Sketches drift, specs go stale
There are two familiar ways to build. Write the spec first and let the code follow, or skip the spec and let the code find its shape. Jain's team tried specs: they went stale as the code moved, and nothing checked that the code still matched them. Without a spec, nothing holds the shape at all. Neither approach learns.
“The intent was written down, but nothing really checked whether we're actually still consistent with the spec.”00:00:29
Evidence: transcript 00:00:29
Every conversation is evidence
What makes learning possible now is that the work leaves a trail. The intent of a change is already in the agent session. Decisions surface when a coding agent and a review agent argue, and every disagreement is written down with its outcome. Evidence that a change works is captured as agents verify it. Sessions, chat logs and issues become a record of what the system is meant to be, without anyone writing documentation.
“And these already live in your sessions.”00:08:38
“Everything that they disagree about, we should write it down.”00:07:08
Evidence: transcript 00:07:08, 00:08:38, 00:13:02
What keeps proving true gets promoted
The record is where invariants are discovered. Jain's team mined their past review comments and found 200 candidate patterns, which distilled into 64 rules written in English. An agent then turned rules into scripts, and 19 became deterministic checks. Settled decisions take the same path and harden into the team's standards.
“Which now essentially become the standards of your approach.”00:14:09
Evidence: transcript 00:10:17, 00:14:09
The system reads itself back
Learning only counts if the record is used. Jain's team pulls the decisions behind a feature or a week and gives them to their agents, which then know what was already decided. The system's memory of why it is the way it is becomes context for the next change. Design-time decision records, like the ADRs the grill-with-docs skill writes, capture intended decisions; this captures the decisions actually made.
“Now, agent knows these past decisions.”00:14:09
Evidence: transcript 00:14:09
What learning looks like up close
What it doesn't solve
The risk is learning the wrong thing. Mined rules reflect past habits, including bad ones, and every rule needs an owner or it goes stale. Jain's answer is that responsibility stays with people: models can't own anything, so people own the guardrails and the rules. It also isn't free: his team spent about ninety days doing double work before it paid off.
“So instead of thinking about owning the code. Own the guardrails. Own the rules.”00:17:18
Evidence: transcript 00:05:48, 00:10:17, 00:17:18, 00:18:42