The conversation about AI in enterprise technology has shifted. We've moved past "should we adopt AI?" and into the harder question: how do you lead effectively when AI is reshaping every layer of your stack?
I've spent the last few years sitting at that intersection — leading engineering teams at AMD while simultaneously building Libra Technologies — and I want to share what I've actually learned, not what sounds good in a conference keynote.
The mental model shift that matters most
Most engineering leaders are still thinking about AI as a feature — something you add to your product. The leaders pulling ahead are treating it as a capability layer that touches everything: how teams work, how code gets written, how decisions get made.
This isn't a subtle distinction. It changes your hiring bar, your architecture choices, your build vs. buy calculus, and your definition of "done."
What changes for engineering teams
Speed expectations are being reset. When a senior engineer with good taste and the right tools can ship in hours what used to take days, the bottleneck moves. It's no longer about raw throughput — it's about judgment, context, and knowing what not to build.
The leverage gap is widening. The gap between engineers who use AI tools fluently and those who don't isn't linear — it's compounding. If you're leading a team and you're not actively closing this gap for every person on it, you're creating organizational debt.
Your architecture decisions now have a half-life. Things that were "good enough" eighteen months ago may already be obstacles. The engineering leaders I respect are building more deliberately for optionality — favoring smaller services, clean interfaces, and the ability to swap in new capabilities without rewriting everything downstream.
What stays the same
The fundamentals of good engineering don't change. Clear thinking. Honest tradeoffs. Systems that are understandable by humans. Code review culture that catches problems before they compound.
If anything, the AI era makes these things more valuable — because there's more surface area for sloppiness to hide, and the pace at which sloppiness scales has gone up.
The actual hard part
The hard part isn't the technology. It's the people and the process.
Getting a team to genuinely internalize a new way of working — not just use a tool when reminded — takes time, repetition, and leaders who model the behavior they want to see. It takes being honest when old patterns aren't serving you anymore.
I don't have this figured out. But I've found that the leaders doing it best share one trait: they're genuinely curious about the technology themselves, not just delegating curiosity to their teams.
If this resonates, I'd love to hear what you're seeing on your end. The best learning I've done in the last few years has come from honest conversations with other people in the weeds.