Writing is thinking, presenting is deciding

One of the things AI is making us confront is that often we don't say what we mean. I guess it's a version of the genie/computers-being-very-literal problem.

So, for instance, when we say someone is 'a coder' we often don't mean 'someone who writes code'. We might mean 'someone who does all the complex social, organisational, emotional, logistical things that means the right code is written and the right people know about it and agree that it's a thing that should be done'. And also writes code.

I'm simplifying but you know what I mean.

Similarly, the other day I asked someone at work to write me a note about a project we were working on. And he did. He came back to me (much too quickly) with a note that had been written by AI. And, obviously, that was my fault. Because really what I meant by 'write me a note' was 'think about this problem, reflect on it for a while and then write down your thoughts, so you can transfer your thoughts to my head and we can think about it together.'

(TANGENT I'D LIKE TO COME BACK TO: I'd like to talk about friction sometime. I suspect we think 'write a note' is enough because the thinking is enforced by the format. Writing takes time, which leaves space for thinking. But we shouldn't fetishise the inconvenience. The thinking is the point, not the friction.)

A few years ago I was doing some consultancy for a lovely company who wanted to some strategy/communications help. I asked them if they had a presentation version of the existing strategy and they said they had - they excitedly sent me a Miro board. My heart sank.

As is often the way with Miro boards it was just a bunch of stuff, a compilation of thoughts, observations, facts, stories and questions. It was useful and interesting but it wasn't a good strategy and it wasn't a good presentation. (The two are often synonymous)

Presentations and presentation software (like strategy) are good because they're constrained. You have a limited amount of time and a limited amount of space. So you have to make decisions. You have to decide what comes first. You have to decide what gets left out. You have to prioritise.

(That's also the first thing that's instantly, obviously wrong with all these AI presentations that get shared under headlines like Claude Has Killed PowerPoint. What it should say is that Claude Has Automated Terrible Presentations. It's just a more efficient way to put too much stuff on slides.)

That's a good way to judge a presentation. Decisions per slide. If every slide represents a bunch of good decisions then it's probably a good presentation. Presenting is deciding.

Anyway.

Here's the video version: