What if the agent that already holds your strategy and your story could write your whole funnel - landing page, lead magnet, follow-up sequence - without you re-explaining a thing?

That's the third move in the Growth Codex series. Once strategy and content already live in your codebase, the funnel stops being a design file in a separate tool and becomes the user-facing rendering of context the agent already has. The split stack — WordPress page here, app there, a form plugin in the middle — is exactly what creates the double sign-up and the attribution black hole. One repo closes the gap.

This episode walks through the real build: Bridesmaid's "First Week After Yes" lead magnet, assembled inside Claude Code straight from the context docs. You'll see why the lead-magnet sign-up IS the product's free-trial sign-up (no handoff, full attribution from first click), how a typography fix on one landing page propagated across the entire site and web app, and how the build fed its own learnings back into the strategy and content skills — so the next funnel starts smarter.

There's also the honest part: a plain PDF was a real consideration, and why the interactive version won anyway.

This is not theory. It's a git-proven build you can run this week with three copy-paste prompts.


IN THIS EPISODE

00:00:00 – Why vibe code your marketing funnel
00:01:10 – Stop building funnels by hand
00:01:25 – The Bridesmaid lead magnet: "First Week After Yes"
00:02:39 – The old way: split stack, double sign-up, zero attribution
00:03:34 – One repo: funnel and product, no gap
00:04:13 – Inside Claude Code: the build becomes a skill
00:05:33 – Promote it: social posts from your content skill
00:06:24 – Nothing to re-explain (and a typography fix that spread everywhere)
00:08:51 – Seamless: the lead-magnet sign-up IS the product
00:10:15 – Interactive vs PDF, vibe-coded emails, and the compounding loop
00:13:09 – Your 3 prompts, what's next, and where to meet us


DEEP DIVE LINKS

📖 Full Article: https://pirateskills.com/insights/vibe-code-funnel

📚 The series so far:
→ Growth in Code (strategy in the repo): https://pirateskills.com/insights/growth-in-code
→ Content Skill (content as an agentic skill): https://pirateskills.com/insights/content-skill

🧪 Vibe Coding Cologne — Wed June 3: meet the community in person in Cologne: https://pirateskills.com/events/vibe-coding-cologne/2026-06-03

🎧 Spotify: https://open.spotify.com/show/7kFl0BJiOYih8yqQTRR9SN

🍎 Apple Podcasts: https://podcasts.apple.com/us/podcast/build-grow-repeat/id1870561487

🗓️ All upcoming events: https://pirateskills.com/events

💬 Connect with me on LinkedIn: https://www.linkedin.com/in/bensufiani/

Shipping a new feature has become cheaper and faster than posting about it. That's ridiculous, right?

The grindiest job in any founder's week is content — and it's the one job we keep doing by hand. This episode shows the way out: take the four questions every brand has to answer (why we do this, what we talk about, how we say it, when we post it), put the answers in your codebase, and wrap them in one agentic skill that drafts a full week of on-brand content for you. The soul stays human. The grind goes away.

This is a fresh, first-time setup — not a polished after-shot. Christina and I sat down for a two-hour interview inside the Bridesmaid repo (the wedding-planning SaaS we're co-building) and built the content skill from zero. You'll watch the skill plan a real week in Claude Code with Sonnet 4.6, see Christina vet the picks as the first-time bride, then go file-by-file through the actual SKILL.md, the story doc, the themes-and-topics doc, the voice rules, and the weekly cadence. Hundreds of lines, all shared.

You'll also see the moment that proves why you can't hand the keys over on day one: the AI confidently misheard Christina's "all-in-one plan" as "hole-in-one" — twice. That's the case for L1 review by default. Treat the skill like a baby: review every piece, give annoyingly explicit feedback, and earn autonomy one format at a time.

This is not theory. It's the same skill Christina and I just shipped inside Bridesmaid — and the exact recipe to build your own.


