I love this post. I think it's spot-on for where I think things are headed.
TikTok indeed is an existence proof that taste can be captured digitally. Where we converge at an equilibrium may literally just be an oligopoly of models where many implicit assumptions are embedded, with different "tastes" which cater to some specific subsegment of the population. Kinda like politics if you think about it with different parties and "taste"
This is grossly over-engineered. Taste isn’t a system or checklist and it’s certainly can’t be taught or given to someone.
Even someone who’s put in their mythical “10,000” hours isn’t guaranteed to have any taste as we all know a millionaire (or billionaire) who has none.
Taste is a feeling. And some folks have it and some will never have it. That’s why we have tastemakers and everyone else enjoys their work.
Trying to distribute it broadly or teach it is exactly the opposite of what it is. It’s impossible to capture in a blog post or put in a slogan or marketing campaign. It just… is.
I actually agree with a lot of this. Taste is real, it matters, and not everyone has it.
My argument is less about teaching taste and more about what counts as taste continually changing. Five years ago, "good taste" in software design meant knowing how to architect a clean system. Today (I'd argue more and more) AI handles that, so taste moved up to knowing what to build and why. The person with great architectural taste didn't lose it; the baseline just absorbed it.
That's why I frame it as alpha rather than a moat. The tastemakers you're describing are real, but their edge isn't static. What made someone a tastemaker last year becomes the default this year, and they have to find the next layer of judgment that the machine can't replicate yet. "They either have it or they don't" is true at any given moment, but what "it" is keeps shifting underneath them.
But that’s the thing. Taste doesn’t change as it’s already and always related to the time in which it is discovered or applied. It is the “ghost,” the shadow of great products and things well-done.
Even you saying “taste moves up” is trying to apply it to some layer of abstraction like an engineering food chain. It is way outside all of that.
This might be semantics. For me it's <gap between what AI produces by default and what "great" looks like>. You could also call it judgment or discernment.
AI on its own does not usually produce "great" things but with 'judgment' it can. The post then is really about the value of human 'judgment' being relative to what AI can do out of the box.
I love this post. I think it's spot-on for where I think things are headed.
TikTok indeed is an existence proof that taste can be captured digitally. Where we converge at an equilibrium may literally just be an oligopoly of models where many implicit assumptions are embedded, with different "tastes" which cater to some specific subsegment of the population. Kinda like politics if you think about it with different parties and "taste"
As you approve non-objective (non-verifiable) "taste"
And we know that non-objective taste tends to polarise
This is grossly over-engineered. Taste isn’t a system or checklist and it’s certainly can’t be taught or given to someone.
Even someone who’s put in their mythical “10,000” hours isn’t guaranteed to have any taste as we all know a millionaire (or billionaire) who has none.
Taste is a feeling. And some folks have it and some will never have it. That’s why we have tastemakers and everyone else enjoys their work.
Trying to distribute it broadly or teach it is exactly the opposite of what it is. It’s impossible to capture in a blog post or put in a slogan or marketing campaign. It just… is.
They either have it or they don’t.
I actually agree with a lot of this. Taste is real, it matters, and not everyone has it.
My argument is less about teaching taste and more about what counts as taste continually changing. Five years ago, "good taste" in software design meant knowing how to architect a clean system. Today (I'd argue more and more) AI handles that, so taste moved up to knowing what to build and why. The person with great architectural taste didn't lose it; the baseline just absorbed it.
That's why I frame it as alpha rather than a moat. The tastemakers you're describing are real, but their edge isn't static. What made someone a tastemaker last year becomes the default this year, and they have to find the next layer of judgment that the machine can't replicate yet. "They either have it or they don't" is true at any given moment, but what "it" is keeps shifting underneath them.
But that’s the thing. Taste doesn’t change as it’s already and always related to the time in which it is discovered or applied. It is the “ghost,” the shadow of great products and things well-done.
Even you saying “taste moves up” is trying to apply it to some layer of abstraction like an engineering food chain. It is way outside all of that.
This might be semantics. For me it's <gap between what AI produces by default and what "great" looks like>. You could also call it judgment or discernment.
AI on its own does not usually produce "great" things but with 'judgment' it can. The post then is really about the value of human 'judgment' being relative to what AI can do out of the box.
That I can agree with… until AI does something that most folks would deem tasteful. I think it’s coming. I’m not sure how we will respond to that!