← Chris Barry

The Meeting That Should Have Been a Diff

March 19, 2025

We were forty minutes into a meeting about a proposed change to how we handled authentication tokens.

Someone had drawn a diagram. Someone else had questions about the diagram. A third person was concerned about backwards compatibility. A fourth was worried we hadn't consulted the security team.

I was staring at my laptop, half-listening, when I realised I had no idea what we were actually changing.

Not the concept. The concept was clear. But the actual change — the lines of code, the files affected, the before and after — none of that was visible. We were discussing an abstraction of a change, not the change itself.

After the meeting, I asked the engineer who'd proposed it if she had a branch. She did. I read the diff. It took four minutes.

The change was smaller than I'd imagined. Most of the discussion had been about edge cases that didn't exist in the implementation. The backwards compatibility concern was already handled. The security implications were obvious from the code — or rather, obviously fine.

We'd spent an hour discussing something that would have taken five minutes to read.

This happens more than anyone admits.

Meetings are good for some things. They're good for building relationships, for navigating ambiguity, for situations where the path forward genuinely isn't clear. They're good when you need to read the room.

But they're terrible for reviewing concrete changes.

A diff is precise. It shows you exactly what's changing and what isn't. It doesn't let you imagine problems that don't exist. It doesn't reward the person who speaks longest or loudest.

A meeting, by contrast, rewards performance. The person who raises the most concerns sounds thorough. The person who asks the most questions sounds engaged. The person who stays quiet might be the only one who actually understands what's happening.

I've started noticing a pattern. The teams that move fastest tend to have a bias toward asynchronous review. They write things down. They share branches early. They treat meetings as a last resort for when written communication has failed.

The teams that move slowest tend to have a meeting for everything. Kickoffs, syncs, reviews, retros. Every decision requires a calendar invite. Every change requires a presentation.

It's not that meetings are bad. It's that they're expensive, and we've stopped noticing the cost.

The next time someone proposes a meeting to discuss a change, it might be worth asking: is there a diff we could read instead?

Sometimes the answer is no. But sometimes the answer is yes, and everyone's been too polite to say it.