Why product managers have a massive engineering opportunity in the AI world?

Photo by Austin Chan on Unsplash

I’ve been thinking deeply about the kinds of trends that are going to emerge in the AI world and one specific dimension is the role of a product manager. I believe that the product managers of today have a massive opportunity on their hands, only waiting to be unlocked!

What am I talking about? I wrote about the courage it takes to be a product manager in my last post. Product managers have to deal with a lot of uncertainty - they have to make decisions on which customers to talk to, which problems to solve, how to approach a solution, how to go to market, etc. They have to constantly make decisions like these based on some probability distribution. If they try and chase 100% certainty, they will never ship anything useful. This is the core of their job and they have to get really really good at it.

Similarly, in the AI world the solution you provide to your customers is going to be a probability distribution. It is quite different from the non-AI programming paradigm which is deterministic (if this happens then do that). Most of the world’s software today is written in a deterministic paradigm.

With new tools and techniques popping up every day, it is also becoming easier and easier to build technology-based solutions. It was much harder to be an engineer a decade or two ago…fast forward to today and you can get a lot of things done just by telling a LLM what you need. It is only going to get easier to build a majority of the software we see in the world today.

So, let’s intersect these three things -

1. product managers ability to be really good at making decisions in a stochastic space,

2. AI-first solutions that are going to be non-deterministic, and

3. the advancement of AI based development tools to build software very quickly.

Based on this, I believe that product managers could be the new generation of engineers who not only understand their customers and business problems but are also the ones who build the actual solutions. They can get comfortable relatively quickly on the solutions they are building, ship and then iterate.

Now you may ask, if AI has a probability distribution, doesn’t it imply that the developer tools will only provide a solution that gets you to a certain point and then you have to make tweaks to fit your need. So, you then you need an actual engineer?

I’m glad you asked! This is the opportunity I’m referring to. I believe product managers can be those engineers themselves. You do need to learn new programming tools and techniques, but it is WAY easier now than trying to learn how memory allocation worked in the old days! I believe the product managers that pick up those skills today are going to be the pioneers who lead us to this new world. Product managers will quickly go from deciding what to build to building those solutions themselves.

I’m bullish about setting this trend and I’m going to pay it forward by training those pioneers. So, keep an eye out for that update coming shortly! In the meantime, don’t forget to subscribe to my newsletter because that is how you’re going to get access to it!