Coinbase's 15-to-1 Memo Created a Player-Coach Pool LinkedIn Can't See
Coinbase, Block, and Meta are deleting "pure managers." The displaced cohort has recent IC chops but no title sourcers actually filter for. Here's how to find them.
Brian Armstrong's May 5 memo did something subtler than cut 700 people. It deleted a job category. "Pure managers" are gone at Coinbase, replaced by player-coaches running 15+ reports inside AI-native pods. Block did the same in February. Meta's applied AI org is running at 50-to-1. If you're still sourcing on the title "Engineering Manager," you're filtering past the most interesting cohort on the market right now.
The specific archetype that just got purged
The pattern is consistent across three companies that don't usually agree on anything.
Coinbase flattened to a maximum of five management layers and now requires every leader to carry 15+ direct reports while remaining a hands-on contributor. The org is being reassembled into "AI-native pods," some staffed by a single person spanning engineering, design, PM, compliance, and CS with AI tooling glued underneath.
Block went further. Jack Dorsey cut roughly 4,000 jobs in February 2026, shrinking the company from about 10,000 to 6,000, and published a manifesto with Sequoia's Roelof Botha titled "From Hierarchy to Intelligence." Block's structure now formally names three roles: ICs, Directly Responsible Individuals (DRIs) who own a problem like merchant churn for 90 days, and player-coaches who replace traditional managers but still ship code. Dorsey's stated near-term goal is to compress Block's five management layers to two or three within the year.
Meta's new Applied AI Engineering team, led by Maher Saba and reporting directly to the CTO, runs at a 50-to-1 employee-to-manager ratio. That's double the 25-to-1 figure most org designers treat as the outer limit of span-of-control. Meta has also started retitling some managers as "organization leads."
These aren't three independent stories. Per Axios reporting, the Coinbase memo is being forwarded inside PE and VC portfolios as a template. Sequoia co-authored Block's version. The archetype getting purged, and the one being hired, is now standardized enough to copy-paste.
Why your "Engineering Manager" search misses them
Here's the practical problem. The people losing seats in this purge are senior engineering managers who, defensively, spent the last twelve months re-sharpening IC skills. They've been shipping prototypes in Claude and Codex, writing code on the side, and quietly preparing for a market that no longer rewards pure people management. They are exactly the candidates a founder building a small, dense team should want.
But:
- Coinbase is calling them "leaders of AI-native pods."
- Block is calling them "player-coaches" and "DRIs."
- Meta is calling them "organization leads."
- LinkedIn still has them flagged as "Engineering Manager," "Senior Engineering Manager," or "Head of Engineering," because most haven't updated yet.
- "Open to Work" badges underindex this group hard. Senior managers with options don't broadcast.
A title search for "Engineering Manager" plus "open to work" returns the wrong slice. You get the pure people-managers (the cohort that, per the new playbook, companies are explicitly trying not to hire). You miss the player-coaches who shipped a Kubernetes migration last quarter and have 40 commits in the last 90 days.
The sourcing signal lives in the commit graph
The real signal isn't the title field. It's the gap between the title field and recent GitHub activity. An EM with current IC chops looks like:
- LinkedIn title: Engineering Manager / Senior EM / Head of Platform (still).
- GitHub: meaningful commits in the last 6 to 12 months, often to AI-coding tooling, internal-looking infra, or personal prototype repos.
- Bio or recent post mentions "player-coach," "DRI," "pod lead," "hands-on," or "shipping with Claude/Codex."
- Tenure at a company in the active delayering set (Coinbase, Block, Meta Applied AI, Crypto.com, Bolt, Gemini, MARA, Snap, Pinterest, CrowdStrike, Atlassian, Cognizant Project Leap, Pendo).
Standard ATS sourcing tools can't express that query. That mismatch is exactly why we built Refolk: you describe the person in plain English ("EMs at Coinbase or Block with 15+ Python or Go commits in the last 6 months and a past Staff+ IC stint") and get a ranked shortlist that pulls from GitHub, LinkedIn, and the open web at once.
How big is this pool, actually
Bigger than the headline cuts suggest, because the archetype concentrates in a small number of companies that are now all delayering at once.
Refolk's own index returns roughly 3,914 current US-based Engineering Managers who list both Python and Kubernetes as skills (a reasonable proxy for "EM who can still ship"). Top employers in that cluster: Datadog, Apple, Google, Rippling, HeyGen, Mintlify. Geographic concentration: Bay Area and NYC. Those are exactly the companies most likely to copy the Coinbase template next quarter.
Layer in the macro. Gallup found the average manager's team size rose from 10.9 in 2024 to 12.1 in 2025, a nearly 50% increase since they started measuring in 2013, with a two-percentage-point jump in teams of 25+ driving most of the change. Kalshi traders give 92% odds that 2026 tech layoffs exceed 447,000. BLS already counted 178,000 information-sector layoffs through March.
Most of those aren't player-coach candidates. But the slice that is, EMs at delayering companies with recent commits, is structurally underpriced because nobody's filter catches them.
