I hit inbox zero this morning in under nine minutes. That used to feel like a win. It no longer does, and understanding why has changed how I manage every channel of incoming communication, not just email.
The old problem with email was effort. Replies took time, and time was finite, so you either answered or you didn't and felt vaguely guilty about it. Inbox zero was a solution to the effort problem. Empty inbox, no outstanding replies, clear conscience. The system worked when effort was the constraint.
Effort is no longer the constraint. The model drafts faster than I can decide what needs a reply. So the problem changed, and the goal needs to change with it.
What the new problem actually is
When replies cost nothing, the scarce resource is not effort. It is attention — mine when I'm deciding what to send, and the recipient's when they're processing what they receive.
The goal of inbox management is no longer clearing the queue. The queue clears itself. The goal is protecting the part of your day that isn't email, and protecting the relationship you have with the people you write to. Both of those require exercising more judgment about what gets sent, not less.
Inbox zero as a concept incentivizes sends. Empty inbox, full outbox. In the old system that was fine, because the cost of sending was real and that cost acted as a filter. Remove the cost and you remove the filter. The result is more email sent by everyone, to everyone, about everything — which is roughly where we are.
A reply you didn't send is not a failure. It is a boundary.
Inbox enough
I now do one pass per day: thirty minutes, same time, non-negotiable window. The model drafts everything I might want to send — replies, follow-ups, acknowledgments, declines. Then I read each one in my own voice and decide.
About a third get sent close to as-written. A third get substantially rewritten, usually because the model was politer than I wanted to be or vaguer than the situation required. A third don't get sent at all. Reading the draft helped me realize I didn't actually need to reply — the email was asking for something I couldn't give, or noting something that required no acknowledgment, or following up on something I'd already handled.
That last group is the most valuable output of the thirty minutes. Every non-send is an email thread that doesn't grow. Fewer threads means fewer replies expected. Fewer replies expected means a lighter inbox next week.
The triage framework
Three categories, evaluated in order:
Requires a decision I can make now — answer it. The model drafts; I adjust; I send. Move on.
Requires a decision I cannot make now — schedule the decision. One line in reply: "I need to check [X]. I'll follow up by [date]." The model drafts this too. The key is the date. Without a date, this is a non-answer that generates another follow-up. With a date, it's a commitment.
Does not require a decision — evaluate whether it requires acknowledgment. Most acknowledgments are social friction with no functional content. "Got it" is often a second email that accomplishes nothing. If the sender would not be confused or concerned by silence, silence is often the right reply.
The inbox is not a to-do list. It is other people's to-do lists for you. Inbox enough means you process it on your schedule, using your judgment, without letting it set your agenda for the day.
What this requires
The hard part is not the tool or the system. It is the belief that not every email deserves a reply, held firmly enough to act on it when the cost of replying has dropped to near zero.
That belief is surprisingly difficult to maintain. The social pressure around responsiveness is real, and it hasn't updated for the new economics. People still judge by reply speed even though reply speed is no longer a meaningful signal of effort or care. It's a signal of availability, which is a different thing.
Protect your availability. The cost of your attention hasn't dropped. Reply because something requires a reply, not because the model made it easy.