# Summary: This CPO regrets that product management exists — Tom Verrilli (CPO of Whatnot)

- Source: Lenny's Podcast
- Video: https://www.youtube.com/watch?v=ruvis-VWg2s
- Companion to: `source-docs/youtube/youtube-ruvis-VWg2s.md`
- Uploaded: 2026-08-02

Tom Verrilli (CPO of Whatnot, ex-CPO of Twitch, ex-director of product growth at Twitter) argues that product management as a specialist role is **overused** — Whatnot's product team is built on the premise "we regret that product management exists."

**Core thesis:** PM shouldn't be a default hire for every team. PMs infantilize engineers and designers who are perfectly capable of making good decisions but never had to because a PM was always there to babysit them. PM work is a *trade/muscle* — the more you abstract engineers/designers from it, the more their muscle atrophies. So PMs should be hired only for specific needs and mapped to *problems/projects*, not teams.

**At Whatnot this looks like:**
- Only ~21–22 PMs (small for their GMV), loosely grouped into buyer/seller/trust-and-risk, frequently reassigned every 6-month planning cycle
- Leaders spend 90%+ of time on **IC work** (Tom ~50%); managers get promoted *back into doing things*, not out of it
- Of 31,832 PM applicants in 2 years, they hired one

**What they hire for now:** trending *down* — candidates who talk about alignment/stakeholder management (i.e. their specialty was "politics"). Trending *up* — **systems thinking** (macro + micro together), impatience to validate quickly, specificity about things they actually built, being a decision-maker/agent of change rather than a babysitter.

**AI's role:** Tom doesn't think the shift is explicitly AI-driven, but AI makes it far easier — data science (Hex Threads), querying the codebase directly via Claude instead of bothering engineers, and real-time live-product debugging. But it's also a broader cultural shift (top-down is good when leaders are genuinely in the weeds; "know then go"; "whole-ass few things").

**On IC work & comp:** Moving back to IC as a VP is *easier* comp-wise — 3 highly-paid senior ICs can beat a big org tree of L5/L7s. His greatest desire is that Whatnot's product bench looks like the ex-CTOs who became plain engineers at Anthropic.

**Advice for PMs job-hunting:** start doing IC work in your current role, stay in the weeds, understand the customer/business/tech and translate all three. Core PM skills (identify what to build, distill requirements, prioritize, understand strategy) are the most durable — but most PMs have spent years rewarded for "product theater" (storytelling/alignment) rather than these skills. There's nowhere to hide from that anymore (a hands-on case study exposes it fast).

**Other memorable takes:**
- **"Play the accordion"** — constantly zoom out to understand knock-on effects, then zoom back into V1; the value is in the iteration loop, not the roadmap OR pure spaghetti-throwing
- **On live/agentic commerce:** agentic AI will be huge for high-intent/programmatic purchases, but most retail is *low-intent* and social — e-commerce never exceeded ~20% of US retail in 30 years, and live commerce is the first format to combine internet scale with the social/cultural experience of physical shopping (buying a live-streamed lobster, shipped overnight)
- **From Twitter:** "most of the time you hear it's really complex, it isn't — leadership's just weak" (9 heads of product in 2 years); and that true product-market fit is felt emotionally, not seen in graphs
- **Fail-corner lesson:** "averages mean nothing to the individual" — a feature with 3% adoption may be 100% of what a core user group does; deprecating it on averages can blow up real businesses
- **Recommended:** *The Hard Thing About Hard Things*, *The Purpose Driven Church* (Rick Warren — engineering emotional investment), *Babel* (R.F. Kuang); motto: "I'll figure it out"

**Bottom line:** fewer, more senior, more autonomous PMs doing real IC work — right for Whatnot, not necessarily everyone, but going back to "actual product work and getting out of the theater" is probably true everywhere.