No public post URL exists for this piece — it is explicitly framed in-email as part of a promotional sequence written for the paid TDG Design Workshop ("these emails will ostensibly be about TDGs, but more broadly about accelerating expertise... even if you're not interested in purchasing a seat"), not a Commoncog essay with its own permalink. Third email in the active Commoncog x Shadowbox sequence following [[2026-08-25-commoncog-why-tactical-decision-games-work]] (the original essay, unsponsored) and [[2026-11-01-commoncog-making-of-common-sense-tdg]] (first Shadowbox-sponsored guest post in this run).
Why this is in the vault
A short but genuinely new frame on an already-tracked thread: TDGs work as "accelerated osmosis of expertise," reframing why apprenticeship is slow and uneven, plus a fresh claim — senior ICs are benefiting disproportionately from AI because judgment is the bottleneck, while juniors struggle — that bears directly on RDCO's agent-deployer thesis and the founder's own career-moat framing.
The core argument
Cedric and Jared open with the sales pitch (seats for John Schmitt's TDG Design Workshop, link repeated twice), then pivot to the actual content. The claimed pattern: over the past year, senior individual contributors (lawyers, engineers, bankers) have benefited disproportionately from AI because they have the judgment to accept/reject/iterate on AI-generated output, while junior people struggle — and some firms are reportedly slowing junior hiring as senior productivity expands. This is stated as early/tentative/anecdotal, not demonstrated.
The explanatory move is new relative to the original TDG essay: apprenticeship is the "ideal model" every company implicitly wants, but it's slow for two linked reasons — (1) picking up everything that makes a senior person good takes a long time, and (2) most people plateau because truly hard, expertise-revealing cases only come around every few months, so you either get lucky with exposure to a strong senior partner on a hard case, or you don't and feel stuck. TDGs compress this by putting people through simulated hard cases every week instead of every few months, with an expert in the room narrating their sensemaking — hence "accelerated osmosis of expertise" rather than the earlier essay's framing (getting novices to see what experts see). No new empirical evidence is offered; this is a sharper restatement of the same mechanism, now with the apprenticeship-failure-mode breakdown as the add.
Mapping against Ray Data Co
The concrete connection is the judgment-gap claim itself, not just the TDG mechanism (already mapped in the predecessor notes). "Senior ICs benefit disproportionately from AI because they have the judgment to triage AI output; juniors don't" is the load-bearing premise behind two things already live in RDCO: the critic/brigade architecture (station-critic, verify-vault-write, verify-strategic-output, verify-dispatch, behavior-critic) exists precisely because raw agent output needs an experienced-judgment gate before it reaches the founder — i.e., RDCO already built the apprenticeship substitute this email describes, for agent outputs rather than junior humans. It also validates the founder's own positioning bet (project_credibility_for_phdata_sales, user_career_commitment_shape): if judgment under AI is the actual scarce resource, then an experienced operator's moat is demonstrating that judgment in production, not learning prompt syntax — which is exactly the credibility argument behind building public agent-deployment work for phData sales.
Where this doesn't add anything new: the TDG-as-training-mechanism mapping (discovery-stage ROI sniff-tests, AI Council rehearsal, critic drills) was already made fully in [[2026-08-25-commoncog-why-tactical-decision-games-work]] and restated in [[2026-11-01-commoncog-making-of-common-sense-tdg]]. This email doesn't extend that application — it extends the premise (why judgment matters more under AI) that motivates it. Being honest about strength: this is a real, specific tie-in, but it's reinforcing an argument RDCO already holds, not new evidence that would change a decision. RDCO is dormant with phData as the main bet (project_rdco_dormant_phdata_main) — no action item here, just a sharper citation for a position already taken.
⚠️ Sponsorship
sponsored: true / sponsor_entity: Shadowbox. This is the explicit continuation of the sponsored sequence disclosed in [[2026-11-01-commoncog-making-of-common-sense-tdg]] — Cedric states outright in this email that it exists because "we're currently selling seats for a TDG Design Workshop by John Schmitt from Shadowbox," with two direct CTA links to the ticketing page (luma.com) and an explicit note that non-buyers still "benefit from the writing." Co-authored by Jared Peterson, a Shadowbox employee. The content (apprenticeship-failure analysis, "accelerated osmosis" framing) is credible as a sharper restatement of the original unsponsored essay's thesis, but its selection and timing serve the $2,000/seat workshop sale — same pattern as the predecessor note, now in its third installment. Treat each future email in this sequence as sponsored-by-default until the sequence ends.
Related
- [[2026-08-25-commoncog-why-tactical-decision-games-work]] — the original, unsponsored TDG essay this email explicitly summarizes and builds on.
- [[2026-11-01-commoncog-making-of-common-sense-tdg]] — the prior installment in this same Shadowbox-sponsored sequence; same sponsor, same workshop, same co-author (Jared Peterson).
- [[2026-04-19-commoncog-expertise-is-just-pattern-matching]] — the underlying pattern-matching claim about expertise that both the "seeing what experts see" and "osmosis" framings restate.
- [[project_credibility_for_phdata_sales]] — the founder's own judgment-as-moat positioning bet this email's premise reinforces.
- [[project_rdco_dormant_phdata_main]] — why this mapping stays conceptual rather than action-generating right now.