Meta’s LlamaCon was all about undercutting OpenAI

On Tuesday, the target held its first conference of AI developers, Llamacon, at its headquarters of Menlo Park in California. The company announced Starting the Meta AI Chatbot application facing consumerswho will compete with Chatgpt, as ia API turned to the developer to access the Lem models in the cloud.

Both editions aim to expand the adoption of the open -ended Llam AI companies, but this goal could be a secondary true Meta motif: Being Openi. The target of AI ambition, in wide moves, stimulates a successful open Ai ecosystem that holds it on “closed” services of AI services such as Openi, who enter their models behind services.

Meta -ova ai chatbot app feels almost as an assumption Openi is a rumor social network. It has a social feed on which users can share their AI chats and offers personalized answers based on the user’s Meta application activity.

As for the API llam, this is a challenge for the Openi -API job. The API API is designed to make developers easier to build applications that connect to the Llam models in the cloud, using only one line of code. It eliminates the need to rely on third party service providers to launch the LLE Model and allows the target to offer a more complete range of tools for AI developers.

The target, like many AI companies, is experiencing an openii as a top rival. Court reports in the case of a meta reveal that the company executors previously obsessed He beat up the OPENAI GPT-4, which was once a top model. Understanding owners of AI models like Openi have long been fundamental to the target of AI strategy. In a July 2024Meta Meta Mark Zuckerberg CEO has sought to compare the target with companies like OpenAI, writing that “the sale of access to AI models is not [Meta’s] Business model. “

Several AI -Techcrunh’s researchers in front of Llacacon have hoped the target Release a competitive model of AI Resonation like OpenAI’s O3-Dom. The company didn’t finish it. But for the target, it’s not about winning in the AI ​​race.

During the interview on stage with Databricks executive, Ali Ghhody during Llalacon, Zuckerberg said he sees any AI laboratory that allows its models to openly accessible, including Deepseek and Alibaba Qwen, as allies in the fight against closed models.

“Part of the open code value is that you can mix and match. So, if the second model, like Deepseek, is better – or if Qwen is better – then, as developers, you have the ability to take the best parts of intelligence from different models and produce exactly what you need,” said Zuckerberg. “This is part of the way I think an open source basically passes in the quality of all closed sources [models] … [I]T feels like a kind of unstoppable force. “

Apart from the arrest of openii growth, the target can also try to push its open models In order to satisfy regulatory carving. The EU AI Law gives special privileges to companies that distribute AI systems of “free and open code”. Meta often claims that his models of llam “are open source”, Despite disagreement in whether they meet the necessary criteria.

Regardless of the reason, the target seems to be pleased with the launch of the launch that strengthens the ecosystem of the open model and limit the growth of the openii-a-occasion to the detriment of top-notch models.