The Software Development Lifecycle Is Dead
March 6, 2026
AI agents didn’t make the SDLC faster. They killed it.
I keep hearing people talk about AI as a “10x developer tool.” That framing is wrong. It assumes the workflow stays the same and the speed goes up. That’s not what’s happening. The entire lifecycle, the one we’ve built careers around, the one that spawned a multi-billion dollar tooling industry, is collapsing in on itself.
I imagine this piece will get quite a bit of pushback from a lot of software engineers who will argue that while this approach might be fine for greenfield projects, for small teams or individuals building something unencumbered by years or decades of existing coding systems, it’s not true of larger, more complex systems, particularly long-standing ones.
Perhaps we’ll see a bifurcation of existing systems, let’s call them legacy systems, which will move toward maintenance mode, and greenfield systems built from day one with agentic programming at their heart.
Perhaps agentic programming will in time take over the maintenance and development of large existing systems. I think it’s far too early to say. But I think perspectives like this are important whether you agree with them or not because they can help shape your own reasoning about the work that you do and the systems you’re responsible for.
The frame of 10x developer always assumed SDLC process stayed constant. But youre pointing deeper—when AI agents can rewrite the problem statement mid-build, theres no cycle anymore, just turbulence. Skill shifts to knowing when to let go.