AI News – Nov 27, 2023
Summary of Recent News
-
OpenAI Launches Agents
OpenAI recently announced the launch of their own agent framework called “GPTs”. GPTs are essentially AI agents, but the naming is confusing as GPT is already a term in the AI world. Along with GPTs, OpenAI also introduced a GPT “app store” where creators can list their GPTs and earn money. However, currently, off-the-shelf GPTs are not in high demand as they are easy to replicate. As GPTs and agents become more sophisticated, it is expected that the demand for pre-built GPTs will increase.
-
Understanding AI Agents
An AI agent is defined as a modular artificial intelligence bot programmed to accomplish specific tasks. They are more powerful when paired with other AI agents as they can work together and check each other’s work. OpenAI’s AutoGen is currently the most advanced AI agent framework, offering features like caching, multi-agent definitions, and enhanced inference. The future of AI agents involves the development of models specifically fine-tuned for agent frameworks, more tools for agents to use, and the need for authentication to allow agents to access websites and apps that require login credentials.
-
Reinstatement of Sam Altman as OpenAI CEO
After a week of confusion and uncertainty, Sam Altman has been reinstated as the CEO of OpenAI. This comes after a series of events including a joint letter from OpenAI employees threatening to quit, the appointment of Emmett Shear as interim CEO, and negotiations involving Brian Chesky. The ongoing saga highlights a clash between OpenAI’s commercial arm, focused on raising capital and bringing AI products to the masses, and its non-profit arm, focusing on responsible AI research. The board’s handling of communications during this period has been criticized. The implications of Altman’s return and the board reshuffling are yet to be seen.
-
Disbanding of Meta’s Responsible AI Division
Meta, formerly Facebook, has dissolved its Responsible AI division as part of cost-cutting measures. The move reflects the company’s focus on innovation and catching up with AI competitors. This decision follows a similar move by Microsoft earlier in the year, highlighting the tension between AI acceleration and AI-safety concerns. The implications of disbanding the Responsible AI division for Meta’s AI projects and ethical considerations are unclear.
