Last month, AI became the number one reason for layoffs in the United States for the first time ever. In the same two-week window, product management job openings hit 7,300 — the highest level since 2022. Both of those facts are true. Both are from reputable sources. And if you think they contradict each other, you’re looking at this wrong.
PM demand is surging, but not for the role you remember
Let’s start with the data that should end the “AI is replacing PMs” panic. According to Lenny Rachitsky’s March 2026 analysis, open PM roles globally are at a three-year high — up 75% from the 2023 low and climbing 20% since January alone. AI PM roles went from 201 to 1,135 in that same period. That’s a 465% spike.
But here’s the catch. Challenger Gray’s April report shows 15,341 AI-driven job cuts in March — 25% of all layoffs that month. AI PM is exploding. Junior and generalist PM is disappearing. The market isn’t growing or shrinking. It’s polarizing.
The full-stack builder is replacing the traditional PM
These 7,300 roles don’t look like the PM job you held three years ago. In December 2025, LinkedIn’s CPO Tomer Cohen killed the company’s Associate Product Manager program and replaced it with a new role: Associate Product Builder — a hybrid role that combines an Engineer, a Designer, and a PM in one.
As Naval Ravikant put it: “Vibe coding is the new product management.” Cohen’s five skills for this new builder? Vision, Empathy, Communication, Creativity, Judgment. Notice what’s missing — writing requirements, managing backlog, sprint planning.
Look at Kilo Code startup story: 14 engineers, zero product managers, 15+ features shipped per week, a million developers in under a year. One of their engineers built a subscription product in just three days. Their people function as mini-CEOs: part strategist, part architect, part product manager.
AI amplifies your judgment, or your blind spots
So if the old PM toolkit is obsolete, what separates the people who thrive from the ones who don’t? Not AI fluency. Business judgment.
If you understand segments, jobs-to-be-done, and unit economics, AI makes you devastatingly effective. If you don’t, it scales your blind spots just as fast. LinkedIn expected AI to be the great equalizer — juniors would close the gap with seniors. The opposite happened. As LinkedIn’s CPO told Lenny Rachitsky: “Top performers adopted AI fastest. They had the judgment to know what to ask for, how to evaluate the output, and where to apply it.”
I see this constantly. A builder recently told me: “I feel like a rocket from vibe coding, but I can’t convert it into results.” Speed without direction. Across the product teams I’ve coached, the pattern is identical: the people who get 10x results from AI are the ones who already knew what to build before the tools showed up.
The ladder that builds experts is breaking
Here’s what nobody tells you about this story: the ladder that used to produce those experts is collapsing. CS graduates now have a 6.1% unemployment rate — higher than liberal arts majors. Seventy percent of hiring managers say AI can do the work of interns. Big Tech new grad hiring has dropped 50% since 2019, and 80% of “entry-level” Bay Area positions now require two or more years of experience.
This isn’t an elevator where everyone rises. It’s an escalator where the top moves up and the bottom moves down. If juniors aren’t learning on real product decisions today, who becomes the senior PM your startup needs in 2030?
50x speed means MANY MORE decisions
This is the landscape we’re operating in. The question is what we do about it.
Development cost has dropped to near-zero. You can prototype in hours, not weeks. That sounds like pure upside until you realize what it actually means: decision points explode. Fareed Mosavat, former VP Product at Reforge, described it well: “I can whip out a spec in 12 minutes instead of 7 hours. But the micro-decisions — product sense, taste, prioritization — are harder now because the realm of possible solutions is wider.”
When you have six AI agents running in parallel, the constraint isn’t build capacity. It’s your ability to decide what to build next. A Harvard Business Review study with BCG found that workers overseeing AI output experience 33% more decision fatigue and make 39% more major errors. Building faster is the easy part. The hard part is knowing what’s worth building. This works when you have the business judgment to navigate it. Without it, 50x speed is 50x ways to build the wrong thing.
What to do this week
Every founder with a credit card and an AI subscription can ship an MVP before lunch. The constraint is no longer building — it’s deciding.
Look at your last three product decisions and ask whether AI made them faster or better. If the answer is only “faster,” you have a speed problem disguised as a productivity win. Fix that before you hire your next PM.
