However, Meta is also facing criticism from within its own ranks. Yann LeCun, Meta’s chief AI scientist, has expressed skepticism about generative AI and believes that the future of AI lies in joint embedding predicting architecture (JEPA), a different approach to training models and producing results. Despite this, Meta is continuing with its plans to release Llama 3 and is also working on Emu, an image generation tool, although no release date has been given for this yet.
Key takeaways:
- Meta has confirmed that it plans to release Llama 3, the next generation of its large language model used to power generative AI assistants, within the next month. The new model will power multiple products across Meta and will have different capabilities and versatilities.
- Llama 3 is expected to be bigger in scope than its predecessors, with capabilities to answer questions more accurately and to field a wider range of questions, including more controversial topics.
- Meta is not yet releasing Emu, its image generation tool, and is taking a cautious approach with AI, especially when it comes to other generative AI beyond text generation.
- Yann LeCun, Meta’s chief AI scientist, expressed skepticism about generative AI and predicted that the future of AI is joint embedding predicting architecture (JEPA), a different approach to training models and producing results.