Writing

Practitioner notes
from production.

Essays and open-source releases from the team. No hot takes, no vendor pitch — just data and patterns from systems we run ourselves.

Latest — Architecture 2026-10

Where a Person Decides: Human in the Loop, by Rule

In 60 days, our AI judge passed 14 of the 19 jobs a person rejected. Why a person still decides, the question the judge never asks, the rules that call a person in, and what a decision becomes.

Read →

Architecture & Method

How the system around the model is built: gates, scaffolds, context delivery, and why structure beats instruction.

Architecture 2026-10

When to Stop an Agent Loop: Convergence, Not a Cap

An AI agent retrying a failed step can loop forever, and a fixed cap only hides it. How BPOS decides after every pass whether another is worth paying for, and why a bounded loop is not a converging one.

Architecture 2026-10

The Quality Gate: How BPOS Decides What Ships

A quality gate for AI agents, rule by rule: criteria judged three times by an LLM judge, a failed declared check scoring zero, the platform sending back its own failures, and a person deciding the rest. With the flaw we have not fixed.

Architecture 2026-10

Procedural Memory for AI Agents: One You Can Read

What procedural memory is, where AI agents usually keep it, and how BPOS keeps it in skill pages and the links between them: learned by practice from every outcome, with the traffic showing what agents need, and every step readable.

Method 2026-10

Nobody Reads It. The Work Asks It.

A wiki is knowledge you go to. Working knowledge is knowledge you look through. The difference decides whether an AI agent uses what your company knows, or ignores it.

Method 2026-10

A Decision Is Not a Fact

Three kinds of claim live in every company. A knowledge base that stores them as one kind will learn nonsense, and so will the AI that reads it.

Method 2026-09

Nobody Types the Weights

We decide how BPOS learns. The work decides what it learns. The three ways a learned weight can move, why one is locked, and the eight days that taught us.

Method 2026-09

A Hand-Written Map Can Be Worse Than No Map

Given no guidance, an AI agent scored 87.50 on a benchmark task. Given an expert’s map of the task, it scored 58.93. What repaired it, and where our own system still falls short.

Essay 2026-09

A Green Build Is Not Knowledge

A passing pipeline proves the code compiled and the tests you wrote passed. It does not prove the work does what was asked, or say why it worked. Across 12,103 recorded gate decisions and 331 sourced rejection notes, that difference is where AI-written software quietly fails.

Essay 2026-07

Don't Build Your Safety on a Chain of Thought You Don't Own

Reasoning traces from a third-party API are a safety surface you rent, not own. Building enforcement on them is faith, not verification. Three beats from our own stack: a trace reader, dependency ordering, and a portable test that survives a provider change.

Architecture 2026-08

You Cannot Out-Coach the Scaffold

We put a mechanical grounding gate in front of our plan generator and watched violations fall from nine to zero across seven generations — without loosening the gate once. The last two defects lived in the generator's own scaffold and in the judge itself.

Essay 2026-06

The Model Is the Smallest Part

We bet the system around the model matters more than the model itself. Over the last few months, five independent practitioners and a frontier lab arrived at the same conclusion — none of them ours.

Library · MIT 2026-04

context-steward

Lazy skill loading for agent systems. Claude Skills are useful when loaded on demand, expensive when loaded up front. context-steward is the npm package that solves the second problem without breaking the first.

Essay 2026-05

The Boring Parts That Ship

Multi-agent delivery doesn't break on the model. It breaks on file conflicts, bad sequencing, and per-job metrics that miss the deployment. The unglamorous machinery that decides whether a sprint ships.

Evidence & Null Results

Measured findings, including the ones that said no. Nulls published next to wins — the method is the product.

Essay 2026-09

Limitations, Perceived and Real

An AI assistant told us our system could not do eight things. Seven of those limits were never there. How to tell a perceived limitation from a real one before you build a workaround.

Evidence 2026-09

Our LLM-as-a-Judge Returned a Perfect Score 92% of the Time

LLM-as-a-judge has a failure mode nobody warns you about: saturation. Our conformance judge returned exactly 1.00 on roughly 92% of non-failed runs, looked statistically precise, and scored a fabricated output and a faithful one identically. The measurement, and the vector that replaced the single score.

Case study 2026-09

A Wrong Theory in the Feedback Makes the Next Run Worse

One incident, fully instrumented: a reviewer wrote a plausible causal theory into rejection feedback, the model obeyed the theory, and the retry regressed. Why the feedback channel carries sourced facts only. An n of one, stated as one.

Null result 2026-07

The Checklists That Did Nothing

We built skill checklists for our agents, believed in them, and measured them anyway. Across 88 arm-randomised scored executions, quality moved +0.002 — inside the noise. The null, the method that caught it, and the boundary on what it does not say.

Register 2026-07

Independent Convergence

A running record of independent practitioners, an academic study, and frontier labs arriving at the conclusions we built on — in their own words, with dates, sources and caveats. Updated as the record grows.

Essay 2026-06

We Deleted 20 of Our Own Quality Checks

We audited 36 mechanical quality checks and found about 20 had caught nothing in a month of production. The scaffolding you build for a weak model becomes a cage for a strong one.

Essay 2026-06

The Arithmetic of Absence

A GDPR-certified consent tool got a client fined anyway — because blocking a tracker doesn't make you invisible, it makes you a specific, reconstructable kind of missing data. On MNAR statistics, Bayesian imputation, and what enforcement actually catches.

Essay 2026-03

The 85% Accuracy Trap

Per-job quality scores in multi-agent systems almost always land around 85%. That number is telling you something different than you think it is. 4,600+ scored executions from our own production, and the pattern we found.

Essay 2026-04

The Median Trap

We ran a controlled A/B across 88 scored agent executions, with and without skill checklists in the prompt. Quality moved +0.002 — inside the noise. Why front-loaded instructions don't land, and what does.

Signals & Operations

Listening to production: telemetry, failure surfacing, and the feedback loops that turn incidents into structure.

Engineering 2026-09

The Click Your Dashboard Filed Under “Direct”

Server-side click attribution: record the click at the link, before the browser can drop it. How HikrLink does it, the one thing it refuses to do, and the three wiring bugs we found joining it to our own tracker — none on the link side.

Method 2026-08

How We Do SEO: Signals, Not Schedules

SEO as an instrumentation problem: four sensors, deconvolution that checks its own arithmetic, a hysteresis machine cleared through a four-run backtest, and a publisher that mechanically says no. Every decision leaves a receipt.

Engineering 2026-07

From 47 Minutes to 6 Seconds

The same failure surfaced twice in one day: 47 minutes in the morning, 6 seconds in the evening — a ~470× difference. Five changes, one principle: listen to the system's own signals.