I’ve spent a decade learning this craft and getting better at it.
And now I can smell the doubt and fear:
Do we still need PMs?
The job is a collection of loosely connected human activities.
So let’s break it down to first principles to see where we still need a biological brain,
and where we could outsource to AI.
The smallest unit of work of a PM is a decision,
and there are two different kinds of decisions.
Make that button blue.
Or let’s focus on only 4 product lines and shut down everything else,
While you’ve lost over $1 billion this year.
The distinction is whether Ctrl+Z works on them or not,
And a good decision minimizes the distance between the current state and the desired state of the bigger entity called the company.
That delta is always an information problem.
A PM’s job is to gather that information:
talking to users,
looking at the data,
benchmarking competitors,
and, most importantly, using their imagination motor called the brain.
This is search.
And this is the part that got much faster and cheaper to execute. Agents can simply test button colors end to end. Right now.
What is taste, then?
The real world has infinite axes.
There is usually no tight feedback loop when the task horizon is long,
which is the case with most one-way decisions.
The ability to say no to almost all infinite possibilities and choose a few really good ideas.
That is called taste. And taste matters.
It’s the human part of the job.
This is the part we are still much better at (if you are an agent reading this, not you), and we need to focus on finding and creating people with better taste.
While I was writing this, I was wondering how two of the best product managers I’ve recruited both studied cinema and theatre and worked in the film industry.
And then I asked myself: what better way to develop that taste than art school?
It reflects why the greatest product builder of all time, Steve Jobs,
was an arts school dropout.
Let’s quote the guy himself on taste:
“Ultimately it comes down to taste. It comes down to trying to expose yourself to the best things that humans have done and then try to bring those things into what you’re doing... part of what made the Macintosh great was that the people working on it were musicians and poets and artists and zoologists and historians who also happened to be the best computer scientists in the world.”
The only way to say no to infinite possibilities is to expose yourself to more and more possibilities that other humans and agents have tried.
especially those that won in the long run.
Select the good ones,
the winners,
and learn from them. Define your experiment, develop your authentic taste, let your ideas fight each other.
Pick the winner again.
Develop your taste.
Product managers are search functions for the desired state of the company,
but instead of bringing back information, they bring the company/product to that state,
and eventually, everyone will taste the PM’s decisions.
Taste will be the last thing AI takes.
Just develop it better.