The Breeze
Private · Empire Edition
Hash check · Session in localStorage

Empire Learning Paths · Edition 1 · 2026-08-09

Four paths
into a one-person
empire.

You can read everything the B3RT empire ships and still not know what to do first. The four paths below are deliberately opinionated: a sequence, a deliverable per step, and a measurable result by the end. Pick one. Walk it. The empire will still be here when you finish.

4 paths 36 total steps 90 days max 0 frameworks required

00

How to read this

Each path has a thesis, a step list, and a "deliverable" for every step. The deliverable is the smallest artifact that proves the step happened. If you finish a step without producing the deliverable, the step didn't happen.

The four paths are not sequential. They are not prerequisites for each other. Path 1 ("Build Your Empire") is the canonical twelve-step path. Path 2 ("Crypto + Handyman Meta") is the local-biz-plus-DeFi variant. Path 3 ("AI Agent Layer") is for operators who already have a working business and want to layer agents on top. Path 4 ("Indie Empire from a Cabin") is the minimum viable stack for someone starting from zero and trying not to over-build.

Every step includes a "deliverable" block. Treat the deliverable as a contract: when the step is done, that artifact exists on disk (or in a public URL) and survives your laptop being turned off. The four paths collectively shipped 36 deliverables, and every one of them is a real file or live URL somewhere in the empire as of August 2026. None are aspirational.

The honest version. The empire you are reading about is one person's work, not a team. The paths below assume you have 4-6 hours a day, a workstation, and a willingness to ship ugly before you ship good. If any of those is not true, cut the path in half. The empire does not need you to do all twelve steps. It needs you to do the first two well.

01

Path 1 · Build Your Empire

Twelve steps over 90 days. The canonical path. Pick this if you have to pick one.

Path 1

Build Your Empire

12 steps90 daysOwner: 👑 Prime

The thesis: a one-person empire is the smallest operating system that compounds. Domain + brand + hub + services + content + social + monetize. Each step produces a file, a domain, a service, or a public surface that survives the step being finished. By step twelve you have a working empire that bills, posts, monitors, and improves itself.

  1. Day 1–4 — Domain & thesisPick the empire name and the customer promise. Buy the apex domain. Write a one-paragraph thesis you can defend in under a minute. Deliverable: a registered yourname.dev and a one-paragraph thesis.md on disk.
  2. Day 5–8 — Brand & colorPick one accent color, one type family, one mark. The empire looks the same on every surface. Deliverable: a 1-page brand-manifest.md with hex, font stack, mark, and "no-go" list.
  3. Day 9–14 — Empire HubStand up a FastAPI service on a free local port (e.g. :8096). It returns a SITES JSON list, an /api/health, and serves a static homepage. Deliverable: a live http://127.0.0.1:8096 with one card pointing at your homepage.
  4. Day 15–22 — Three servicesAdd three small FastAPI services, each with a different shape: a kanban, a knowledge graph, a portfolio sim. Each gets a port, an /api/health, and a SITES card. Deliverable: three /api/health endpoints all returning 200 with distinct service names.
  5. Day 29–36 — Editorial hubWrite the first 1,500-word editorial page on something you actually know. Post it at /essays/<slug>. Make the typography good. Deliverable: one published editorial page on the hub, mobile responsive, 1,500+ words.
  6. Day 37–46 — CF Pages deployPick the most public sub-page, deploy it to yourname.dev via wrangler. Verify with a real curl https://yourname.dev/ from outside the LAN. Deliverable: a live public subdomain returning 200 with security headers.
  7. Day 47–54 — Skill libraryWrite five SKILL.md files covering the procedures you have already used three times. The empire now has a memory. Deliverable: five SKILL.md files with frontmatter, procedure, verification, and pitfalls sections.
  8. Day 55–62 — First agent personaPick one repeatable task (morning brief, lead intake, watchlist probe) and give it a persona. The persona has a name, a deliverable, a Discord channel, and a cron schedule. Deliverable: one persona, one skill, one scheduled run, one Discord message.
  9. Day 63–70 — Content cadenceShip two editorial pages a week. Same word count, same template, same masthead. Cadence beats craft at this scale. Deliverable: eight editorial pages in eight weeks.
  10. Day 71–78 — Social surfacePick one social platform (X, LinkedIn, your own blog). Cross-post the editorial pages there with a 3-line hook. Track clicks. Deliverable: eight social posts, four of which got >0 engagement.
  11. Day 79–86 — First revenuePick the lowest-friction monetization (Stripe checkout for a digital product, a paid newsletter, a single bookable service). Ship it end-to-end. Test the failure path. Deliverable: one live checkout or one bookable service with a real URL.
  12. Day 87–90 — Imperator briefWrite the first 6-section imperator brief: status, money, services, content, action, risk. Deliverable: one morning-brief.md delivered on Telegram and saved to ~/.hermes/morning-brief.md.
02

Path 2 · Crypto + Handyman Meta

Eight steps. The local-service-plus-DeFi-treasury variant. Pick this if you have a working service business and want to keep the cash in your own name.

