"The Great Reshuffling" — @CJ Gustafson

Why this is in the vault

Field report from CJ's SF week (25 CFO dinners, 50 finance-pro beers, 57 banker/investor coffees) on the cross-cohort angst rippling through early-to-mid-career SaaS finance leaders as the SaaS-to-AI transition reshuffles the wealth distribution. The dispatch hint framed this as a "State of F&A Talent" data report; it isn't — it's a qualitative, anecdote-driven essay. But it IS load-bearing as a sentiment-anchor for a labor-market regime shift the founder is watching across multiple theses (investing capital-cycle, Sanity Check audience, phData enterprise-buyer mapping).

The most useful claim is non-consensus: in 2026 the mispriced career lever is experience scope, not title or equity at the hot AI co. Counter-trade to what 90% of CJ's network is currently doing.

This is also a self-promo issue for CJ's recruiting arm (Mostly Talent) rather than a paid third-party sponsor placement — relevant for tracking the Mostly Metrics business model evolution (he's monetizing the audience directly now, not just renting it to Brex/Abacum/Samsara).

The core argument

CJ frames a generational reshuffling event among elder-millennial SaaS finance leaders:

The wealth bifurcation — roughly 10,000 tech people have hit $20M+ retirement-wealth events in the last 5 years (Anthropic's recent secondary capped per-person at $30M; 75 people hit the cap). A much larger cohort of SaaS operators in their early 40s thought they'd hit the lottery and didn't. They have "great jobs, well paid" but not F-you money — and they're comparing themselves to peers 5 years older who joined HubSpot pre-IPO.

Three anecdotes from the SF trip:

1. $50M ARR SaaS CFO considering jumping to an AI co rather than playing for exit. Six years in, board meeting about hiring a banker, exit valuation now closer to $300M than the $1B he expected 2 years ago. (Implicit: who is buying a $50M ARR SaaS co right now?)
2. Cybersecurity banker now doing more defense deals than SaaS-security deals, dragged there by his CFO network. Defense tech = more near-term IPO candidates than cyber. Adjacency play.
3. Investor whose IC won't fund anything with "SaaS" in the name. AI, Defense, or Hardware/Compute only. Investment-brand positioning, not necessarily unit-economic conviction.

The three career-comp levers (CJ's framework, drawn from his own CFO career):

• Maximize W-2 cash comp (do this at a steady-eddy 30%-growth, $500M+ revenue co — late-stage startup or recently-IPO'd-but-down-big)
• Maximize total equity (do this at a Series B/C with real conviction; give back salary if you have to)
• Maximize title + experience scope (take a CFO seat at a smaller / less-shiny shop; get the keys to run around)

People are bad at picking which game they're playing — they spray and pray across all three and maximize none.

The non-consensus claim — where the arbitrage lives in 2026:

CJ identifies two paths and clearly favors the second:

• Path 1 — Work at a fast-growing AI co with real technical advantage and weak finance/GTM DNA. "Equivalent of where you went to college." Higher cash AND equity than SaaS historically paid. But expectations are 6-days/9-hours, AI-pilled, frontier-eats-people risk.
• Path 2 — Be the #2 at a company that's lost its sheen. The "broken toy in a flyover state" trade. Craig Conti (Verra Mobility CFO): "Don't go work at that hot AI company. Go work at that paper mill that's going out of business." Bruno Aniq (Wellhub CFO) did this at AOL post-glory-days — "the cockroach of the internet" — and it accelerated his path to first CFO seat. John Malone at TCI = canonical example.

CJ's punchline: "Nobody is running their career math on experience. That's the lever that's currently mispriced."

What this is NOT

• NOT a labor-data report. No proprietary survey, no scraped JD analysis, no salary benchmarking against the F&A talent market. The dispatch hint about "State of F&A Talent (May)" appears to be Gmail-snippet inference, not a section of the body.
• NOT an AI-replacing-F&A-labor quantification. The piece names the angst — Q1: "Is AI going to diminish (or take) my job?", Q2: "Can I learn it fast enough to be valuable?", Q3: "If I do, will the value accrue to me or the company?" — but provides zero data on actual displacement. This is sentiment-color, not labor-stats.
• NOT sponsored. No Brex/Intuit/Samsara/Rivian/MLB/Abacum third-party ad. The top-of-email block and bottom CTA both promote CJ's own Mostly Talent recruiting arm (mostlytalent.com, typeform hiring/candidate intake forms). Reader pool stated as 75,000.

