The Armageddon Moment for Product Teams
AI is forcing product leaders to ask a question the movie industry answered in 1998: which direction is easier to train?
There’s a scene in the DVD commentary for Armageddon where Ben Affleck asks Michael Bay a simple question: why is it easier to teach oil drillers to become astronauts than to teach astronauts to become oil drillers?
Bay didn’t have a great answer. But the question stuck.
AI is putting product teams in a similar moment.
The Armageddon Question
As AI coding tools get better, product-shaped folks can now ship working code. Not great code. Not production-ready code. But real, functioning code. Vibe-coded pull requests from your UX designers are a thing.
At the same time, engineers have always had the option (although not always the invitation) to develop product instincts. Spend more time with users. Think about outcomes instead of outputs. Some do. A lot don’t. Engineering orgs usually reward shipping code over developing product judgment.
So here’s the Armageddon question for your product team: what’s actually more efficient? Training a PM to ship code, or training an engineer to think like a product person?
Why This Isn’t a Simple Answer
“But Tony…” you say incredulously, “how didn’t you think of just hiring product-minded engineers? Or building a culture where everyone thinks about the user? Isn’t that clearly the right answer?” In a pitch-deck perfect world, yeah.
We live in an imperfect reality. One where Chick-fil-A is closed on Sundays, and product thinking and engineering thinking are importantly different modes.
Some of the best product designers I know can’t read a stack trace. Some of the best engineers I know can rebuild your infrastructure and still miss why users are churning.
AI lowers the barrier to execution. A PM with the right tools can prototype, iterate, and even ship surface-level features faster than ever. This is awesome. Everybody should be building right now.
But it doesn’t replace the judgment required to decide what should exist in the first place. Or how something behaves in edge cases. Or whether the interaction model actually makes sense for the people using it.
The oil drillers made it to space in the movie, but they still needed NASA.
What This Means for Product Teams Right Now
The teams getting this right aren’t haphazardly trying to collapse one role into another. They’re asking a sharper question: where does human judgment create the most leverage?
In practice that usually looks like:
- Engineers spending less time on boilerplate and more time on systems thinking
- PMs prototyping earlier and validating assumptions before anything gets built
- Designers spending more time thinking about the product’s concepts, taxonomy, and all the areas where workflows fall apart. Basically, focusing on the decisions AI can’t make: what this should feel like and why
The roles aren’t going away, but the boundaries between them are shifting. The teams that figure out the new shape of those boundaries early will have the real advantage.