My read on pushing collaboration is that it's a hedge against doing the wrong thing. In startups (say, <100 people and in grow-or-die mode), I think you have to

My read on pushing collaboration is that it's a hedge against doing the wrong thing. In startups (say, <100 people and in grow-or-die mode), I think you have to focus on producing revenue on a 3-9 month timeframe. If you have a bunch of people working on their own, a lot of them will do things that are valuable, but won't show value to the company for 2-5 years. That not really OK, unless everybody agrees it's OK. So, everybody collaborates constantly, so when you do something like run a pen-test before you have a customer with a CISO, there's agreement that it's OK to do it early. Otherwise, you might end up with a bunch of interesting work that doesn't have any coherence, and with only 25 people, that's not enough to sustain a company. Of course, if everybody is working on the right things, and you have manager drive-bys via Slack/email/daily standups 3 times a day, that's a separate issue. It's the fear of working on the wrong thing that's motivating that pattern.