Path 2

Crypto + Handyman Meta

8 steps~60 daysOwner: 💰 Cypher / 🔧 Fixer

The thesis: a local service business produces durable cash, and a DeFi treasury preserves the upside. The "meta" is running both on the same workstation, with the service business as the audit trail and the treasury as the optionality. The crypto does not fund payroll. The crypto funds ambition; the service funds the operator.

  1. Day 1–5 — Service backofficeStand up a small FastAPI on :8091 for your service business. Intake, jobs, photos, invoices. SQLite, no auth (single user). Deliverable: a live local backoffice with three real jobs entered from your last week.
  2. Day 6–10 — Customer record of truthMove every customer record into the backoffice. One source of truth. Never again a spreadsheet. Deliverable: 100% of active customers in the system, 0% in a spreadsheet.
  3. Day 11–15 — Stripe checkoutAdd a Stripe Checkout link for a fixed-price service (a turnover clean, a single repair, a one-time consultation). Test the failure path. Deliverable: one paid test checkout, $0.50 to your account.
  4. Day 16–25 — Treasury splitOpen a separate treasury wallet (hardware or multisig). Route 10% of every service payment to the treasury. The split is automatic and visible. Deliverable: a treasury wallet with at least one incoming transfer, visible on-chain.
  5. Day 26–35 — Stablecoin yieldMove the treasury's stablecoin into a yield-bearing position (lending, LP, or staking) with a hard cap of 5% of net worth. Document the risk, the exit, the monitor. Deliverable: a one-page treasury-policy.md with risk, cap, exit, monitor.
  6. Day 36–45 — Invoice chaserAutomate the overdue-invoice chaser. The script reads the backoffice, finds invoices > 7 days overdue, sends a single email, logs the send, and reports to Discord. Deliverable: one cron job, one chaser email sent, one Discord notification.
  7. Day 46–55 — Operator dashboardOne FastAPI page at :8099 that shows: open jobs, overdue invoices, treasury balance, treasury yield, last chaser run. Single screen, single morning. Deliverable: one dashboard that you actually open daily.
  8. Day 56–60 — Public siteShip a mobile-first public site for the service. One page, one phone number, one contact form, one Stripe link. Deliverable: a public subdomain with the contact form wired to the backoffice.
03

Path 3 · AI Agent Layer

Ten steps. The "I already have a business, now I want agents" path. Pick this if you have ops to automate and you trust yourself to verify the agents' work.

Path 3

AI Agent Layer

10 steps~75 daysOwner: 🦙 Prime

The thesis: an agent is a procedure that survives context loss. The empire's agent layer is ten procedures, each with a persona, a deliverable, and a verification. Local LLM is the default; paid models are the exception. Agents never make a customer promise — they only produce artifacts a human inspects.

  1. Day 1–5 — Local LLMStand up a local 35B model on your AMD or NVIDIA box, --ctx-size 32768, exposed at :8084 as OpenAI-compatible. Test it with five prompts. Deliverable: curl http://127.0.0.1:8084/v1/models returns 200.
  2. Day 6–10 — First skill, not first agentWrite a single SKILL.md for a procedure you have used 3+ times. Do not start with an agent. Start with a skill. Deliverable: one SKILL.md with frontmatter, procedure, verification, and pitfalls.
  3. Day 11–18 — First personaGive the skill a name and a deliverable contract. The persona has a Discord channel, a cron schedule, and a single artifact per run. Deliverable: one persona, one cron-driven run, one artifact in the channel.
  4. Day 19–26 — Cron, not chatMove the persona from "called on demand" to "called by cron, output to channel". The chat interface stays as a debug tool; the production interface is a clock. Deliverable: a cron job that runs the persona unattended and posts the result.
  5. Day 27–34 — Knowledge graphBuild a tiny SQLite-backed knowledge graph: nodes for "skills", "personas", "deliverables", "domains"; edges for "owned by", "produces", "depends on". The graph powers a "what to delegate next" view. Deliverable: one kg.sqlite3 with 30+ nodes, a force-laid-out canvas view, and one query that surfaces the highest-leverage next task.
  6. Day 35–42 — Second personaAdd a second persona with a different shape (research vs. build, or monitor vs. write). The pair tests the persona boundary. Deliverable: two personas, two channels, two cron jobs, two artifacts per day.
  7. Day 43–50 — Briefing layerBuild a morning-briefing skill that aggregates all persona outputs into a 6-section markdown brief, delivered to Telegram at 08:00 ET. Deliverable: one Telegram message every morning for five consecutive days, with the same shape.
  8. Day 51–58 — Drift detectorWrite a drift detector that compares expected state to actual state (cron jobs, services, public sites, free RAM) and alerts when something degrades. The detector runs every 5 minutes via cron. Deliverable: one drift detector with a real Discord alert for one real degradation.
  9. Day 59–68 — Subagent disciplineAdopt the subagent contract pattern: every delegated task has an input, an output, a budget, a verification, and a then-clause. Update every persona's skill to use the contract. Deliverable: five subagent runs with contracts, all of which produced verifiable artifacts.
  10. Day 69–75 — Self-evolutionEnable the self-evolution loop: every Friday, the empire reviews the week's drift, the week's drift-resolved, the week's new skills, and proposes one skill to add, one skill to prune, one drift to formalize. Deliverable: one Friday review post in #empire-meta, three concrete next-week actions.
04

