@rohanpaul_ai: A firm’s judgment does not live in its archives; it lives in the changes (diffs) senior people make before work ships. …
Summary
This article argues that a firm's true judgment is not found in archived documents but in the edits (diffs) made by senior employees before work is finalized, and that AI should learn from these differences to capture institutional knowledge.
View Cached Full Text
Cached at: 06/26/26, 08:08 AM
A firm’s judgment does not live in its archives; it lives in the changes (diffs) senior people make before work ships.
Farsight calls this a system of judgment, i.e software that preserves those edits across real work, can turn repeated decisions into measurable rules.
The next enterprise AI moat is not stored knowledge, but stored judgment.
If AI is going to learn the work of professional firms, it cannot learn only from polished outputs.
AI needs the gap between first draft and final draft, because that gap contains the firm’s private standard of what good means.
Kunal Tangri (@TangriKunal): “Capture your institutional knowledge” has meant the same thing for 30 years: index the documents, search over them.
But a document is the output of judgment with the judgment removed.
The judgment is in the diff. The markup between the first draft and the one that ships. And
Similar Articles
@rohanpaul_ai: AI can make junior professionals look much better at the work without making them better at the underlying judgment. Th…
A randomized trial with 133 patent lawyers found that AI assistance improves drafting quality, especially for juniors, but does not enhance underlying judgment, as shown in a judgment-heavy task without AI.
What is human judgment doing in AI’s “last mile” that models still can’t?
The author reflects on whether human involvement in AI's 'last mile' has shifted since the 2019 book 'Ghost Work', exploring current unique roles of human judgment and its potential evolution.
Use AI for reviewing code especially when the diff is huge
The article argues that human code reviewers should use AI to handle large diffs, and instead contribute their out-of-distribution knowledge and high-level context.
@hnshah: https://x.com/hnshah/status/2066761276945211442
Hiten Shah reflects on how AI is transforming GitHub from a repository of evidence into an environment where non-coders can apply product judgment directly to software workflows, bridging the gap between customer understanding and code.
@heyshrutimishra: Any AI workflow a company wrote down three months ago is probably already out of date. That isn't my hot take. It's Aar…
A commentary on Box CEO Aaron Levie's viral post arguing that AI progress is accelerating rather than plateauing, making company AI workflows outdated within months and widening the gap between frontier users and everyone else.