In other AI news, Microsoft has signed a deal with HarperCollins to train an upcoming model on the publisher's non-fiction books, Google has rolled out a memory feature for paying Gemini users, and AI pioneer Yoshua Bengio has called for urgent AI regulations, citing OpenAI’s o1 model. The Stanford Institute for Human-Centered AI has also released a new index ranking countries on the vibrancy of their AI ecosystems, with the U.S. taking the top spot, followed by China and the U.K.
Key takeaways:
- Rabbit, the maker of the AI-in-box device R1, has launched a new capability called Teach mode. This allows users to instruct their devices on how to perform specific actions using natural language, with the aim of creating an app store for model actions.
- Microsoft has signed a deal with HarperCollins to train an upcoming model on the publisher's non-fiction books. The details of the model are not yet known, but authors will need to opt into the training program.
- Google has rolled out a memory feature for paying Gemini users, which allows the chatbot to remember facts about users' preferences and interests for more tailored responses. However, there are concerns about the potential for this feature to be exploited by hackers.
- AI pioneer Yoshua Bengio has called for urgent AI regulations, citing OpenAI's upcoming o1 model. He argues that the advancement of AI towards human-level reasoning presents both opportunities and risks, including the potential for deception and the development of biological weapons.