Replace the Trivial, Augment the Difficult
A rule in the code-review engine I built caught an idiom I've been writing for years and can no longer see. That's the small half.
Rhys Brett-Bowen — Coding in the age of AI
-
-
A Flip Is a Missing Rule
A colleague sent me the cleanest bug report I've had all year. He'd taken an open change, squashed it, and opened it again — same content, byte for byte, as…
-
How a Team's AI Coding Flow Should Be Wired
A change lands in review with everything green. The rules pass, the types check, the tests are close to the ones I'd have asked for.
-
The Other Intent
The business intent, what we're accomplishing, goes in the PR description. Solved problem; we've had a text box for it since forever.
-
You Can't Save a Layout You Don't Own
The request was: let me arrange these panes how I like, give the arrangement a name, and hand it back to me tomorrow on a different laptop.
-
The Agent Will Invent the Bar
One evening an agent working in this repo killed an idea of mine for failing a rule. The kill was well argued and it cited the rule by line number, the way…
-
A Pipeline Is What It Refuses to Forward
The first stage of my code reviewer hands the next one a two-field object: a rule id and a citation. Not its reasoning. Not the file it just read. Two fields.
-
Nobody Broke It
It looks at a deprecation deadline declared in a schema, calls new Date(), compares, and if the deadline has passed it fails.
-
Buy Back the Coverage
I had already published the warning: if a model only judges what a scanner shortlisted, a site the scanner fails to flag is a false negative nobody ever sees.
-
Intermittent Competence Looks Like Coverage
My reviewer caught a deleted flush call three times out of three, and said nothing about eight of thirteen removals in any run.
-
A Constraint You Can Quietly Lift Isn't One
I had a harness measuring whether a set of skills helped a cheap model. It came back negative: no lift worth the tokens. I believed it.
-
Correct Is Not Usable
I asked an agent which of three approaches to take. It gave me three — each described accurately, tradeoffs and costs and failure modes, all of it right.
-
Build the Operator, Keep the Operands as Data
A good rule set grows. Mine has grown — tested against a whole corpus of real PRs and code — into something too big to hand a model all at once, and that's…
-
The Harness Is the Product
The reviewer is the visible product: it reads a diff, applies a rule corpus tested against real PRs and code, and ships the findings worth a person's attention.
-
The Raw Is the Trust Boundary
Ask a model to summarise a document and you can hold the result against the document — the raw is right there.
-
Train on What the Gate Throws Away
My code reviewer does far more work than it ships. Scouts cast a wide net — anything that might be a violation.
-
Keep the Factory, Ship the Result
I built an eval to tell me whether my agent's skills were earning their place. For a while it said they weren't — clean, green, confident.
-
Fluent in a Dead Idiom
A newer engineer left a note on my pull request: this array should be readonly. He was right.
-
A Big Rule Set Is an Attention Problem
Sixteen minutes and sixteen dollars to surface four low-severity nits, most of it spent scanning for rules rather than judging code.
-
Where the Reps Come From
Advice to keep reading assumes you built the judgment first. If AI writes the first draft, the cohort coming up is asked to keep what it never built.
-
Rules Are the Alpha
Two reviewers, the same commit — a fat change, a few thousand lines. One was a deep reviewer I own, running a portfolio of judgment rules tested across a…
-
A Cue Finds More of What It Can Already See
An AI reviewer finds three bugs on a real PR. Tell it to expect four or five and it finds more — but only more of what it could already see.
-
Familiarity Isn't Trust
I caught myself telling an AI, "you're starting to sound like me." My cadence, my hedges, the figures I reach for — it had picked them up and was handing them…
-
The Agent Reaches for the Nearest Metaphor
I had a figure I liked. Writing about giving the AI the why instead of the what, I'd reached for a bespoke image — mine, a little odd, exactly the shape of…
-
Shrink, Reorganise, Relocate: Why Synthetic Wins Fail on Real PRs
I spent a stretch trying to make an AI code reviewer cheaper without letting it miss more. Cost and recall, two dials pulling against each other.
-
Augmented Reading
The reflex, when a model meets something long, is to make it shorter. Summarise the PR, TL;DR the paper, have the agent review the diff and hand back the…
-
You Can't Grade Your Own Misses
Ask a code-review tool how good it is and you'll get a precision answer: of the things it flagged, how many were real.
-
Fidelity Was a Vibe, Cost Was a Number
I told the agent fidelity came first. It optimised for cost anyway, for an embarrassing reason — cost had an estimator and a guardrail, and fidelity had a…
-
The Nit Was Never Optional
Read that definition again. It's one of the most widely-cited code-review standards telling reviewers how to flag a small issue: say the author should fix it…
-
Your Merged-PR History Is Survivorship-Biased Training Data
In the Second World War, the Statistical Research Group was asked where to add armour on bombers, and handed the damage data from the planes that came back …
-
Confidently Wrong
A model handed me a clean answer once. Well structured, sure of itself, and wrong — and nothing in how it was written gave that away.
-
Noticing Is Cheap, Judgment Is Scarce
The goal was never a cheaper reviewer. It was finding everything that is really there while posting nothing that isn't — and that objective, not the budget…
-
Your Skills Are a Codebase, Not a Junk Drawer
I asked the agent to review a diff. Two skills woke up and started fighting over whose job it was.
-
Skills are one-offs. If you want an ecosystem, you need plugins.
You wrote a skill. It's good. Now you want it to lean on another skill — call it, require it, ship the two together so installing yours pulls in the one it…
-
"Context Debt" Is a Strawman
Here's the pitch, and it's a good one. Vibe-coded slop is only the symptom; the disease underneath is context debt — the steady loss of what the code means…
-
Evals Aren't the Moat
A fast eval told me a restructured module was clearly better. Comfortable margin. I ran it again to be sure, got the same verdict, promoted the new version…
-
A Pool of Models Narrows the Overlap — It Doesn't Erase It
Sakana AI shipped Fugu with a clean pitch: one endpoint that routes your task across a pool of models — calling itself recursively — and handles selection…
-
Teaching the Reviewer
I was building a code-review tool. Nothing fancy — read the diff, flag what's wrong, the job I do by hand a dozen times a week.
-
Thinking at the Speed of Code
A while back, I stopped reading the AI's end-of-task summaries and started watching the diffs. They came through faster than I could have typed them.
-
A Description Can't Replace the Code
A description is a compression, and it drops the omissions it doesn't know it made — the case not handled, the check not added.
-
Give Away the Writing, Keep the Reading
Linus Torvalds vibe-coded the visualizer and hand-wrote the audio math. One repository, the whole argument — writing code was never the scarce skill.
-
The Productivity Mirage
Sixteen experienced developers finished 19% slower with AI and believed they were 20% faster. The feeling of a fast session tracks fluency, not distance.
-
Your Codebase Is the Prompt
The AI reads from your codebase before it writes into it. Whatever's already there — your conventions, your shortcuts, your best work and your worst — is the…
