As I shared previously, I don’t think AI is killing software engineering; it does however transform it. The rise of AI coding assistants—like GitHub Copilot, Cursor, and others—has fundamentally changed how we build software. But more importantly, it’s changing who we are as engineers.
We’re no longer just writing code. We’re shaping product—fast, and often with a healthy dose of ambiguity.
From Vibe Coding to Product Shaping
Let’s talk about “vibe coding”: if you’ve been using Cursor or Copilot, you’ve probably found yourself throwing a vague prompt into your editor—”add support for dark mode” or “make this onboarding smoother”—and getting something surprisingly workable back. It’s not always perfect, but it’s good enough to nudge you forward. You iterate, refine, prompt again. Suddenly, you’re not thinking in terms of writing code line-by-line. You’re shaping product experiences in real time, directly in the editor.
This way of working doesn’t rely on long, detailed specs. It leans heavily on intuition, rapid feedback loops, and the ability to make small product decisions on the fly. In other words, the skillset of a product manager.
The Disappearing Middle Layer
Historically, engineers have always dabbled in product. We’ve made micro-decisions every day: should we cache this result? Should we debounce that API call? Should we preload a bit of state so the UI feels snappier?
These are product decisions at the technical edge—subtle but impactful. But with AI accelerating our ability to generate and iterate on code, our sphere of influence is expanding fast. We’re not just choosing caching strategies anymore—we’re designing flows, adjusting onboarding logic, exploring different feature variants before a PM even sees them. The decisions we’re making are creeping upward, into what was once the realm of junior or mid-level product managers.
AI accelerates iteration. And iteration is the currency of product development.
As a result, that middle layer of product management—writing out stories in Jira, shepherding granular UI flows, specifying exact button behaviors—is starting to blur into the engineering role. Not disappear, but dissolve, as engineers with strong product intuition become the ones driving those decisions in real time with AI as their co-pilot.
Embracing Ambiguity
This new model rewards engineers who are comfortable operating in ambiguity—those who don’t need every requirement handed to them, who can take a loose idea and mold it into something real. It favors engineers who think in terms of UX impact, who care about user outcomes, who can write just enough to get feedback and pivot quickly.
In this world, it’s not just about writing efficient or clean code. It’s about writing directionally correct code—then refining it through interaction, usage, and iteration. AI makes that possible, but the mindset shift has to come from us.
The Future: Engineer as Builder-Shaper
If this trend continues—and there’s every reason to believe it will—the future of software engineering looks less like a factory floor and more like a design studio. Engineers won’t just be implementers. They’ll be builders-shapers—rapidly prototyping ideas, driving UX decisions, and owning bigger pieces of the product vision.
This is not something new, we have seen for a while now the CPTO (Chief Product and Technology Officer) role cropping up more and more in various companies; AI is now accelerating this transformation. And product management isn’t going away. But it is moving up a level—focusing on vision, strategy, and alignment. Meanwhile, the line between engineering and product at the tactical level is blurring fast. And the engineers who thrive will be the ones who embrace that blend.
