Noam Shazeer, the Google DeepMind researcher whose work on transformers helped ignite the generative-AI boom, is joining OpenAI, according to The Information. Shazeer was a key author of the original transformer paper, the architecture behind OpenAI’s GPT models, and at OpenAI he will work on new model architectures and on evolving the transformer itself. His exit pulls a pretraining lead off Google’s Gemini team, two years after Google brought him back through a $2.7B deal with his startup Character.AI. The hire is a notable talent-war win for OpenAI as it races to catch Anthropic ahead of both companies’ anticipated IPOs. Shazeer’s career has drawn controversy, including lawsuits Character.AI settled this year over allegations its chatbots contributed to teen suicides, and inflammatory posts he made while at Google.
