The Skip

The PM career framework for AI

PM intuition built at FAANG is going stale because AI collapses the headcount-scaling logic that created the PM function in the first place. One founder hit $50M ARR in 18 months before hiring a first PM; five years ago that would have meant a team of ten. Companies now want one of two things: transformation expertise (legacy enterprises needing someone to be their AI strategy) or execution speed (AI-native firms like Anthropic or Sierra that already have vision and need someone to ship 10x faster). Misreading which bucket you're entering is the most common move failure.

Three constraints narrow which door is actually open: compensation (FAANG pay is 2–3x market, so early-stage equity is often unaffordable), location (80% of AI-native roles require Bay Area), and pace (9-9-6 vs "online after hours"). Layered on that is a Builder-vs-Manager identity split — the comfortable middle is disappearing. Wannabe Builders taking prompt-engineering courses don't cross over; real Builders ship side projects compulsively. Map constraints + identity honestly and the right door becomes obvious.

The product skill you must now master: Reinvention

The 20-year PM playbook has inverted: moving up and away from building used to be the goal; the people in demand now are hands-on, opinionated, high-agency builders. Those who climbed by mastering coordination, stakeholder alignment, and facilitation — the "Shadows of Superpowers" A-students of the old game — are most exposed, because AI is absorbing exactly that work.

Five patterns from real career questions: don't wait for dust to settle (reinvention is now a permanent baseline, the hill only steepens); build with executive presence, not just technical depth; mourn the lost role briefly, then choose upgrade over nostalgia deliberately; treat non-standard backgrounds as wedges — function depth plus product mindset plus AI fluency (the "sales builder," "HR builder") is currently the rarest and most-sought combination; and stop deliberating, take a bridge role to let current work recalibrate your view. On ageism: it's real, but the actual variable is appetite for reinvention, not age.

The PM Career Framework for AI (Part 2): The Large Tech Playbook

Big Tech golden handcuffs are real and usually worth keeping — but at a price: your brand erodes while compensation inflates. If you're within a few years of a financial milestone, harvest deliberately and set an exit date. A decade-long IC at big tech can still jump to a hot AI startup if interviews are going well; the "step down" feeling is the old playbook. APM programs are a genuine early-career on-ramp but teach inside-the-building PM, not the builder instincts AI-native companies want — plan your exit accordingly.

Public enterprise tech (Salesforce, Oracle, Adobe) is a weaker version of Big Tech: easier to enter, more management room, but lower talent density and a playbook that transfers poorly. Only move there from an ex-growth startup, not from a company with real secondary liquidity.

Private late-stage companies — Stripe, Databricks, Rippling — are the quality middle: strong compensation, near-liquid equity, high talent density, and a bias to ship. Competitive to enter and often require in-person presence, but the best pocket in the market for experienced PMs who can clear the bar. Recent IPOs offer the same profile with fully liquid, more volatile stock.

Part III covers AI labs, hot AI startups, ex-growth mid-stage, and early-stage founding roles.

The PM Career Framework for AI (Part 3): AI Labs to Founding

AI labs (OpenAI, Anthropic) don't want AI experts — they want builders who can productize research. IC hands-on credibility is the entry price; middle management doesn't transfer. Bay Area presence is mandatory. Wide funnels signal open-mindedness, not low bars — getting through is the hard part. Pace resembles Meta more than hot startups.

Hot AI startups (Sierra, Decagon, Harvey, Glean) move harder. Ten well-funded competitors race in every vertical; founders work six days and expect the same — 9-9-6 is real. The pace filters out people with young children or aging parents, skewing teams younger and more male until the company matures. Joining early gives experienced PMs outsized leverage since PM is often a newly installed function.

Ex-growth mid-stage is the easiest door to open but the weakest in trajectory. At year seven in a flat-growth company, story value is marginal and ZIRP-era equity rarely pays out. If better options exist, the trigger to leave has been pulled. If it's the best available after a real search, take it — being in motion beats erosion.

Founding is an emotional decision, not a pragmatic one. Founders can't imagine doing anything else — the other doors feel closed. Spreadsheets on hourly rates and success odds signal the wrong wiring. Ex-founders are valued afterward for wide scope and forced fluency in fundraising, but only if the pull is genuine.