IN THIS EPISODE

00:00:00 – Wrap your content production in a skill (no AI slop)
00:00:47 – Live demo: the content skill plans a full week in Claude Code
00:02:53 – Christina vets Monday, Tuesday, Friday as the first-time bride
00:03:57 – Why content marketing became the grindiest job in the company
00:06:18 – The four founder calls: Why · What · How · When
00:10:53 – From two-hour interview to one agentic skill
00:11:35 – Inside SKILL.md: the operating contract, step by step
00:18:11 – The four reference files: story, themes, voice, cadence
00:23:46 – The "hole-in-one" mis-hear: why L1 review is the default
00:26:31 – Treat your skill like a baby: the four autonomy levels
00:31:24 – Pirate Lab · Vibe Hackathon · Pirate Forge


DEEP DIVE LINKS

📖 Full Article: https://pirateskills.com/insights/content-skill

🎧 Spotify: https://open.spotify.com/show/7kFl0BJiOYih8yqQTRR9SN

🍎 Apple Podcasts: https://podcasts.apple.com/us/podcast/build-grow-repeat/id1870561487

🗓️ All upcoming events: https://pirateskills.com/events

💬 Connect with me on LinkedIn: https://www.linkedin.com/in/bensufiani/

Your product lives in the codebase. Your growth strategy lives in Notion, in slide decks, or in conversations you had on a walk. What happens when you move it all into the same repo your agents already read?

The setup is simpler than it sounds. You open Claude Code in your product's repo and answer prompts out loud through Wispr Flow – the agent writes the markdown for you. Every session after that is dual-purpose: it answers the question in front of you AND patches the underlying doc when something turns out to be wrong. The strategy layer self-updates.

This episode walks through bridesmaid.love live – the setup Ben ran with the Pirate Forge cohort last Wednesday. You'll see docs/context/ populate from a founding-interview prompt, a real cookie-consent question turn into a fully-contextualized Linear ticket, and the agent assign next week's work between Ben and Christina from the OKRs alone.

Not theory. A real strategy folder, a real Linear board, and three prompts that ship the entire setup in about an hour.

IN THIS EPISODE

00:00:00 – Why growth strategy belongs in your codebase
00:00:45 – No context, no help
00:02:31 – Half the company's brain was invisible
00:04:51 – Build and grow in one repo (Bridesmaid + Pirate Forge cohort)
00:06:00 – Live: the founding interview that seeds docs/context/
00:07:27 – Inside the strategy folder: business, OKRs, bottlenecks
00:08:20 – The knowledge base that writes itself
00:11:41 – Live demo: a question becomes a Linear ticket
00:16:32 – Strategy that lands in this week's Linear cycle
00:19:40 – Who tackles what – the agent assigns the week
00:23:27 – Run the setup this afternoon: three prompts


DEEP DIVE LINKS

📖 Full Article: https://pirateskills.com/insights/gro...

🎧 Spotify: https://open.spotify.com/show/7kFl0BJ...

🍎 Apple Podcasts: https://podcasts.apple.com/us/podcast...

Paul Graham named founder mode. Brian Chesky added AI to it. Jack Dorsey laid off 4,000 people to get there. All three are catching up to where small-but-mighty teams already operate.

AI Founder Mode isn't a new playbook for big companies. It's the long walk back to where two-person teams already are. We never built the middle layer — so "stay in the details, intelligence at the center, founder close to the work" isn't a restructuring for us. It's our default state.

The mechanic: a four-step founder loop. Do the task once — you guiding an agent through it. Run /skill-creator in the same session. Hand the next run to a human or another agent. When they hit friction, they apply the feedback and update the skill itself, right there. The skill gets sharper every run. No SOP wikis no one reads.

The shift that goes with it: company knowledge migrates out of Notion and into the codebase — because the team members who actually run the work, the agents, live there. The skill is the new onboarding. The agent is the new hire that doesn't need three months and a salary to contribute.

