You're identifying a real shift, and honestly, the answer isn't clear-cut yet.

**What you're actually doing:**
- Architectural decisions (what to build)
- Quality gates (is this correct?)
- Intent translation (turning vague ideas into precise prompts)
- Debugging the AI's logic, not syntax
- Knowing when to trust it vs. reject it

That's legitimately more *design-adjacent* than it used to be. Less "implement the spec" and more "shape the implementation."

**But here's the thing:**

**It's not product management** because you need deep technical judgment to:
- Know if the AI's approach will scale
- Catch subtle bugs in generated code
- Understand the tradeoffs the AI missed
- Know *when* to override it

You can't PM your way through that.

**It might be closer to:**
- **Architecture** (deciding structure, patterns, approach)
- **Code review elevated** (except you're reviewing before it's written)
- **Embedded QA with domain expertise**

The uncomfortable truth: *the title hasn't caught up to the work yet.*

Some companies will call this "Staff Engineer" or "Architect." Others will just compress it into fewer hiring slots and call the survivors "Senior Developers" while paying them like mid-level because "AI does the coding now."

**Real question:** Do you prefer this work? Because if you do, lean into it. If you don't, that's a signal too.