The Post-IPO PM Playbook Is Being Rewritten

Recently public companies face a PM challenge distinct from AI-first startups and big tech: adapting a running system under public-market accountability. Three CPOs — Dheerja Kaur (Hims & Hers), Anneka Gupta (Rubrik), Yuhki Yamashita (Figma) — converged on five findings.

Going public doesn't require swapping product-minded PMs for operators. Robinhood's GM model and Rubrik's five years of product QBRs both showed the existing team can absorb business accountability if you change their objective function, not their headcount.

AI removes the engineering bottleneck and therefore expands PM workload. At Rubrik, bugs now resolve in days instead of two-week cycles — more decisions, more surface area, more output to review. Pure people-management diffuses; every PM becomes a player-coach directing agents and humans alike.

At scale, the scarce skill is systems thinking, not speed. Figma needs PMs who simplify as they add; Anneka's proposed interview: live pair-coding with Claude to test whether candidates sequence problems and ask the right questions.

Planning compressed, not died. Figma dropped annual planning; Hims & Hers teams build while deciding. Quarterly accountability remains — the transformation is tighter cycles, not no cycles.

AI adoption stalls unless the CPO goes first, visibly and imperfectly. The unlock is the joy of building, not mandates.

The Promotion Mistakes That Derail PM Careers

Most PM promotion failures are mindset failures. Post-ZIRP, orgs have fewer senior seats and deliberately slow promotions — at Meta everyone felt promoted 6–12 months late; that's a feature. Treating a constrained org as hostile clouds judgment; the reframe is "constrained, not rigged."

The competitive frame is equally damaging: racing peers makes delays feel like losses. Early promotions come fast because each level is a bigger version of the same job; leadership requires different skills — executive presence, indirect influence, zooming out — that take years to build. Three years without promotion at director is the expected difficulty curve.

When feedback accompanies a denial, don't dismiss it. Managers see execution; skips observe cross-org presence in rooms managers never enter. If your manager never pushed you, you were coasting.

Own your growth regardless of manager expectations. Ask whether you want the next level — VP at a large company is mostly alignment meetings; staying close to craft at a smaller firm may suit you better.

5 Career Questions Your Old Playbook Can't Answer

The advice that made you a senior leader is often exactly wrong for the hardest questions you face from that seat.

When a manager's top bet becomes unmanageable, two superpowers hit their shadows: the manager keeps coaching; the performer keeps bulldozing — now at you. Naming that collision is 80% of the fix. If they can't accept you as manager rather than sponsor, graduation is the move.

Senior product execs fielding 2-3x comp frame it as AI-bubble risk: take the safety. Elite builders likely get more valuable as tools improve. Big-company harvest roles pay above market to navigate internal friction — which matters less to future employers than building something new. Real question: do you want the mercenary role, or just permission for the paycheck?

Three short stints in four years is a pattern, not the market. Stop reaching for the frontier. Take something you can dominate — lesser brand, later stage — one win resets the read; a fourth short stint cements a gap story.

Optimizing for a great manager is early-career advice. Management quality declines as you climb; most executives got there on other skills. Better question: what environment requires least pretzeling? Find 3–5 non-negotiables.

Managing someone more senior: the imposter move is treating them as a peer, becoming a cheerleader. Diagnostic: are you safe giving real critical feedback? Your advantage is company context and tenure — exactly what an overqualified new hire needs first.

The Burnout Paradox in Tech

Tech burnout is self-inflicted by success: Nikhyl Singhal (coach, The Skip) argues high-performing companies systematically identify their best people and pile on more scope, so the reward for solving problems is more problems. One PM he coached had been "surging for just another few months" for five consecutive years. The resentment trap follows: unlimited personal investment into a structurally constrained employer (limited promotions, flat compensation relative to expanded scope) breeds bitterness, not recognition. Startups aren't inherently worse than large companies — bureaucracy burns too. Tactical fixes: color-code your calendar red/yellow/green to surface stress patterns, protect a solo hobby as a "single-player game," and think in decades so a maternity leave or slowdown period doesn't register as permanent loss. The mindset shift: work expands to fill the space you give it, so define that space first and build a career inside it rather than sacrificing present life against a future payoff that never arrives — what Singhal calls "downstream future regret."

