On Tuesday, Meta hosted its inaugural AI developer conference, LlamaCon, at its headquarters in Menlo Park, California. During the event, the company unveiled a consumer-oriented Meta AI chatbot app, which will rival ChatGPT, and a developer-focused API for accessing Llama models in the cloud.
These releases aim to broaden the adoption of Meta’s open Llama AI models. However, a key underlying motive appears to be countering OpenAI. Meta’s broader AI strategy involves fostering an open AI ecosystem to challenge “closed” AI providers such as OpenAI, which restricts its models behind paid services.
The Meta AI chatbot app includes a social feed where users can share their AI interactions and provides personalized responses based on a user’s Meta app activity. Meanwhile, the Llama API simplifies the process for developers to build applications that connect to Llama models in the cloud using just one line of code. This eliminates reliance on third-party cloud providers and enhances Meta’s toolkit for AI developers.
Court filings reveal that Meta executives have been focused on surpassing OpenAI’s GPT-4, a once state-of-the-art model. Undercutting proprietary AI model providers like OpenAI has long been central to Meta’s AI strategy. In a July 2024 letter, Meta CEO Mark Zuckerberg contrasted Meta with companies like OpenAI, stating that “selling access to AI models isn’t [Meta’s] business model.”
During LlamaCon, Zuckerberg emphasized collaboration with AI labs that openly share their models, such as DeepSeek and Alibaba’s Qwen, positioning them as allies against closed model providers. He highlighted the value of open source in allowing developers to combine the best features from various models.
Meta may also be pushing its open models to align with regulatory frameworks. The EU AI Act offers privileges to companies distributing “free and open source” AI systems, a label Meta frequently applies to its Llama models, though this classification remains debated.
Ultimately, Meta seems focused on strengthening the open model ecosystem and limiting OpenAI’s growth, even if it means not always delivering the most advanced models itself.
— new from TechCrunch