№ 08 — FIELD NOTES

Notes from the workshop

Working notes, research, and build logs. Some figures are reserved for subscribers.

MEMBER ACCOUNT →

2026-06-17How to run CLI coding agents without burning your budget or your codebaseRunning a CLI coding agent effectively requires treating it as a contractor, not a vending machine. Separate planning from execution, gate every session with a structured project-brain file, watch your context percentage like a fuel gauge, and set a definition of done before you type the first word. Done that way, agentic coding delivers a compounding productivity gain. Done without discipline, it generates thousands of unreviewed lines, burns through a month's token budget in an afternoon, and leaves the codebase worse than before.≈ 11 minREAD →
2026-06-17Building a personal-brand content engine that compounds: pillars, waterfall, and the trust ladderIn an interest-based algorithm, a small account's single post can outperform a large one — which means the right content architecture matters more than follower count or posting frequency alone. This playbook covers the authority-heavy pillar split, the content-waterfall repurposing system, the short-form trust ladder, the CAPSTONE long-form structure, the comment-to-DM conversion mechanic, and the outlier-audit verification loop — with a 4-hours-per-week budget, a 3x-per-week posting floor, and the 90-day reset rule that defines what "consistent" actually requires.≈ 12 minREAD →
2026-06-17Building your autonomous desktop layer: skills, routines, and the always-on assistant that actually worksAn autonomous desktop layer is not a product you install — it is a configuration you version-control. The durable bet is capability-as-config: each skill is a plain-text file that tells the agent exactly how to do one repeatable job, in your voice, with your guardrails, and against your persistent business context. Skills load on demand, swap without code changes, travel across chat interfaces and code environments alike, and accumulate into a system that gets sharper the longer you run it. This playbook covers the three-folder co-work station, the DBS skill framework, local versus remote routines, MCP reach extension, outcome-first prompting, and every failure mode worth knowing before you build.≈ 16 minREAD →
2026-06-17How to build automation workflows that survive production: the fork, test, and monitor disciplineThe most reliable way to build n8n workflows that stay running in production is to fork a working community workflow, test it against real input before scheduling it, and monitor execution logs on a fixed cadence. The highest-value automation work is deterministic data plumbing — speed-to-lead routing, document extraction, follow-up sequences — not LLM-heavy agent wizardry. This playbook covers the 6-step build method, the item-array mental model, the reusable node idioms, execution reliability targets, the 20%-failure switch rule, self-hosting economics, and the specific failure modes that kill otherwise sound automation projects.≈ 15 minREAD →
2026-06-17How to rank the best recent videos on any topic — a weighted, recency-aware methodA reusable method for pulling the best *recent* videos on a topic — not just the most-viewed all-time. It scores every candidate on five weighted levers (recency decay, view velocity, channel over-performance, damped popularity, engagement), gates on age/views/duration first, caps per channel for diversity, and gets the publish dates it needs from a bulk metadata provider (since plain search can't rank by date). Every parameter is tunable; the defaults favor the last few months.≈ 4 minREAD →
2026-06-17How to package and price an AI automation service that clients actually pay forThe one qualifying question for any AI automation offer is whether the client can see the system pays for itself. If they can, you have an offer. If they cannot, you have a feature. Everything else in this playbook — the manual-first validation sequence, the boring-offers catalogue, the niche-selection rules, the three-tier pricing model, and the productized-SaaS transition — is a direct consequence of that single filter. Build what makes or saves money visibly, price it against the labour cost it replaces, validate demand before writing a line of code, and commit to one niche for 90 days before you consider expanding.≈ 14 minREAD →
2026-06-17When to use a workflow and when to use an agent: the decision that determines everything elseThe field has reached consensus: reach for a deterministic workflow first, every time. An agent — an LLM wired to tools, memory, and goals that runs its own observe-think-act loop — earns its place only when the process is genuinely unpredictable and a workflow cannot be written. Getting this decision wrong costs months: over-agentified pipelines are slower, harder to debug, more expensive, and less reliable than clean rule-based logic. This playbook documents the scoring method, the accuracy-threshold sequence, the four-part agent anatomy, the definition-of-done constraint that most builders skip, and the failure signatures that tell you when you built the wrong thing.≈ 12 minREAD →
2026-06-17How X/Twitter growth actually works in 2026 and the system to earn it without gaming itX growth in 2026 is not an engagement game — it is a relevance game, and the rules changed when the platform shifted from heuristic scoring to AI-driven content evaluation. Grok reads every post and judges whether it is genuinely useful to a specific audience; engagement pods, mass-follow tactics, and copy-paste repetition now register as noise at best and trigger suspension at worst. The system that works is methodical: identify proven posts from peers at the right scale, extract their structural templates, apply them to your own expertise at a daily cadence, and invest one hour every day in genuine peer engagement before your follower base is large enough to carry you on its own. This note lays out the full playbook — strategy, four-lever taxonomy, template cycle, content mix, copywriting disciplines, verification signals, failure modes, and the automation boundary.≈ 14 minREAD →
2026-06-16How to get cited by AI answer engines in 2026A field-tested playbook for getting a site found and cited by ChatGPT Search, Perplexity, Claude, and Google's AI Overviews in 2026 — the full set of levers ranked by impact-for-effort, every number traced to its source, the myths that waste your time, and a teardown of how I applied each lever to this very site.≈ 8 minREAD →
2026-06-16The complete agentic operating system I'm building on Windows — full plan and executionWORK IN PROGRESSA Windows-native build plan for a complete agentic operating system that turns my real automation work into content, funnels, and clients. Every tool is verified for Windows 11 as of mid-2026 and wired to what I already run — Claude Code, a knowledge graph, a live members funnel on Neon and Vercel, a 14,099-post corpus — with mermaid architecture, honest cost economics, and a phased build that earns autonomy before it takes it.≈ 15 minREAD →
julian-goldie-communities-and-distribution≈ 6 minREAD →