Google’s DeepMind is feeling the heat in the race toward Artificial General Intelligence (AGI), and co-founder Sergey Brin isn’t holding back. In an internal message, Brin emphasised the need for increased focus, productivity, and commitment to keep up with the competition.
He sees a 60-hour workweek as the “sweet spot” for maximising efficiency, warning that underperformance could slow the team down. Here are the key takeaways from Brin’s push to keep DeepMind ahead in the AGI race.
60-hour workweek for peak productivity
Brin believes that working 60 hours per week strikes the right balance between output and burnout. While he acknowledged that some employees might push beyond that limit, he warned against overwork, which could be counterproductive. However, those working significantly less were criticised for dragging the team down.
Face-to-face collaboration is key
Virtual meetings aren’t enough. Brin urged full-time in-office work, stressing that face-to-face interactions lead to better communication and fewer inefficiencies. He also advocated for simplifying team structures and reducing unnecessary reporting layers across locations.
Prioritising speed and simplicity in AI development
Brin emphasised the need for faster iterations and fewer unnecessary complexities in AI models. He suggested avoiding excessive post-training tweaks when simple solutions, like better prompting, could achieve similar results. Small-scale experiments should be prioritised before scaling up.
Efficient code is at the heart of AGI
Brin sees self-improving AI systems as the gateway to AGI, making code quality a top priority. He urged the DeepMind team to focus on writing highly efficient code, as he believes this will be the defining factor in AI evolution.
Brin’s message underscores the intense pressure on Google to stay ahead in the AGI race. While the 60-hour workweek and in-person mandates might stir debate, one thing is clear — Google isn’t taking any chances.