What PM Hiring Managers Actually Screen For

The old PM job was picking the best thing to build; the new one is expanding what the team can deliver. Hiring managers at Netflix (Mckenzie Lock), Rippling (Sarah Koo), and EvenUp (Sam Stone) screen for momentum across parallel workstreams, not triage skill. For case studies, structure is table stakes — AI makes polished decks trivial. Differentiators are pre-solution user research (shockingly rare) and willingness to update assumptions under challenge. Pedigree: trajectory outweighs brand names; domain expertise has actively burned EvenUp. Referrals are binary — a genuine voucher is a pre-interview reference check; cold applications are treated equally, and a direct email with a real product opinion skips to round two. Borderline decisions tip on energy and curiosity.

Three Job Searches, Three AI Roles: What Actually Worked

AI companies hire for complement skills, not AI credentials. Ben Dreier (Netflix creative technology), Julia Roberts (OpenAI growth), and Janie Lee (Abridge VP Product) all lacked AI backgrounds; each filled a product skill gap the company had. Julia's growth expertise mattered more to OpenAI than AI depth. Interviews have shifted to live demonstration — over 90% of Julia's process was real-time problem-solving. Janie built a prototype and commercialization plan over a weekend on an unfamiliar healthcare topic; now reviewing 120+ candidates at Abridge, she says depth behind a prototype is what separates.

Julia moved from Pinterest director to OpenAI IC — a role type change, not a step back. Her search stalled for months targeting a larger version of what she had; asking what she wanted differently unlocked it. Ben's offer traces to a cold application seven years earlier — no hire, but a relationship maintained until it opened a door.

10 Job-Search Rules That Just Broke

The senior job-search playbook has broken in at least ten ways, per post-mortems with Dana Ingraham (Harvey), Briana Ings (Atlassian/Loom), and Pei-Chin Wang (now founding).

AI tools have flipped onboarding from headwind to tailwind: connecting agents to a company's Slack and docs on day one collapses months of ramp to days. Founding is fast (validation in a quarter), fun, and skill-additive even if it fails. Remote cuts 70–80% of frontier-lab pipelines; hot mature companies like Atlassian that grant remote exist but require search effort. Filtering on "AI company" is unnecessary — growing established companies are heading there.

Credentials have lost ground to artifacts: Pei-Chin's CPO title didn't open doors; shipped work did. Dana opens cold outreach only when it includes an interactive prototype. Interviews probe hands-on AI fluency, not theory. Take any interview at a desired company — level recovers, company quality doesn't. Non-traditional resume: teach the interviewer how your work maps to their problems. Build real things instead of taking classes; fluency compounds in use. Signal family-time constraints; watch whether a boss internalizes them. Professional identity is already being reformatted — all three found "smiling exhaustion" on the other side.

Six steps to effectively tell your career story

Poor storytelling holds back even strong careers: a less-qualified but more politically savvy person advances because they narrate their work compellingly. The antidote is a six-step framework, illustrated through a fictitious Gmail PM role rebuilt from a dense, jargon-heavy monologue into a two-minute conversational story.

Step 1, project context: open with market conditions, not what you built — mobile was threatening Gmail's web dominance. Step 2, role context: establish your footprint — "I was the only PM for 18 months" signals scope without a detail rabbit hole; spend no more than 45 seconds on setup. Step 3, the challenge: articulate the hard problem before presenting the solution, or the work looks trivial. Step 4, what you built and why: use phases instead of a punch list; invite dialogue mid-story. Step 5, what worked and what didn't: opinions are the payoff — hiring turns on judgment, and sharing what you'd do differently shows wisdom, not weakness. Step 6, conclusion: tie the exit to a career theme.

Facts should "travel well" — short enough that the listener can retell them.

Beyond the resume: master the art of telling your career story

Most professionals fail their own 30-second career pitch — and the fix is structural, not cosmetic. Nikhyl Singhal (founding CEO of two companies acquired by IBM and Google; later lead PM for Hangouts and Google Photos; then Credit Karma and Meta) argues the pitch should work like a movie trailer: open with a career theme, not the beginning of your resume, then deliver two or three pivotal highlights that make the listener want to know more.

