Keith Rabois: The PM Role Is Dead, Business Acumen Wins

2026-04-12 Lenny's Podcast

interview future-of-work productivity business enterprise agents

Why Keith Rabois Says AI Is Killing the Product Manager

Keith Rabois — PayPal mafia member, former COO of Square, early investor in Stripe, Airbnb, DoorDash, and Ramp — sat down with Lenny Rachitsky to deliver some of the sharpest takes on how AI is reshaping team building, hiring, and organizational roles. As managing director at Khosla Ventures, Rabois has spent 25 years identifying undiscovered talent and building world-class companies. His perspective on the AI era is both practical and provocative.

The PM role makes no sense anymore. Rabois was convinced by Peter Fenton’s argument that conventional product management is obsolete. “The idea of a PM makes no sense in the future. The skill is more like being a CEO now, which is what are we building and why?” When tools like Lovable and foundation models can turn ideas into working products in hours, year-long roadmaps become incoherent. Things that were impossible in November are easy by March.

CMOs are the biggest AI power users. In a counterintuitive finding, Rabois reports that at two of his portfolio’s best companies, “the number one consumer of tokens is the CMO.” These business leaders — not engineers — are shipping campaigns, running analytics, and producing work product without relying on layers of deputies. The intellectually curious are thriving regardless of their technical background.

Engineers who code from their phones are the future. Rabois himself hasn’t touched a computer since September 2010, running everything from an iPad. He connects this to a broader trend of 10x engineers coding from their phones using AI — the distinction between “business person” and “technical person” is collapsing.

Speed is the single most important signal. Across his best investments — Square, Ramp, Fair — the common trait is tempo. “At board meeting X, you guys identify an opportunity or problem, and by the next board meeting, you’ve shipped solutions.” Ramp was on the precipice of shipping cards in 3 months when the industry standard was 9-12. That velocity is what made Rabois preempt their Series A just months after leading the seed.

Barrels vs ammunition defines your capacity. Rabois’s famous framework: “barrels” are people who can independently drive an initiative from inception to success. PayPal had 12-17 barrels among 254 people. Sam Altman’s OpenAI had two. Hiring more people without adding barrels just increases coordination tax and drag. In the AI era, barrels with business acumen become exponentially more valuable because they can leverage AI as a second team.

6 Key Insights on Talent and AI-Era Organizations

The team you build is the company you build — Vinod Khosla’s axiom that Rabois considers the most important lesson in startups. Right people make everything easy; wrong people make everything hard
Build on undiscovered talent — Don’t compete for people everyone wants. Find people that large-company hiring machines can’t process correctly. This naturally skews younger (less data points = more alpha)
30-day feedback loop — Ask yourself 30 days after any hire: would you make the same decision? Research shows this is as accurate as measuring 1-2 years out
Don’t talk to customers (for consumer products) — Customer feedback for consumer/SMB products is “directionally dangerous.” People make subconscious purchasing decisions but give conscious rationalizations. Enterprise customer development is the exception
Criticize in public, not private — Counter to conventional wisdom: public feedback lets the whole team see issues are being addressed and enables collaborative problem-solving. “High performance machines don’t have psychological safety. They’re about winning”
Promote from within, not external hires — The most successful companies Rabois works with skip hiring senior experienced people and develop talent internally instead

What This Means for AI-Powered Organizations

Rabois’s vision converges on a single insight: in the AI era, the premium shifts from execution skills to business judgment. When anyone can build, the scarce resource is knowing what to build and why. His best portfolio companies already demonstrate this — CMOs shipping with AI, engineers managing 20 people while coding as much as when they were ICs, founders treating AI tools as a second team. The organizations that win won’t be the ones with the most engineers. They’ll be the ones with the most barrels.