Field Notes · Insights
Long-form notes on building AI that earns its trust.
Pieces from Scott Darrow on what's actually happening inside companies betting on AI — and what it takes for a solo founder with one AI collaborator to ship a system an enterprise can verify. Field reports, not vendor decks.
The field notes
One Founder. One AI. One Production System.
How Claude Code and I built a multi-tenant, 12-stage AI hiring intelligence platform — and shipped the audit system that proves every score it produces.
The minimum viable team for building a sophisticated SaaS platform is collapsing — not because engineers are becoming less valuable, but because the leverage available to a skilled technical leader has increased by an order of magnitude. This is a field report from inside that collapse.
Read the field notes →The Seniority
Penalty.
The most experienced candidates are now the most likely to be discarded before a human sees them. Five documented mechanisms, three architecture classes that carry them — and the design decisions that make each one structurally impossible.
Interview conversion collapsed from 1-in-7 to 1-in-33 across ten million applications, and 88% of employers admit their screening filters out qualified candidates. Every claim cited. One statistic deliberately refused.
The Rubric Foundation
Explained.
Candidate scoring, explained for the rest of us. What the Rubric Foundation is, why it exists, and what it unlocks — in plain English, with the gaps left in.
Every number a résumé score produces points at something you can read, challenge, and verify — including the honesty label that says which values are calibrated and which aren't. No "coming soon." The real status board, gaps included.
The AI Replacement
Theatre.
What's actually happening inside the companies making the headlines.
The headlines say customer service, driving, and coding are over. The METR data, the Gartner survey, and the Klarna reversal say something different. Here's what's actually happening at Spotify, Uber, IBM, Salesforce, Google and Meta — once you read past the earnings-call quotes.
Companies examined
Most résumés never
actually get read.
What a “talent intelligence platform” actually does — and the one thing it changes about hiring that I think genuinely matters.
A hundred résumés land for one role and your team gives a real look to maybe thirty. The other seventy aren’t rejected — they’re never read. An honest, plain-language tour of what changes when every applicant finally gets a fair read.
Scoring the
Scorers.
How effective is an ATS score, really? A visual field report rating résumé scoring against what effectiveness actually means — and where the whole category sits below the bar.
The number looks like a measure of candidate quality. It isn't — it measures how much a résumé resembles a job description. Seven dimensions, two grades, and the structural reason the strongest predictors of performance sit out of reach.
The Devil Is in
the Details.
The plain-language version was the what. This is the how — one résumé followed through all twelve stages, down to the one screen it was all for.
Fingerprinting, evidence-bound extraction, skill normalisation, external-signal fusion, five fail-closed gates and an immutable snapshot — ending at the candidate dashboard. What “reads every résumé” actually costs, stage by stage.
We were built
for this.
Why the AI-native era of IT runs on exactly the way the dyslexic mind is already wired.
AI is dismantling the parts of technology work that used to gatekeep it — the syntax, the documentation, the literacy-heavy layer — and elevating the parts dyslexic minds do natively: architecture, pattern recognition, the lateral leap. The era finally rewards the operating system some of us were running all along.
Information is
the Pitch.
The investor one-pager. Market, exclusion engine, vendor liability timeline, seniority-aware scoring and the audit-packet moat — the whole thesis on a single page.
Hiring runs on a number nobody can explain — and that number is now being litigated, regulated and audited. Every claim on the page cites its source, McCandless-style.
Stop reading CVs.
Start choosing people.
The one-pager for talent leaders. Every application read, scored and explained — ranked shortlists, evidence-linked reasoning, risk flags and interview prompts, above the ATS you already run.
One hundred résumés on a busy role, seventy never read. With Intelletto: 100 of 100 — and a one-click audit report behind every recommendation. Defensible to anyone who asks.