Writing

What the role is, where it came from, what GTM engineers build, from someone who did the job before it had a name.

The unicorn hire for companies standing up a GTM engineering org, and why verification is the hardest part of the job.

The postmortem: what broke, why every run still passed, and the regression gate that stops it.

Open-sourcing the outbound stack we run for clients. Keep the data vendors, drop the dashboards. The interface is the agent.

Claude 5 models change which scaffolding matters: operational knowledge, calibrated judgment, and controls the model cannot improvise around.

12 models, one rubric, 243 funded AI companies. Above $22 per thousand accounts, price stopped buying accuracy.

Most GTM teams running AI cannot tell you whether it works. Why errors compound, why failures are silent, what a working eval looks like.

Sales lists have been filtered on a taxonomy built for government statistics. What changes when the filter moves to what a company actually does.

The six building blocks (skills, subagents, hooks, MCP, plugins, routines) mapped to every stage of GTM. Written companion to the masterclass.

What used to take six weeks now takes an afternoon. The live demo from the Claude for Marketers & GTM event in Toronto.

AI SDR tools churn at 50 to 70% a year because they cannot hold your internal context, and context is the whole game.

Four types of candidates show up for GTM engineer roles, and each is missing something critical. The gap is structural.

The quickest way to get started, and what to focus on to get comfortable with terminal-based AI development.

The four most common objections GTM leaders give, and why none of them hold up.

Compound context, not prompting skill, is the real unlock for GTM teams.

Why Claude Code is in a different league for GTM: code plus context. A single GTM engineer now has the leverage of an engineering team.

The Signal, Bridge, Voice framework, and why code is the natural tool for the whole GTM pipeline.

The models are smart enough. The bottleneck is whether they have the right context at the right time.

Three types of sales orgs ranked by data hygiene, and the four-step system to stop wasting reps' time on stale contacts.

Brex spend data from 35,000 companies: the fastest-growing tools are almost all GTM engineering.

A reasoning model, not a base-model upgrade. How that changes token economics and when to use it in GTM workflows.

Why reasoning models fit GTM work, a four-part prompt structure, and whether o3-Pro is worth it.