Path 4 · Indie Empire from a Cabin

Six steps. The minimum viable stack. Pick this if you have 90 days, one workstation, and no team.

Path 4

Indie Empire from a Cabin

6 steps~45 daysOwner: 🏔️ Operator

The thesis: the smallest empire that compounds is one workstation, one domain, one public page, one paid product, one local cron, and one morning brief. Skip everything that does not directly serve one of those six. Resist the temptation to add a seventh. The cabin empire is a tempo, not a feature list.

  1. Day 1–7 — The workstationBuy or repurpose a single workstation. Local LLM optional, but reserve 32 GB of free RAM and one GPU if you have it. Stand up one FastAPI on one port. Deliverable: one local service with one /api/health returning 200.
  2. Day 8–14 — The domain and the pageBuy the apex domain. Deploy one editorial page with a real masthead, one accent color, one type family, and one call to action. That is the entire public surface for now. Deliverable: a live yourname.dev with one 1,500-word page and a working contact form.
  3. Day 15–25 — The paid productPick the smallest possible paid thing. A Stripe checkout for a $5 PDF, a $20 fixed-price service, a $50 one-hour consultation. Ship the checkout. Test the success and failure paths. Deliverable: one paid test order, $0.50 to your account, the receipt emailed and logged.
  4. Day 26–35 — The local cronWrite one cron job that does one thing unattended. A morning brief, a watchlist probe, a chaser email. The cron should be small enough that you can read it in one screen and explain it in one sentence. Deliverable: one cron job, one artifact per run, one channel receiving the artifact.
  5. Day 36–42 — The morning briefWrite a 6-section morning brief that aggregates: status, money, services, content, action, risk. The brief is one markdown file, no more than 1,200 words. Deliverable: five consecutive mornings with a brief delivered to your inbox, all under 1,200 words, all readable in 3 minutes.
  6. Day 43–45 — The skip listWrite down the things you decided not to build in the first 45 days. Examples: not a second service, not a Discord server, not a newsletter, not a podcast, not a content calendar, not a second domain, not a third persona. The skip list is a feature of the cabin empire. Deliverable: a skip-list.md on disk that survives a future version of you wanting to over-build.

What "from a cabin" actually means. It is a metaphor for a constraint: bandwidth, attention, and equipment are scarce. The cabin operator's edge is that every addition is forced to earn its place against a real cost in focus. The opposite of the cabin empire is the "platform empire" that builds eleven dashboards before the first dollar. The cabin empire builds one dashboard, ships, then decides whether the second one is a real improvement or a vanity.

05

Cross-cutting habits

Five habits that apply to every path, regardless of which one you picked.

  1. One receipt per claim.Every "the empire is up" needs a curl or a SELECT COUNT(*) behind it. No vibes-based status.
  2. One file per skill.If a procedure repeats, it gets a SKILL.md. The library never lies about itself.
  3. One cron per persona.Production interface is a clock. The chat interface is a debug tool.
  4. One page per editorial day.Cadence beats craft at the start. The first eight pages are not the portfolio; the next eighty are.
  5. One skip list per quarter.Write down what you decided not to build. The skip list is the only thing that protects the next quarter from the previous quarter's enthusiasm.
06

The proof

The B3RT empire is on the receiving end of every one of these paths. The receipts are on disk.

Path 1 shipped the b3rt.dev apex, the 11 live FastAPI services, the 19 public subdomains, the editorial cadence. Path 2 produced the BE Repaired backoffice at :8091, the Sparkling CRM at :8093, the Stripe boardroom checkout, the treasury policy. Path 3 produced the 20+ legion personas, the 424+ skills, the cron-driven morning brief, the drift detector. Path 4 produced the skip list that made the other three possible.

If you walked a different path and arrived at a different shape, that is the point. The four paths are not a curriculum. They are four receipts for the same operating system, written by the same operator, against the same constraints. Take the parts that fit. Leave the parts that do not. The empire is not the path; the path is the empire.