The contrarian case for hiring them now
Three reasons this cohort is undervalued, beyond the title-filter problem.
1. Most of them aren't being cut for capability
Sam Altman has publicly warned about "AI washing," companies blaming unrelated layoffs on AI. Boston College's Aleksandar Tomic told Fortune some CEOs use AI restructures to spin cost cuts as forward-looking strategy. That means a meaningful share of displaced player-coach candidates are perfectly competent EMs who got caught in a span-of-control reshuffle, not performance management. If you can read past the press release, you're shopping at a discount.
2. The AI-replaces-managers thesis is overstated
Current and former Block employees told the Guardian that roughly 95% of AI-generated code changes still require human modification, and AI tools cannot yet lead in regulated areas like banking or money transfers. The org charts have moved faster than the technology. Companies are betting on player-coach productivity that hasn't been proven at scale. If you're a founder who needs someone who can both architect a system and coach two juniors through shipping it, that's a real human, not a pod-of-one.
3. Delayering reverses
Tech-driven flattening happened hard in the 1980s and 90s, then reversed into what researchers later called "the explosion of middle management." There's no reason to think this cycle is different. Founders who pick up displaced player-coaches at 2026 prices will see the market re-bid for them in 18 to 24 months when the first wave of 15-report pods starts producing burnout, attrition, and quietly-restored middle layers.
Companies are betting on player-coach productivity that hasn't been proven at scale. That bet is your hiring opportunity.
A sourcing playbook for the next two quarters
Five concrete moves, in rough priority order.
Stop anchoring on title
Build searches around the gap. Past title contains "Manager," "Head of," or "Director." Recent activity (commits, talks, blog posts, repo creation) within the last 90 to 180 days. Current employer in the delayering set. Bonus signal: the candidate has used "player-coach," "DRI," "pod," "hands-on," or "shipping with AI" in their own writing within the last 12 months.
This is the kind of multi-source query that breaks Boolean tooling and is exactly why Refolk takes plain-English requests across GitHub, LinkedIn, and the open web in a single pass.
Watch the named orgs, not just the company
Maher Saba's Applied AI Engineering org at Meta is a specific, nameable target. Coinbase's AI-native pods are named in the memo. Block's DRI structure is public. Source by the org, the alumni network around the org, and the people who've written about working in it.
Monitor PE and VC portfolios
If Axios is right that the Coinbase memo is being forwarded inside funds, Q3 and Q4 2026 will produce waves of the same archetype from portfolio companies. Sequoia, a16z, Founders Fund, Thrive, and Coinbase Ventures portfolio pages are worth a saved alert. The downstream layoffs will be smaller per company but more numerous, and most sourcers won't catch them because the individual companies aren't household names.
Mine the communities, not just the platforms
Rands Leadership Slack, LeadDev and LeadingEng conferences, Gergely Orosz's Pragmatic Engineer reader base, the Staff+ book community, and active commit history on Anthropic and OpenAI tooling repos are all places where EMs with current IC chops self-identify even when their LinkedIn still says "Engineering Manager." A six-line outreach referencing a specific talk or PR will out-convert ten "saw your profile" InMails.
Move before the title catches up
The window closes the moment these candidates update their LinkedIn headlines to "Player-Coach," "Founding Engineer," or "Independent." Once the label catches up to the practice, every other sourcer's filter starts working again and the arbitrage is gone. Refolk's plain-English search is built for exactly this window, when the signal is in the gap between what someone is called and what they actually do.
FAQ
How do I tell a real player-coach from a manager who's coasting on the title?
Look at commit recency and surface area. A real player-coach has merged PRs in the last 90 days, ideally in repos that aren't their own personal scratchpad, and can speak to specific technical decisions in recent talks or blog posts. Managers using the label as a defensive rebrand usually can't, and their public footprint is conference panels and hiring posts rather than code or design artifacts.
Why don't "Open to Work" badges surface this cohort?
Senior managers with strong networks don't broadcast availability. They get warm intros first. By the time someone with ten years of EM experience flips the badge on, they've already talked to a dozen founders. The badge is a lagging indicator for this cohort, useful for junior and mid-career ICs and almost useless for player-coach-shaped candidates.
Is the Coinbase memo really being forwarded inside PE and VC portfolios?
Axios reported it on June 12, and the structural evidence (Block co-authoring with Sequoia, Meta adopting similar language, Crypto.com and Gemini cuts citing AI restructuring) is consistent with template-level adoption. Treat it as a strong working hypothesis worth pre-positioning for, even if the exact forwarding paths aren't all documented publicly.
What's the right outreach angle for someone whose company hasn't formally cut them yet?
Don't pitch the job. Pitch the org design. EMs at delayering companies are watching their span-of-control double and their IC time get squeezed. A short note that names the specific tension ("I noticed Coinbase is asking pod leads to carry 15+ reports while staying hands-on, curious how that's actually playing out for you") gets replies. A generic "great opportunity" InMail does not.