The AI Standoff: Why Software Engineers Must Become Product Managers
Engineers, designers, and product managers all think AI has made the other two redundant. But the real winners will be those who learn to wear all three hats.
Engineers, designers, and product managers all think AI has made the other two redundant. But the real winners will be those who learn to wear all three hats.
The introduction of AI coding agents has created a Mexican standoff between engineers, designers, and product managers, with each believing they can replace the other two roles. To truly leverage the power of AI, software engineers must step out of their silos and learn to architect projects with the mindset of a product manager.
There's an interesting trend happening in tech right now. In a recent episode of Lenny's Podcast, Marc Andreessen summarized the current dynamic as a "Mexican standoff" between the product manager, the designer, and the coder.
They are all pointing guns at each other, secretly believing that AI has made the other two redundant:
They could all actually be right.
The old way of staying in your own little box and just doing one job is over. The winners of this next era won’t be the people aggressively protecting their titles or gatekeeping their departments.
The winners will be "superpowered individuals."
Basically, if you’re a coder who uses AI to design, or a designer who uses AI to actually ship a fully functioning product, you become a "triple threat."
If you aren't using these tools to expand what you can do, you're the one standing in the circle with no ammo.
Eventually, the day will come when product managers and designers can fully wear a software engineer's hat and ship complex, scalable, secure applications end-to-end.
But that day is not today.
Right now, knowing how to structurally architect a codebase and gracefully direct a software engineering project is critical to the success of an AI-assisted software product. AI coding agents are incredibly powerful, but they still necessitate engineering intuition to build robust systems from scratch without collapsing under their own weight or spiraling into architectural spaghetti.
This gives Software Engineers an unfair advantage in the standoff—if they choose to use it.
However, to really experience the full benefit of AI Coding Agents, Software Engineers need to understand how to interact with designers and product owners as though they are a product manager.
It is no longer enough to just grab a ticket, build it to spec, and merge your pull request. You have to understand why you are building it. You need to develop a sense of product taste, user empathy, and business logic.
When you sit down with an AI, you aren't just writing code anymore. You are acting as the Tech Lead, the PM, and the Developer all at once. You have to synthesize the design requirements, the product goals, and the technical constraints into a coherent prompt strategy.
As a Software Engineer learns how to wear that Product Manager hat, something happens. They stop just outputting code, and they start building better software faster that more users are delighted by.
So lower your weapon in the standoff. Stop worrying about who is going to replace whom, and start learning how to do their jobs instead.