Mapping against Ray Data Co

Medium mapping, with one strong thread:

• Sanity Check editorial angle (STRONG) — The mispriced-experience-lever framing is exactly the non-consensus territory Sanity Check should own. The "everyone is running comp math; nobody is running experience math" frame can be lifted as a re-frame for a data-team-leader piece: "In 2026, your career math should weight scope-of-decision more than title prestige." This works because (a) it's counter to the LinkedIn-crush instinct, (b) it's testable against the founder's own data-team-leader audience (where the analogue is "VP-of-data at a hot AI co vs head-of-data with real autonomy at a Series C"), (c) it lands without restating CJ (per [[feedback_no_derivative_sanity_check_pieces]] — re-frame for data leaders, not finance leaders). Founder-actionable: candidate angle for the Sanity Check research brief queue.

• Investing thesis — SaaS-to-AI capital migration (MEDIUM) — Anecdote #3 (investor IC won't fund anything with "SaaS" in the name) is sentiment-corroboration for the SaaS-multiple-compression / AI-rotation we've been tracking. Anecdote #2 (cybersecurity-to-defense-tech adjacency, banker network following CFO network) is a useful capital-flow datapoint for the defense-tech adjacent reads. Not strong enough alone to anchor a thesis update, but worth filing against [[01-projects/investing/markov-capital-cycle/]] as crowd-positioning evidence for "investors actively building anti-SaaS brand positioning."

• phData enterprise-buyer mapping (MEDIUM-WEAK) — F&A leaders ARE budget signers for data-marketplace plays; their current angst about own-displacement may make them more receptive to "tools that help me look indispensable" pitches (Brex/Abacum-style). Worth filing as buyer-psychology context for the upcoming phData start (2026-05-26 per project memory), even if it doesn't direct any specific action this week.

• Founder-personal — small business operator (WEAK) — As a one-person operating biz, the comp-lever framework doesn't apply, but the meta-claim (experience scope > title) maps to the founder's own decision-architecture: he's already optimizing for scope at RDCO over title elsewhere. Useful as a sentiment-anchor that his structural bet is aligned with where CJ's network thinks the arbitrage is.

No DECISION-tier item. This is anchor-sentiment evidence, not a thesis-shifter or a paper-trade trigger.

Hard data points

• ~10,000 people in tech hit $20M+ retirement wealth in the last 5 years (CJ's estimate; not sourced)
• Anthropic's recent secondary capped at $30M per person; 75 people hit the cap
• 75,000 Mostly Metrics readers
• CJ's three personal CFO data points: Veeam (max cash), Snyk (max equity, took salary cut), PartsTech (max title + scope)
• Exit-valuation regime shift anecdote: $50M ARR SaaS co exits "closer to $300M than the $1B" expected 2 years ago

Methodology note

Pure CJ-network qualitative. Conversations across one SF week with ~125 finance/banker/investor contacts. No proprietary data, no panel, no scraped sources. CJ's own caveat: "anecdotes are anecdotes… a point in time, not a dataset. But there's a lotta truth in anecdotes." File accordingly — useful as sentiment-color and cohort-positioning evidence, not as base-rate data.

Related

• [[06-reference/2026-05-14-mostly-metrics-q1-2026-startup-benchmarks]] — companion Mostly Metrics piece on the macro state of SaaS benchmarks; reinforces "$50M ARR exits at $300M not $1B" claim with broader multiples context
• [[06-reference/2026-05-24-mostlymetrics-pizza-shop-ai]] — same author, SMB-AI labor-displacement angle; pairs with this piece's elite-tier angst as the bookends of the AI labor reshuffle
• [[06-reference/2026-05-19-mostly-metrics-when-to-allocate-overhead]] — CJ's normal sponsored shape (Brex top-of-email); useful contrast for tracking that this issue is self-promo not third-party-sponsored
• [[06-reference/2026-05-11-backfill-discovery-mostlymetrics]] — Mostly Metrics whitelist context + known sponsor rotation
• [[feedback_no_derivative_sanity_check_pieces]] — discipline note for the Sanity Check re-frame angle above
• [[01-projects/investing/markov-capital-cycle/]] — investing thesis where the "investor brand-positioning against SaaS" anecdote files as crowd-positioning evidence