The term distillation has become fashionable lately thanks to... well, those guys. But distillation is more than that; maybe most importantly of all, it's the process by which we are allowed to enjoy (if properly conducted) some good old poured moonshine.
In a more metaphorical context, it's the process by which we get to the core of something, to what's really valuable about it. To what really matters.
Startup founders have lots of ideas. Guerrilla startup founders are not the exception. Let's say the velocity of those ideas is x100.
What in the world is the velocity of an idea? This is the frequency in which these ideas are generated. x100 velocity means nothing by itself but hey, we need to start reasoning somewhere.
So what to do with those ideas? Nothing for the moment. Just have them. Take notes if you please (phase 1).
The thing is, understanding how those ideas fit in our product needs some sort of extra process. We will call it distillation. The distillation process is slow — you must consider DevOps constraints, commercial constraints, marketing constraints, and so on...
So to distill an idea and make sure (or try to) understand if it fits in our products has a different velocity than idea generation. Let's say that this distillation process has a velocity of x10.
Sounds arbitrary, and it is. But think about it. I can come up with 40 ideas in a minute on how to improve our product. Should I try to implement them right away? Bollocks! That would be a mess; time and resources would be wasted on 40 random ideas that I think are fantastic. And the next day (or minute) I will have 40 more. A guerrilla startup does not have those kinds of resources to waste.
So we have to distill (phase 2).
Distilling takes time, discussions, damn meetings and stuff. Ok, fair.
Now we need to decide on which distilled ideas we are actually going to execute next (phase 3). Executing is a much slower process. Let's say its velocity is x1. In the AI era, executing ideas has, say... doubled? So x2. A staggering increase of 100% productivity in idea implementation time.
That is fantastic!
Let's just assume 0 technical debt, which is stupid, but ok — we're playing around, trying to get to the point.
Ideas have a x100 velocity, distilling ideas x10, implementing x1–x2. And that is to say, ideas don't matter much. In fact, they can get in the way. Distillation is not free; meetings are not free. It's time a guerrilla startupper could be there executing really valuable, distilled ideas.
You can have 1,000 ideas, 1B ideas. Nobody cares.
Distilling is what matters, and you can increase distilling velocity aided with AI (which also has a cost to add). Executing is very costly and slow; it used to be even more costly and even slower (it takes tokens, time, and general resources, plus it accumulates the wasted time of all the distillation process, in case it fails).
So next time you have a great idea for your product, first do not value it so much until properly distilled; prioritize the surviving ones correctly. And execute only when necessary.
There is a 4th phase of the whole process that is to adjust future ideas. But that is a whole other matter.
Phase 1: Dreams/Ideas (x100 velocity) Phase 2: Distillation (v1, v2, v3 features) (x10 velocity) Phase 3: Execution (v1) (x1 velocity before AI, x2 velocity with coding agents) --- Phase 4: Course correction → Back to Phase 1