Comment by cubefox

I assume: 1) They work at Google Research, not Google DeepMind. The latter seems to be more focused on developing AI products now, but Google Research seems to be still quite research based. 2) Google is overall much more liberal with researchers being able to publish their results, compared to OpenAI. I don't think this is necessarily rational from a business perspective where you get disproportionate payoff if you are a step ahead of everyone else. There is a reason why OpenAI has outpaced Google in terms of AI. Maybe those Google researchers were hired under the condition that they are allowed to publish their results?