Small isn't a phase you grow out of. It's the strategy.

IN THIS EPISODE

00:00:00 – AI Founder Mode for the Rest of Us

00:00:25 – Why Paul Graham's Founder Mode hit different in 2024

00:04:38 – Three founders, one direction (and Dorsey's 4,000 layoffs)

00:06:50 – Why SOPs go stale (and what to do instead)

00:07:42 – The video upload trap: "only I can do this"

00:09:14 – The four-step founder loop in Claude Code

00:12:17 – The codebase is the knowledge base, not Notion

00:13:33 – If you won't move to Claude Code, this isn't for you

00:15:06 – Give your team (humans and agents) full access

00:16:21 – Three roles at Pirate Skills: Ben, Christina, ChristAIna

00:18:11 – Your Monday morning: 10 tasks, one skill, one loop

DEEP DIVE LINKS

📖 Full Article: https://pirateskills.com/insights/ai-founder-mode

📑 Source material:

• Paul Graham — Founder Mode (Sept 2024): https://paulgraham.com/foundermode.html

• Brian Chesky — Invest Like the Best EP 470 (May 5, 2026): https://open.spotify.com/episode/5ODjrzaczFgI5I0nVtkuzb

• Jack Dorsey — From Hierarchy to Intelligence (Mar 31, 2026): https://block.xyz/inside/from-hierarchy-to-intelligence

🎧 Spotify: https://open.spotify.com/show/7kFl0BJiOYih8yqQTRR9SN

🍎 Apple Podcasts: https://podcasts.apple.com/us/podcast/build-grow-repeat/id1870561487

🗓️ All upcoming events (free): https://pirateskills.com/events

💬 Connect with me on LinkedIn: https://www.linkedin.com/in/bensufiani/

What if 80% of every weekend hackathon was already built before you sat down?

Every new agentic SaaS project starts at the same wall: auth, database, payments, deploy, AI streaming, env-var roulette. By the time the magic interaction is in reach, the weekend is over. The Pirate Launch Kit is a skill that scaffolds that whole 80% from an empty folder to a deployed, working agentic app shell — landing page, login, streaming chat — on your own URL, before you write a single line of product code.

The bigger shift behind it: SaaS itself is changing shape. PowerPoint gave humans tools. Google Slides put those tools in the cloud. Gamma and Claude Design crossed the line — the human is no longer the worker, the human is the supervisor. The new shape of SaaS is a UI built around an agent that does the work. The Kit is opinionated about that shape and ships with the primitives it needs.

The video walks the four layers — Frontend, Backend, Agent, Builder — tool by tool, with the why behind every pick. Then a full live demo in VS Code: empty folder, install the skill, answer one question, and watch a private GitHub repo, a Vercel project, env vars, password gate, AI Gateway, and a streaming chat agent appear on a live URL — in minutes, not a weekend.

This is not theory. The Kit is being built in public. Send feedback after you try it.


IN THIS EPISODE

00:00:00 – The 80% scaffolding tax
00:00:33 – Demo preview: empty folder to deployed agentic app
00:02:30 – Bridesmaid hackathon + the Forge pattern
00:05:00 – The new shape of SaaS (PowerPoint → Gamma → Claude Design)
00:08:50 – Layer 1: Frontend (Next.js, Tailwind, shadcn/ui, AI Elements, Playwright)
00:13:43 – Layer 2: Backend (Vercel, Clerk, Neon + Drizzle, Stripe, Resend, PostHog)
00:18:50 – Layer 3: Agent (Vercel AI SDK, AI Gateway, Workflow SDK, Chat SDK)
00:22:55 – Layer 4: Builder (Claude Code, Vercel plugin, MCP servers, CLIs)
00:28:55 – Live demo: empty folder to live agentic app
00:42:45 – Phase 2 preview + how to send feedback


