---
type: "synthesis"
sources: ["agentic"]
tags: ["talent", "apprenticeship", "future-of-work"]
id: "cross-apprenticeship-erosion"
sourceVaultSlug: "hbr-seg-agentic"
originDay: 6
articleStem: "hbr-seg-agentic"
sourceUrl: "(unified vault: 9 sources)"
sourceTitle: "HBR — Firm Ⅱ-A · Agentic operating model / brand code / workstreams"
---
A quiet but corpus-wide worry: if agents do the entry-level work, **where does human judgment come from?**

A026 names it the [[concept-invisible-pipeline]] — the apprenticeship in which juniors absorb tacit rules by doing "grunt work." Automate it and you get [[claim-eroding-governance-capacity]] ([[quote-automate-judgment]]): *"automate the work that built judgment and you erode the capacity to govern the systems that replaced it."* A026's fix is [[action-protect-practice-ground]] — red-team rotations where juniors audit and break AI decisions — though A026 honestly asks whether that scales ([[question-scaling-apprenticeship]]).

A027 answers from the opposite direction: don't just protect the pipeline, *reinvent* it. Ramp's model in [[action-train-employees-to-build]] trains new hires from day one to build their own agents — cultivating [[concept-thought-doer|thought-doers]] rather than button-pushers.

A017 shifts hiring criteria entirely: [[action-hire-for-agency]] and [[claim-hiring-for-agency]] screen for judgment and ownership over rote execution.

**The tension worth holding:** A026 fears the loss of slow, high-volume practice that builds intuition; A027 bets judgment can be *taught faster* via building and debate ([[framework-scenario-based-extraction]]). Both agree the old ladder is broken; they disagree on whether the replacement suffices. This feeds directly into [[cross-executor-to-judge]] — you cannot staff a firm of "judges" if you have destroyed the mechanism that produces judgment.