April 29, 2026
On my way to the agenda committee meeting for NDC Oslo this morning, I randomly had what felt like an insight. What happened (I think) is that my brain took a proto-insight that had been floating around in its sea of neurons for quite some time and suggested a couple of words to describe it. On playing those words back to itself, my brain quickly decided that my brain was onto something.
The insight was this: many prospective speakers seem to think that the job is to find a topic to talk about, but the real job is to articulate and convey a message. It’s not enough to figure out what you want to talk about, you also need to figure out what you want to say about it. Unfortunately many speakers seem to stop short of this important step.
When you read submissions, there are some tell-tale signs that a speaker has found a topic but not a message. Often you can see it right in the title. There’s a certain kind of “frilled-up generic topic” title that is easily spotted when you start noticing the pattern.
Say the speaker decides they want to talk about refactoring, a tried and true topic that has been around for a long time. They can’t simply put “refactoring” as the title of their talk, so they need to dress it up. But how, and with what? Clichés, unfortunately. Hence we get talk titles like “The hitchhiker’s guide to refactoring”, “Refactoring for fun and profit”, and “So you think you know refactoring?”. But of course this fools no-one but the prospective speaker. When I read a title like that, my brain automatically filters out the fluff, leaving “The hitchhiker’s guide to refactoring”, “Refactoring for fun and profit”, and “So you think you know refactoring?”. We’re back where we started, with the naked topic exposed.
Instead of dressing up topics, prospective speakers should take the difficulty of coming up with a meaningful title that can stand on its own as a sign that they don’t yet have a message. If you want to talk about refactoring, what is it that you want to say? Here you basically have two choices. Either your message needs to be “here’s my practical experience with refactoring in my concrete context” (draft title: “How we do refactoring at my place”) or you need to find something original and worthwhile to say about refactoring in general, something that hasn’t been said a million times before (good luck!).
Finding something original to say is hard and can probably not be forced. Instead of you finding it, it should probably find you! It may appear in the form of a proto-insight that is planted in your brain through your practical experience, simmers for a while, and then suddenly starts manifesting itself in your brain in the form of words. It may be possible to accelerate, nurse and midwife this process a bit by taking walks or sitting in a hammock and talking to yourself in your head. That is, through thinking. At least it’s worth trying. If nothing happens, maybe there wasn’t a seed in your head, or you need to give it more time.
Generative AI tools are dangerous here, because they are really good at taking a topic - any topic - and transforming it into a plausible-looking abstract with a plausible-looking title. On the face of it, it looks just like an acceptable talk submission. But what about the message? Your message? With just a generic abstract about a generic topic, chances are your submission won’t be good enough to make the cut. The tools seem to help you, but really they bereave you of the process you need to pass through in order to find your message.
As a final complication, once you think you have a message, you need to engage in another little bit of introspection. Is it your message or are you just parroting something you heard? Have you heard many others saying the same thing? Could it, in fact, be the marketing message of certain big actors in the tech industry? If so, you need to reconsider. We don’t need more of that. Go back to the hammock and think some more. I’m sure there’s something genuinely you in there.