In an internal message from Google’s Deepmind AI Division, the company’s co-founder Sergei Brin announced that employees needed to improve focus and productivity to stay ahead in a rapid race towards Artificial General Intelligence (AGI). Addressing the growing competition, Brin urged the AI ​​teams to urge the 60-hour work to adopt as a “sweet spot” for the maximum output, while warning to reduce employees who can obstruct the overall progress of the team.

In the last two years, Brin has made it clear with Google’s AI efforts that bets have never been more in search of developing AGI. He insisted that Google’s Deepmind Division has been referred to as GDM, two years ago since the launch of its Gemini program, there has been great progress, the pressure to succeed has intensified as the contestants have increased their efforts.

Brin wrote in communication, “The last race for AGI,” Brin wrote in communication, which The Verge has seen. He expressed confidence in Google’s ability to win, but also said that the team would need to “turbocharged” their efforts to maintain the lead. The central code performance for Brin’s message was important, as they believe that AGI is probably likely to emerge when AI systems can improve themselves. “Code matters the most,” he said, highlighting the need for efficient coding practices within the company’s own system.

A remarkable aspect of Brin’s message was his stand on working hours. He mentioned that 60 hours per week is the optimal balance for productivity, based on his experience. Accepting that some employees may work beyond this limit, they warned against overaxiting, which could lead to burnout. On the other hand, he criticized the lesser working people and said that they could destroy the rest of the team. Brin said, “A small number to achieve bare minimum,” Brin said, called more commitment to all employees.

Brin also highlighted the importance of in-tradition cooperation, urged team members to work in office every week. They believe that face-to-face communication is far more effective than virtual options, which can lead to misunderstandings and disabilities. Additionally, Brin advocated reducing reporting lines at various places to simplify team structures and ensure smooth workflows.

In terms of technical direction, Brin advised against adding unnecessary complications to his models, such as post-training or highly complex solutions, when simple approaches such as prompting may work. “Let’s use simple solutions where we can,” he said.

Brin also urged the team to focus on speed and efficiency in terms of product development and internal equipment. He suggested that small scale experiments should be preferred, allowing quick recurrences and beliefs before scaling.