DEEP DIVE LINKS

📖 Full Article: https://pirateskills.com/insights/pirate-launch-kit

🧪 Vibe Coding Cologne — Wed May 6th: free monthly meetup (offline + Zoom): https://pirateskills.com/events/vibe-coding-cologne/2026-05-06

⚓ Pirate Lab — Wed May 13th: free Zoom session, this week's deeper dive on what we shipped: https://pirateskills.com/pirate/lab/2026-05-13

🎧 Spotify: https://open.spotify.com/show/7kFl0BJiOYih8yqQTRR9SN

🍎 Apple Podcasts: https://podcasts.apple.com/us/podcast/build-grow-repeat/id1870561487

🗓️ All upcoming events: https://pirateskills.com/events

💬 Connect with me on LinkedIn: https://www.linkedin.com/in/bensufiani/

Your SaaS has a ceiling it will hit — and it's already programmed in by two numbers.

Back from MicroConf Portland with the slide that reframed everything: Max MRR = New MRR ÷ Churn. Jason Cohen showed Kit's real revenue chart flattening at $150k/month, and underneath the math was the two-sided formula I'd been teaching for years without realizing it. Build · Grow · Repeat is not a tagline — it's the calculator.

This episode walks through both halves with the speakers who made them concrete. Defense: Jason Cohen on obsessing about churn, Alex Pham on pricing as a retention lever (C.R.E.A.M.), Anthony Eden on operational excellence as a silent moat. Offense: Gia Laudi's painkiller positioning, Nick Disabato's Pain-Dream-Fix homepage (including a /nickdfy skill I built that rewrote our homepage in his style), Amanda Natividad on publishing for zero clicks, Einar Vollset on the 20% growth-rate cliff that halves your exit multiple.

The personal arc underneath: before vibe coding I could only pull Grow levers. Now the Build half is reachable too — and the math says that's where the biggest unlock sits.

Live from MicroConf Portland. Bring your own numbers.

IN THIS EPISODE

00:00:00 – Back from MicroConf Portland

00:00:48 – Two halves, one ceiling: Jason Cohen's Max MRR formula

00:04:19 – Build · Grow · Repeat: the two halves vibe coding unlocks

00:07:23 – Defense: obsess about churn first

00:10:10 – Alex Pham: pricing as a retention lever (C.R.E.A.M.)

00:12:18 – Anthony Eden: operational excellence as a silent moat

00:13:07 – Offense: Gia Laudi on painkiller positioning

00:14:45 – Nick Disabato's Pain-Dream-Fix (and the /nickdfy skill)

00:16:32 – Amanda Natividad: publishing for zero clicks

00:18:39 – Einar Vollset: growth rate as a valuation lever

00:22:06 – Run the Max MRR calculator on your own business

DEEP DIVE LINKS

📖 Full Article: https://pirateskills.com/insights/build-grow-mrr

🧪 Forge Preview — Wed Apr 22, 18:00 Berlin: Free live walk-through of Builder Forge + Growth Forge — see which half of the ceiling to push first: https://pirateskills.com/pirate/forge/preview/2026-04-22

🎧 Spotify: https://open.spotify.com/show/7kFl0BJiOYih8yqQTRR9SN

🍎 Apple Podcasts: https://podcasts.apple.com/us/podcast/build-grow-repeat/id1870561487

🗓️ All upcoming events: https://pirateskills.com/events

💬 Connect with me on LinkedIn: https://www.linkedin.com/in/bensufiani/

What if the "Zero-Human Company" everyone's hyping is actually a trap?

The frustrations behind it are real – managing humans is distracting, most early hires can barely pull their own weight, and getting to traction is often easier with 3 people than 15. But removing all humans strips out the ambition, the taste, the soul.

This episode lays out the counter-position: the Augmented-Human Company. A small, focused crew where AI removes what's holding each person back – so they can give the contribution only they can give. Real examples from Papermark, Typefully, Oleve, and Gamma. Plus the thread from Rob Walling's "Start Small, Stay Small" to Karpathy's Iron Man Suit metaphor.