Five failure modes recur: chronological recaps that start too early, grandiose abstractions without achievements, project descriptions that omit stakes, jargon that alienates non-specialists, and blurring individual from team contributions. A recruiter spends 7.4 seconds on a resume — the spoken version faces the same compression.

The structural fix: lead with a one-sentence theme, anchor achievements to recognizable names and real numbers (LinkedIn JobSeeker Premium from $30M to $100M revenue), and plant "conversation hooks" — references that invite follow-up. Singhal's own revised pitch demonstrates: founding CEO exits to IBM and Google acquisitions, then Hangouts, Photos, Credit Karma's "money button," Facebook Video and Groups — each line is a door, not a summary.

How to find a mentor and build a rewarding relationship

Companies optimize for their own growth, not yours — outsourcing career development to a manager or internal program is a structural mistake. Build a "board of directors" instead: former managers (candid, knows your work), dotted-line managers (company context without direct politics), peers trading wins and struggles, and friends who care. Cold outreach fails; advisors invest in people they have reason to help. Thirty minutes per week compounds to 500 conversations over a decade — the engine of professional luck.

When no network exists, free content on LinkedIn, YouTube, and podcasts substitutes; a small professional community adds real-time exchange.

For senior leaders, paid resources earn their cost. Soft-skill gaps and emotional blind spots — burnout resets, cultural mismatch, identity rewiring — require trained expertise. Executive coaches bring structured accountability and pattern-matched solutions; therapists handle deeper heart-versus-head conflicts.

To be coached well: send a focused agenda, name one specific gap, process contradictory signals from your whole board rather than following any single voice, and summarize takeaways after each meeting. Mentoring others sharpens the same skills — it is the fastest way to improve at being coached.

Red flags: coaches who failed as operators, contrarians pushing risks they never took, and advisors recycling what worked for someone else. The non-negotiable: the coach listens long enough to understand your situation before prescribing.

How to Navigate Org Drama

Org drama has intensified not because people got worse at politics, but because flattening, structural layoffs, and AI-driven restructuring concentrate friction on ICs and frontline managers whose overwhelmed bosses can no longer advocate for them.

The default stay/leave rule: stay if the role keeps you current with how modern teams build; leave if it lets you fall behind. High comp can justify staying, but only if you use that time to remain current. Prestigious-but-static companies are becoming resume liabilities.

Don't put people on the spot until you've earned the right — directness without leverage just makes managers close down. Let work create leverage before raising ambitions or going to a skip-level.

During uncertainty, delivering quietly tips 50/50 decisions. Build cross-functional relationships now, not during the crisis.

If the promotion contract breaks and burnout follows, take 10–15% off the throttle and protect what matters outside work before the breaking point arrives.

Hiring Your First Head of Product

Most first Head of Product hires fail not because founders picked the wrong person, but because they got timing, role definition, and profile wrong before the search.

The inflection point arrives after product-market fit, when operational details multiply faster than founders can handle while building the company. Before that point a product leader distances founders from market signals; after it, resisting the hire constrains momentum.

Five failure modes dominate. Delegating strategy too quickly: the first Head of Product should own execution and grow into a strategic partner — founders who hand over the roadmap on day one inevitably claw it back. Prioritizing resume over readiness: FAANG-pedigreed executives built for layered orgs often can't thrive doing daily customer interviews and writing specs. Leaving the scaling-vs-expansion question open: most early-stage companies need someone deepening the core product, not incubating new lines. Hiring a change agent before the org is ready: if founders privately want the status quo, a process-builder hits a wall. And mistiming in either direction.

The job-spec process is itself the alignment tool. Working through "Pixels, Process, People" (product gaps, operational bottlenecks, leadership voids) forces agreement on what the role is before candidates hear conflicting answers from interviewers. Hire for the current 12-month problem; candidates who outgrow the role signal success.

The strongest candidates come through investor and community networks, not postings. CEO-led references reveal problem-solving patterns formal interviews miss.