock-in ever. Speed, ease of use, customization, and no lock-in are the marquee features of the Llama API,” Manahar Palari, a research scientist at Meta, says. 

The introduction of the Llama API marks a strategic move by Meta to expand the reach and usability of its open-weight Llama models. By offering developers an accessible, cloud-based way to integrate Llama into their products and workflows, Meta is positioning itself as a serious contender in the developer ecosystem. The Llama API enables more modular and customizable use cases, allowing companies to fine-tune the model or build specialized agents without needing to host the model themselves. This flexibility, combined with Meta’s open-source ethos, could attract developers looking for alternatives to more closed and proprietary AI platforms.

Palari discloses that startups are testing the Llama API today. “The early feedback we’re getting is how easy it is to use, the speed, and quality of responses.” When launched, Meta’s Llama API would join similar other feeds that developers can choose from, including ones from OpenAI, Google DeepMind, Mistral, Cohere, and xAI.

All that being said, Meta’s LlamaCon had some interesting reveals. Still, some people contend that the event was lackluster and hoped the company would release more models, such as a reasoning one that would rival OpenAI’s o3-mini. At least one review argues that the conference showed that the company is still playing catch-up. There’s no doubt Meta has the resources to compete with the top players in AI, but its models and developer tools may leave some wondering just how much confidence they should place in its technology.

Still, one of Meta’s key advantages is its long-standing relationship with the developer community. From its early days with Facebook Platform and the F8 developer conferences to more recent investments in open-source AI, Meta has consistently provided tools, documentation, and infrastructure to help developers build on its ecosystem. This developer-first approach, now paired with the open-weight Llama models and the newly launched Llama API, could foster a surge of creative experimentation. By lowering the barriers to entry and encouraging transparency, Meta is setting the stage for a wide range of applications—from personalized AI agents to domain-specific tools—that might not emerge as easily within more closed ecosystems like OpenAI’s or Microsoft’s.

Don’t Miss out on Future Issues of ‘The AI Economy’

The AI Economy is expanding! While you’ve been getting weekly insights on LinkedIn, I’m gearing up to bring you even more—deep dives into AI breakthroughs, more interviews with industry leaders and entrepreneurs, and in-depth looks at the startups shaping the future. To ensure you don’t miss a thing, subscribe now on Substack, where we’ll be rolling out more frequent updates.

Don’t worry; the weekly newsletter will still be published on LinkedIn, but other stories will be available on Substack.

Stay ahead in the AI revolution—sign up today!

Subscribe to The AI Economy

Quote This

“We’re trying to anchor our north star on the product value that people report to us, what they say that they want, and what their revealed preferences are, and using the experiences that we have. Sometimes these benchmarks just don’t quite line up. I think a lot of them are quite easily gameable.”

— Meta CEO Mark Zuckerberg on the Dwarkesh Patel podcast. He was asked to comment on his company’s Llama 4 model performance evaluation on Chatbot Arena. Zuckerberg criticized open-source benchmarking, claiming it’s “often skewed toward a very specific set of use cases, which are often not actually what any normal person does in your product.” He pushed back at Patel’s insinuation that product value wasn’t something readily comparable across different models, saying it might be possible.

This Week’s AI News

🏭 AI Trends and Industry Impact

🤖 AI Models and Technologies

✏️ Generative AI and Content Creation

💰 Funding and Investments

☁️ Enterprise AI Solutions

⚙️ Hardware, Robotics, and Autonomous Systems

🔬 Science and Breakthroughs

💼 Business, Marketing, Media, and Consumer Applications

🛒 Retail and Commerce

⚖️ Legal, Regulatory, and Ethical Issues

💥 Disruption, Misinformation, and Risks

🔎 Opinions, Analysis, and Editorials

End Output

Thanks for reading. Be sure to subscribe so you don’t miss any future issues of this newsletter.

Did you miss any AI articles this week? Fret not; I’m curating the big stories in my Flipboard Magazine, “The AI Economy.”

Follow my Flipboard Magazine for all the latest AI news I curate for "The AI Economy" newsletter.
Follow my Flipboard Magazine for all the latest AI news I curate for “The AI Economy” newsletter.

Connect with me on LinkedIn and check out my blog to read more insights and thoughts on business and technology. 

Do you have a story you think would be a great fit for “The AI Economy”? Awesome! Shoot me a message – I’m all ears!

Until next time, stay curious!

Subscribe to “The AI Economy”

Exploring AI’s impact on business, work, society, and technology.

Subscribe

Leave a Reply

Discover more from Ken Yeung

Subscribe now to keep reading and get access to the full archive.

Continue reading