Published live from MicroConf Portland – the very community built on intentional smallness.

WHO THIS IS FOR

• A founder tempted by the "zero human" hype but sensing something's off
• A bootstrapper who wants to stay small without scaling back ambitions
• A vibe coder building their first product and wondering how far a tiny team can go
• Someone who's tried hiring and found it harder than doing it with fewer people

IN THIS EPISODE

• 00:00:00 – The Zero-Human Trap
• 00:00:46 – Why the frustrations behind it are real
• 00:02:00 – The brutal trade-off small teams always faced
• 00:03:06 – Building alone gets old fast
• 00:03:49 – Karpathy's Iron Man Suit metaphor
• 00:04:52 – What an Augmented-Human Company actually looks like
• 00:06:50 – The intellectual lineage: Walling → Naval → Altman → Shipper
• 00:10:04 – Real examples: Papermark, Typefully, Oleve, Gamma
• 00:12:54 – The Repeat Layer: Build, Grow, Repeat
• 00:14:49 – The new way to build companies
  
  

DEEP DIVE LINKS

• 📖 Full Article: https://pirateskills.com/insights/augmented-human
• 🧪 Builder Lab – Apr 15: https://pirateskills.com/build/lab/2026-04-15
• 🎧 Spotify: https://open.spotify.com/show/7kFl0BJiOYih8yqQTRR9SN
• 🍎 Apple Podcasts: https://podcasts.apple.com/us/podcast/build-grow-repeat/id1870561487
• 🗓️ All events: https://pirateskills.com/events
• 💬 LinkedIn: https://www.linkedin.com/in/bensufiani/

What happens when you stop treating AI as an assistant — and start onboarding it like a real teammate?

The leap from "AI power tool" to "AI teammate" had nothing to do with smarter models. The missing piece was infrastructure and onboarding — a workspace, communication channels, tool access, and clear responsibilities. The same things you'd invest in for a human hire.

Meet ChristAIna — our first OpenClaw agent. She runs on a €12/month Hetzner server, has access to Slack, Gmail, Calendar, GitHub, Linear, Attio, Vercel, and Resend, and she works while we sleep. Morning briefings at 7:30 AM, CRM pipeline management, code deploys, email monitoring, and nightly memory consolidation — all on a schedule.

This is the full setup after one month of real usage: the tools, the skills, the memory system, the 7 Slack channels, the automated routines — and everything that still breaks. Not theory. A two-person team that added an AI as crew member #3.


IN THIS EPISODE

00:00:00 – Meet our first OpenClaw teammate
00:00:44 – The problem: assistant vs. teammate
00:02:09 – The crew: Ben, Christina, and ChristAIna
00:03:19 – The foundation: OpenClaw + Hetzner + openclaw-infra
00:05:50 – Onboarding an AI teammate: tools, skills, and memory
00:13:48 – Coding: Claude Code on the server
00:15:30 – What she actually does: 7 Slack channels
00:16:43 – The morning briefing deep dive
00:20:51 – Automated routines: cron jobs and heartbeats
00:23:59 – What still breaks
00:25:47 – Where "Repeat" gets interesting


DEEP DIVE LINKS

📖 Full Article: https://pirateskills.com/insights/first-openclaw-teammate

🎧 Spotify: https://open.spotify.com/show/7kFl0BJiOYih8yqQTRR9SN

🗓️ All upcoming events: https://pirateskills.com/events

💬 Connect with me on LinkedIn: https://www.linkedin.com/in/bensufiani/

What happens when a non-technical couple locks themselves into a 48-hour hackathon with an engagement ring and an idea?

Most people think building an AI product requires months of learning or a technical co-founder. Christina spent weeks scrolling through Bridebook and WedStay — generic catalogs, no personalization, nothing that understood her taste. Then one shower conversation turned frustration into a product vision: what if you could talk to a wedding planner that actually gets you?

