The fastest way to slow a team down is to route every technical decision through one person. Here's how I avoid being that bottleneck.
When I took my first lead role, I reviewed everything. Every pull request, every schema change, every library choice waited on me. The team was busy and I was the bottleneck — and for a while we mistook my busyness for rigor.
What "the edge" means
The edge is the person closest to the problem — the engineer who has the failing test open, who has read the ticket three times, who knows the weird thing the legacy cron job does at 2am. They almost always have more context than I do. My job is to make sure they also have the authority to act on it.
I push technical decisions down to the people closest to the problem.
In practice that means I set the constraints — the deadline, the budget, the non-negotiables around security and data — and then I get out of the way. I review outcomes, not keystrokes. When someone makes a call I wouldn't have, I ask why before I ask them to change it. Half the time their reasoning is better than mine.
This is not abdication
Pushing decisions down is not the same as disappearing. I still own the architecture that ties the pieces together, the hiring that decides who's at the edge in the first place, and the air cover when a bet doesn't land. The difference is that the team stops waiting on me to think, and starts thinking out loud where I can hear it.
A team that can decide without me is the only kind that scales. It's also the only kind I'd want to manage.