It really feels like it’s time for a positive perspective. So here it goes.
On a personal level, seeing the recent posts about great people leaving DeepMind hits home. Leaving DM was incredibly hard, and that's entirely down to the amazing people: both the friends you closely know and all the other brilliant minds giving internal talks. Every person who leaves is a missed opportunity for a great hallway conversation.
But looking beyond the personal level, I’m not remotely worried about the team's ability to deliver. The key to the impact of Gemini 3 wasn’t individual heroes. A cult of personality doesn't work (alone) for frontier research... and by now, the industry has seen clear examples of that.
Success comes from an organization and processes optimized for tight iteration and hillclimbing amazing models. (One reason why I think we'll reach again a time of increased publication density for individual method advances.) And that is something the company has known how to do better than anyone at smaller scale for a long time.
I’m immensely looking forward to Gemini 3.5 and 4!
That's not to say there aren't good reasons to try something new - no matter where you are! Btw we're hiring 😉!