This is the full behind-the-scenes breakdown of every product decision and tech choice — from picking next-forge over a simple template, to the moment fal.ai generated a visualization of our dream wedding and Christina was astonished. Ten product needs, ten tech decisions, each explained from both the non-technical and technical perspective.

Honest disclaimer: we didn't follow our own Pirate Launch Plan. We set up Stripe and auth too early. The visualization feature worked 30 minutes before the deadline. And production deployment almost broke everything on Saturday night.

But we made it: https://www.bridesmaid.love/de is live.

WHO THIS IS FOR

- A founder who's stuck between "I use AI tools" and "I shipped a product with AI"
- A non-technical builder wondering what it actually takes to go from idea to live product
- A vibe coder preparing for a hackathon who wants to see what's realistic in 48 hours
- A couple or team where one person has the domain knowledge and the other has the tech instinct

IN THIS EPISODE

00:00:00 – How we built an AI startup in 48 hours
00:00:34 – Scratch your own itch: Christina's wedding app frustration
00:03:15 – The product we wanted to use: chat-first wedding planning
00:05:54 – 10 product decisions, 10 tech choices
00:07:15 – AI SDK + AI Gateway: a conversation that does things
00:09:57 – Real-Time Scribe: talk, don't type
00:10:39 – fal.ai: show me my dream wedding
00:12:08 – Clerk, Stripe, Claude Code: the rest of the stack
00:14:54 – Christina's hackathon experience (in her own words)
00:17:12 – The honest weekend recap: what went wrong
00:19:46 – What's next: events, Builder Forge, and your turn


DEEP DIVE LINKS

📖 Full Article + All 10 Tech Decision Cards: https://pirateskills.com/insights/startup-in-48-hours

🏗️ Builder Forge Q2 April 1st — Build your own AI product with guided support: https://pirateskills.com/build/forge

🎧 Spotify: https://open.spotify.com/show/7kFl0BJiOYih8yqQTRR9SN

🍎 Apple Podcasts: https://podcasts.apple.com/us/podcast/build-grow-repeat/id1870561487

🗓️ All upcoming events: https://pirateskills.com/events

💬 Connect with me on LinkedIn: https://www.linkedin.com/in/bensufiani/

Everyone's using AI. Nobody knows where they actually stand.

There's no shortage of AI tools, tutorials, and hype. But there's no map. No way to know: am I a beginner who thinks they're advanced, or someone further along than I realize? And what should I actually learn next?

I couldn't find a clear progression path, so I built my own. The AI Skill Path is a 6-level framework — from Landlubber to Admiral — that shows you where you are, what you already unlock at that level, and what it takes to reach the next one. Each level covers both the Build track and the Grow track.

This isn't about collecting tools. It's about knowing which capabilities matter at your stage — and ignoring everything else until you get there.

WHO THIS IS FOR

A founder who uses ChatGPT daily but isn't sure what the next level looks like

A vibe coder who ships apps but hasn't touched AI marketing yet

A marketer experimenting with AI content but not building anything technical

Someone drowning in AI FOMO who wants a clear, honest progression path

IN THIS EPISODE

00:00:00 – The problem: there is no map

00:01:11 – The principle underneath every level: relationship to initiation

00:02:39 – The 6 levels overview

00:02:52 – Level 1: Landlubber — the ChatGPT user

00:04:36 – Level 2: Deckhand — persistent context and better prompting

00:07:19 – Level 3: Navigator — vibe coding simple tools and automations

00:09:30 – Level 4: Corsair — owning your stack and building real SaaS

00:13:43 – Level 5: Captain — proactive agents working for you

00:17:29 – Level 6: Admiral — orchestrating multi-agent systems

00:20:36 – Do you really need to go all the way?

00:24:48 – Find your level and start from there

DEEP DIVE LINKS