Elena Verna was midway through a stellar run at SurveyMonkey when the Netflix recruiter called. She was excited—until the first question landed: "How many people are you managing?" She had three reports. The role required fifteen. Interview over. Not her output. Not the budget she controlled. Not her ability to solve the problems Netflix actually faced. Just headcount. That single metric disqualified her before she could say a word about the work.
Today, Verna is an IC again. No direct reports. She ships growth campaigns, codes her own pricing pages, runs partnerships solo at Lovable. She's paid like a VP. And she's not an outlier—she's the leading edge of a career inversion that most hiring managers haven't yet noticed, even as it reshapes who gets hired, who gets promoted, and which middle managers quietly discover their role no longer exists.
The inversion is structural, not aesthetic. AI tooling has collapsed the coordination tax that made management the only path to leverage. The old ladder assumed that one person couldn't own a full growth campaign or ship a pricing page without handing off to designers, engineers, analysts, and a PM to stitch it together. Coordinating those handoffs was the job; the title you earned for doing it well was "manager." But when a single IC can execute end-to-end—write the copy, prototype the page, instrument the analytics, deploy to prod—the ladder doesn't just lose its appeal. It loses its economic justification. The people driving that output are overwhelmingly ICs with enough context and tooling to move from idea to shipped feature in days, not quarters.
The candidates who optimized for climbing the ladder spent their energy on coordination theater: the decks, the alignment meetings, the carefully managed stakeholder maps. They didn't maintain their craft.
The hiring crisis is that almost no one outside AI-native companies is screening for this. The old interview playbook—headcount managed, title progression, cross-functional influence—still dominates most loops. Verna's Netflix story isn't an isolated anecdote; it's the default filter at most large tech companies, and it's now selecting for exactly the wrong people. The candidates who optimized for climbing the ladder spent their energy on coordination theater: the decks, the alignment meetings, the carefully managed stakeholder maps. They didn't maintain their craft. Verna's own boss at SurveyMonkey gave her an ultimatum when she hit senior manager: no SQL queries, no Jira tickets for six months, or she wouldn't be ready for director. She did it, and at the time it was the right move. Management was the only way to have impact. But that bargain has expired.
The people who stopped writing code, stopped running queries, stopped designing pages—they're now facing a market that wants builder skills, and they don't have them anymore.
LinkedIn just scrapped its Associate Product Manager program and replaced it with an Associate Product Builder track. The signal is in the name. The company also introduced a Full Stack Builder career ladder running parallel to the traditional PM track, and the interview process now reflects the shift. Jaclyn Konzelmann, an AI PM director at Google, published the five questions she asks every AI PM candidate: they test system design, model selection, agent orchestration, and failure-mode handling. Not feature lists. Not stakeholder management. System design.
Meanwhile, Ankit Virmani, who recently closed offers at Uber, Cisco, and Atlassian, reports that seventy to eighty percent of his interview loops were still traditional behavioral and product sense—but the one round that showed up at every top AI company was AI system design, and that round decided his level placement, not just pass/fail. The companies that have adopted this filter are hiring faster and placing candidates more accurately. The companies still running the old playbook are rejecting high-impact ICs because those candidates can't answer "how many people did you manage?" with a number that clears the threshold.
Elena Verna's cowboys-and-farmers hiring grid captures the underlying talent taxonomy, and the grid's most revealing feature is what it de-emphasizes. Pincus maps candidates along two axes: cowboys versus farmers, and old guard versus new guard. Cowboys push boundaries; farmers scale operations. Old guard brings scar tissue and business instinct; new guard brings AI-native fluency and the confidence to ignore legacy constraints. The grid has four quadrants, but Pincus explicitly fades two of them: young farmers and old cowboys. Young farmers lack the context to build scalable infrastructure—operational work requires experience, and moving fast in ops breaks things you can't easily fix. Old cowboys lack AI-native instincts, and while they can learn, they're not literally native to the tooling.
The high-impact IC role lives in the other two quadrants: old-guard farmers who understand how to build systems that scale, and new-guard cowboys who know how to ship with AI tooling. Both can execute end-to-end. Both compound skills faster than they accumulate reports.
The structural risk is that middle management is already obsolete in companies moving at this velocity, but most hiring teams haven't updated their filters. Companies with lean teams reportedly now ship at output levels previously requiring much larger headcounts, and the traditional manager role—filtering information up and down, coordinating handoffs, turning strategy into tasks—has become low-leverage. The coordination cost has collapsed. Decisions are faster because they're cheaper to reverse; you can prototype, test, and pivot in a day instead of waiting for the next quarterly planning cycle. The employees who need that level of oversight are the wrong employees. The ones who don't need it don't need a manager whose primary job is supervision. They need context, access to information, and autonomy to act.
Companies that haven't figured this out are still hiring for the old ladder, and the mismatch is acute. They're filtering out the Elena Vernas—the people who can ship a pricing page solo—and hiring people who know how to run alignment meetings but can't execute without a team.
The trust piece explains why command-and-control cultures can't unlock this velocity even when they hire the right people, but the hiring failure happens upstream. If your interview loop still asks "how many people are you managing?" as the primary signal, you're not even getting the right candidates into the room. The candidates who maintained their craft, who kept writing SQL queries despite their boss's warning, who can now build an entire growth campaign with Claude and ship it in a week—they're getting disqualified in the phone screen, and they're going to the companies that redesigned the loop.
The companies that win the next three years are the ones rewriting job descriptions and interview rubrics right now. They're testing for end-to-end execution: can you design the system, write the prompt, evaluate the output, instrument the metrics, and ship the feature? They're asking candidates to vibe-code during the interview, not to prove they can write production-grade code, but to prove they can think in systems and translate product instinct into working prototypes. They're looking at candidates who spent the last five years as directors and asking not "how many reports did you have?" but "what did you ship, yourself, in the last six months?" The candidates who can't answer that question are getting filtered out, and the ones who can are getting offers with comp premiums that reflect the scarcity.
Mahesh Yadav's trajectory—Microsoft at $120K to Google at $1.3–1.4M in under a decade, with Mahesh Yadav reports friends at Nvidia clearing $2–2.5M—isn't typical, but it's real, and it's being paid specifically for builder skills that let companies move faster with fewer people.
The career advice that follows is almost too obvious to state, but the evidence suggests most people haven't internalized it: skill compounding beats title chasing, but only if you can prove the skills in an interview designed to test them. If you're a senior PM or director and you haven't shipped something yourself—coded a prototype, built a dashboard, designed and deployed a page—in the last six months, you're not ready for the market that's coming. If you're a hiring manager still using headcount as a primary signal, you're losing the candidates who can actually move your metrics, and you're hiring people who will need two years to learn skills they should have walked in with. The ladder is dying. The people climbing back down are the ones getting paid, and the ones still climbing up are discovering the rungs no longer lead anywhere.