In a notable industry shift, Character.AI co-founder and CEO Noam Shazeer is returning to Google after a departure in October 2021 to establish the a16z-backed startup. During his previous tenure at Google, Shazeer led the team that developed LaMDA (Language Model for Dialogue Applications), a key component in conversational AI.

Joining Shazeer in the transition back to Google are Character.AI co-founder Daniel De Freitas and several other employees. Dominic Perella, Character.AI’s General Counsel, will step in as interim CEO. Despite these changes, the majority of Character.AI’s team will remain with the company.

As part of this move, Google is making an undisclosed investment in Character.AI. In exchange, Character.AI will provide Google with a non-exclusive license to its technology.

“I am thrilled to return to Google and join the Google DeepMind team,” Shazeer shared in a statement to TechCrunch. “I am incredibly proud of what we have accomplished at Character.AI over the past three years. The funds from our non-exclusive licensing agreement with Google, combined with the outstanding Character.AI team, will ensure our continued success.”

Google expressed enthusiasm in their statement: “We’re excited to welcome Noam back, a leading figure in machine learning, to Google DeepMind’s research team, along with a few of his colleagues. This agreement will enhance Character.AI’s growth and support its mission to develop personalized AI products for users globally.”

Character.AI has garnered over $150 million in funding, primarily from a16z.

In their blog post announcing the move, Character.AI stated: “When Noam and Daniel founded Character.AI, our vision for personalized superintelligence required a comprehensive approach. We had to pre-train models, refine them to power unique experiences, and build a platform capable of reaching users worldwide. Over the past two years, the landscape has evolved with the availability of more pre-trained models. This shift allows us to leverage third-party LLMs alongside our own, dedicating more resources to post-training and new product development for our expanding user base.”