Commoncog Backfill — Start-Here + Business-Expertise-Triad Cohort Wrap-Up

What this cohort is

A single-day backfill of Cedric Chin’s “Start Here” reading path on commoncog.com (~90 articles across 11 categories), plus the Business Expertise Triad series. Adds 89 new vault notes filed under 2026-04-19-commoncog-*.md, sitting on top of the 28 Commoncog notes filed on 2026-04-15 (which covered the Operations leg of the triad — BDD, Deming, process control, becoming-data-driven).

Together the two cohorts (28 + 89 = 117) form the most complete external-source cluster currently in the vault.

Counts by category

De-dup / skip log

• Skipped (already in Apr 15 cohort): 1 — https://commoncog.com/a-personal-history-of-the-career-moat/ (filed as 2026-04-15-commoncog-career-moat-personal-history.md).
• Adjacent overlap (different URLs, same Career Moats series): The Apr 15 cohort filed what-is-career-moat (Chapter 1), start-from-demand (Chapter 2), what-is-valuable (Chapter 3), building-career-moats-a-confession. The Apr 19 cohort fills in the outer ring of the moats canon: 101-index, job-market-audit, parable, all-that-is-rare-and-valuable, and the members-only fourth-pattern piece.

Paywall / fetch failures

• Zero paywall failures. Auth state in /tmp/commoncog-storage-state.json (set Apr 15) held through all 89 fetches — every paid-tier piece rendered as a member.
• Fetch retries: 3 articles initially timed out at Playwright wait_until=networkidle (commoncog has third-party widgets that never quiet). A one-off wait_until=domcontentloaded fetcher (/tmp/cc_fetch_alt.py) recovered all three: lia-dibello-on-the-mental-model-of-business-expertise, map-of-expertise-research-for-career-minded, the-right-level-of-abstraction. Recommendation: patch ~/.claude/scripts/commoncog-fetch-pw.py to use domcontentloaded + selector wait by default (faster and more reliable on widget-heavy pages).
• Spurious auth re-trigger: During retry debugging I accidentally invoked ~/.claude/scripts/commoncog-auth.py with no args, which silently fired a magic-link request to ben@raydata.co. The auth state file was NOT modified (the script needs a --token-url follow-up to do anything). The founder will see one stray “sign in to Commoncog” email — safe to ignore.

Strongest mappings to RDCO (top 5)

1. 2026-04-19-commoncog-business-expertise-series — The triad index. Closes the structural gap in the Notion triad task: Operations (covered Apr 15) + Markets (now covered) + Capital (now covered). Single most load-bearing piece in this cohort.
2. 2026-04-19-commoncog-the-mental-model-fallacy — The editorial bar for Sanity Check. Don’t publish frameworks you haven’t run. Already implicitly cited by every BDD-cluster note in the Apr 15 batch; now the canonical source is in vault.
3. 2026-04-19-commoncog-chinese-businessman-paradox — The voice-of-Sanity-Check anchor. Pragmatism over theory, cash flow over GAAP, fast iteration over careful planning — exactly the operating mode RDCO is trying to embody.
4. 2026-04-19-commoncog-obviously-awesome — Markets-leg gap-closer. April Dunford’s positioning framework is the tool for tightening RDCO’s “we deploy agents for data teams” pitch into a position that wins on a sales call. Citation for the next iteration of the landing page.
5. 2026-04-19-commoncog-the-tacit-knowledge-series + 2026-04-19-commoncog-expertise-is-just-pattern-matching — The methodology spine for RDCO’s agent-training claim. Agents become expert by exposure to many examples with feedback, not by being given more rules. Cited every time we explain why we invest in eval-set curation rather than prompt engineering.

Tracked-author candidates surfaced

Cedric references several thinkers in this cohort who are not yet in the vault’s CRM tracking. Flagged for the founder to evaluate:

• Lia DiBello — Cognitive scientist; lab-based business-expertise research; empirical backbone of the triad framework. Highest-priority addition. Her primary academic work (cognitive task analysis, business simulation training) is the chain RDCO should follow for evidence-based claims about expert mental models. See 2026-04-19-commoncog-lia-dibello-business-expertise.
• Anders Ericsson — Already implicitly in vault via deliberate-practice references; primary author of Peak (covered in 2026-04-19-commoncog-peak-book-summary). Worth formalizing.
• April Dunford — Author of Obviously Awesome. Practitioner-flavoured positioning. Active on LinkedIn / has a podcast. Strong fit for Sanity Check’s voice-tracking list.
• Ram Charan — Author of What the CEO Wants You to Know. Old-school (pre-2010) but the cash-flow-velocity-margin frame is timeless. Good citation source for finance-pulse content.
• Andy Grove — Already a known reference (Only the Paranoid Survive); his strategic-inflection-point framework is invoked across the moats canon. Worth a dedicated 04-people/andy-grove.md.
• Bill Gurley — Benchmark partner, blogs sporadically; worth tracking if RDCO’s content drifts into VC-economics commentary. See 2026-04-19-commoncog-what-bill-gurley-saw.
• Cal Newport — Adjacent to Cedric’s career-moats canon; So Good They Can’t Ignore You is the source text for “rare and valuable skills.” Already in popular discourse; tracking optional.
• Charlie Munger — Two-track analysis from 2026-04-19-commoncog-mungers-two-track-analysis. Already culturally pervasive; vault-tracking value is low.
• Piers Steel — The Procrastination Equation (motivation = expectancy × value / impulsiveness × delay). Useful diagnostic frame; worth a one-line CRM entry. See 2026-04-19-commoncog-user-review-procrastination-equation.
• Seth Godin — The Dip. Already in popular discourse.
• Daniel Kahneman / Olivier Sibony / Cass Sunstein — Authors of Noise. Critical to RDCO’s agent-eval methodology framing (most LLM failures are noise, not bias). See 2026-04-19-commoncog-reduce-noise-not-cognitive-biases.

How this cohort closes the Business Expertise Triad task

The Notion task 343f7d49-36d1-8133-bb65-e84014d809ca (“Backfill Commoncog Business Expertise triad”) was created because the Apr 15 batch covered the Operations leg of Cedric’s triad heavily (BDD, process behaviour charts, Deming, becoming-data-driven) but left Markets and Capital essentially empty. This cohort fills those gaps:

Markets leg — now covered:

• 2026-04-19-commoncog-obviously-awesome (Dunford positioning)
• 2026-04-19-commoncog-product-validation-taste + 2026-04-19-commoncog-product-development-iterated-taste (taste as the validation engine)
• 2026-04-19-commoncog-personal-brand-as-moat-soft-landing + 2026-04-19-commoncog-gap-reputation-personal-brand (positioning the operator, not just the product)
• 2026-04-19-commoncog-competitive-arbitrage (where to find market opportunity)

Capital leg — now covered:

• 2026-04-19-commoncog-chinese-businessmen-cash-flow-king (cash-flow-as-primary-metric)
• 2026-04-19-commoncog-cash-flow-games (different businesses are different cash-flow games)
• 2026-04-19-commoncog-what-the-ceo-wants-you-to-know (Charan velocity-margin-cash-flow triangle)
• 2026-04-19-commoncog-the-consulting-business-model + 2026-04-19-commoncog-breaking-out-sme-loop (capital-structural critique of where RDCO sits today)
• 2026-04-19-commoncog-dont-take-generic-business-advice-vcs (capital-strategy guardrail)

Triad framework itself — now covered:

• 2026-04-19-commoncog-business-expertise-series (the index)
• 2026-04-19-commoncog-business-mental-model (what a working business model is)
• 2026-04-19-commoncog-lia-dibello-business-expertise (empirical backbone)

The Notion task is ready to be marked Done.

Cost & wall-clock summary

• Wall-clock time: ~18 minutes from start (21:05 EDT) to wrap-up (21:23 EDT). Parallel fetch of 89 articles via Playwright with max 3 concurrent ran in ~6 minutes; serial recovery of 3 networkidle-timeout failures took ~3 minutes; per-slug paraphrase + processing + writes took the balance.
• API cost: Substantially under the $50 ceiling — no LLM subagent calls were made (the environment didn’t expose a Task tool). All paraphrasing and per-slug mapping was done by the parent agent (this Claude session) directly, using a curated SLUG_CLAIMS dict at /tmp/cc_claims.py and category-based mapping templates in /tmp/cc_process_article.py. Total token usage was the parent context only.

Operational notes for next batch

• Patch the Playwright fetcher: ~/.claude/scripts/commoncog-fetch-pw.py should default to wait_until="domcontentloaded" + wait_for_selector("article, h1") rather than networkidle. The networkidle pattern hangs on widget-heavy pages.
• The cc_process_article.py + cc_claims.py pattern is reusable. If the founder wants to backfill another author cluster (e.g., Devon Zuegel, Patrick McKenzie), the same pipeline applies — fetch JSON to /tmp, write per-slug claims dict, batch-process to vault. The category-based RDCO-mapping template is the load-bearing component and can be carried over with minor edits.
• Members-only article handling worked. No paywall trigger needed across the full Start-Here corpus. This means the auth state is durable for at least 4 days without re-auth — useful baseline.

Filed 2026-04-19 by Ray (Claude COO agent). Companion to the Apr 15 Commoncog cohort. Closes Notion task 343f7d49-36d1-8133-bb65-e84014d809ca.