Key Takeaways
- Meta held its first AI developer conference, LlamaCon, but didn’t announce anticipated new models like its reasoning or “Behemoth” models.
- The company rebranded its Meta View smart glasses app into a standalone “Meta AI” mobile app.
- The new Meta AI app features voice chat and a social feed showcasing user interactions with the AI.
- Meta confirmed its open-source Llama AI models have been downloaded 1.2 billion times.
- A new Llama API is being previewed for developers, offering more privacy for enterprise use.
- Despite these updates, the event suggested Meta is still trying to catch up with competitors in the generative AI space.
Meta’s recent LlamaCon event, its first conference for AI developers, brought updates but didn’t deliver the major new AI models some were anticipating.
Instead of unveiling a rumored reasoning model or the large “Behemoth” model, Meta focused on consolidating its current position in the competitive AI landscape.
A key announcement was the transformation of the Meta View app, previously for smart glasses, into a dedicated Meta AI mobile app. This new app allows users to chat with Meta’s AI, similar to other chatbot apps.
An interesting feature of the Meta AI app is a “social discover feed.” Unlike typical social feeds, this one displays posts from random users showing how they’re using Meta AI, including prompts and generated images.
Meta’s Chief Product Officer, Chris Cox, highlighted the popularity of their open-source approach, revealing that Llama models have been downloaded 1.2 billion times, according to reporting by CNET.
While Meta integrates its AI into Facebook, Instagram, and WhatsApp, the LlamaCon announcements felt more like catching up than leaping ahead. Competitors like OpenAI already have established mobile apps and advanced reasoning models.
The company did reiterate details about its current Llama 4 models, Scout (smaller, efficient) and Maverick (more powerful), though there was previous discussion around benchmarking comparisons with competitors.
For developers, Meta introduced a preview of its Llama API. This platform aims to make building applications with Llama models easier, faster, and more customizable.
Crucially, Meta stated that data processed through the API (user inputs and AI outputs) will not be used for training Meta’s models, addressing privacy concerns for businesses.
Overall, LlamaCon showcased Meta’s ongoing commitment to AI and its open-source strategy. However, the lack of groundbreaking model releases left the impression that the tech giant is still working hard to match the pace set by others in the rapidly evolving AI field.
The new Meta AI app and its unique social feed are developments to watch, potentially offering a new way for users to interact with and discover AI capabilities.
