Google co-founder urges AI staff to go all in on Gemini model

Google co-founder Sergey Brin is urging employees at the company's AI division to intensify efforts on its Gemini model, signaling a renewed push to close the gap with leading competitors such as Anthropic and OpenAI. This directive comes as Alphabet, Google's parent company, seeks to reassert its dominance in the rapidly evolving artificial intelligence landscape.

Recent months have seen Gemini face stiff competition. While Google's model briefly surpassed rivals in November, new advancements from Anthropic and OpenAI have since placed Gemini back in a catch-up position. Internal testing conducted by Google revealed that the performance of Gemini's new flagship version continued to lag behind competitors, particularly in coding capabilities. This led to a two-month delay in its release, originally scheduled for August.

In response to these challenges, Google announced a significant leadership restructuring within its Google DeepMind AI division on August 5. Demis Hassabis, the division's chief, transitioned to the role of chair, relinquishing day-to-day management responsibilities to his deputy, Koray Kavukcuoglu. Concurrently, the two original technical co-leads for Gemini departed the company to establish their own startup.

These strategic shifts indicate Alphabet's commitment to achieving model supremacy, despite some industry speculation about Google potentially exiting the AI race. The reorganization also signifies a further reduction in DeepMind's autonomy, a trend that has been ongoing since Google acquired the London-based lab in 2014. The company aims to integrate AI more deeply into its software and enhance its commercialization strategy.