Meta sees no commercial problems in sharing its AI technology. This was stated by Meta’s AI scientist, Yann LeCun, at a meeting recently.
Meta has been busy in recent months releasing open-source versions of its LLMs, including Llama 2. According to Meta AI scientist Yann LeCun, the tech giant does not see these open-source versions as a threat to commercials, Bloomberg writes. Instead, the tech giant wants to earn from the products and services built on top of the AI models.
Afraid of copies
Meta hereby deviates from the vision held by other providers of AI models, such as OpenAI and Google Deepmind. These providers do not have open-source versions of their AI models and make users pay for them.
The main reason for the lack of open-source versions is that these AI developers are mostly afraid that competitors can steal their technology.
Meta sees it differently
For Meta, this is not a problem at all, LeCun continues. In fact, the tech giant sees a great advantage in offering open-source versions of its AI models. This ensures that a broader community develops the models together. LeCun points out that technology can improve faster than if Meta had had to do it itself because of this. “When you do this development in secret, you fall behind,” the Meta AI scientist tells Bloomberg.
Payment will follow
Not all AI technology is free at Meta. According to CEO Marc Zuckerberg the tech giant does plan to charge cloud providers to use its AI models in the longer term.
Also read: Meta’s LLaMA 2 AI models available through Amazon SageMaker JumpStart