Why this is in the vault
Series premiere ("The CFO Operating System," Part I of V) names two concepts — altitude control (choosing the right zoom level: aerial scan vs. on-the-ground deep dive) and attention quality (System-1 scan review vs. System-2 deep review) — that describe, almost exactly, the triage discipline an AI COO needs for its own task queue.
Mapping against Ray Data Co
The author's "scan vs. deep review" split is the same distinction the /process-newsletter triage already encodes: cheapest-test-first, content-triage-on-subject-line-alone-where-obvious (a scan) before fanning out a sub-agent for full extraction (a deep review) — and his warning that "constantly switching altitude and attention quality is wildly inefficient" is a direct argument for why that skill's triage order exists (don't deep-read everything, don't skip triage either). His broader frame — that a role's complexity eventually breaks an organic, ad hoc system and demands a deliberate toolkit for directing attention, not just more effort — maps onto AGENTS.md hard rule 4 (route long artifacts through subagents; context rot is real) and onto why Ray runs on skills + state files + sub-agent fan-out rather than one long-lived context trying to hold everything. His "what should live in your head, and what absolutely shouldn't" (previewed for Week 4) is the same question the vault's memory-vs-working-context split already answers operationally; worth re-checking that split against his framework once Week 4 lands. No immediate action — flagging the series to follow, since Weeks 2-5 (four operating modes, core rhythm, information triage, a worked personal-calendar example) will likely sharpen this further.
The core argument
A new Group CFO role outgrew the author's decade-old personal system (a one-page A3 priority sheet) because breadth, numeric-context load, and project count exceeded what any single static tool could hold. He surveys how different senior executives solve this (lever-arch files, daily-takeaway notebooks, hyper-automated digital stacks) to establish there's no universal answer, only universal need. He introduces two operating concepts the series will return to: altitude control (a CFO must deliberately choose zoom level — aerial scan vs. ground-level deep dive — rather than defaulting to one) and attention quality (Kahneman's System 1/System 2 mapped onto "scan reviews" vs. "deep reviews" of other people's work). He flags AI as the most promising lever for personal-operating-system transformation specifically (more than for dashboard automation), conditional on codifying how you actually work first. Previews a 5-week series: operating modes ("Four Es"), core rhythm, information triage, and a worked personal-calendar example.
⚠️ Sponsorship
Sponsor: Campfire (ERP/close software) — the long-documented, openly-disclosed relationship (author is an investor/user; first flagged 2026-05-11, recurs across most Playbook issues). Not a new sponsor; the rotating third-party pool (Zip, Pulley, Stuut, Una, Ledge, Summation, Nominal, CloudZero, Aleph) did not appear this issue. One clean sponsor block plus the standing "sponsor this newsletter" footer CTA — no other promotional content.
Related
- [[2026-02-23-every-chatgpt-memory-context-rot]]
- [[2026-05-06-osmani-cognitive-surrender]]
- [[2026-09-05-cfo-secrets-inheriting-shitshow-finance-function-i]]
- [[feedback_targeting_system_prioritization_filter]]