DiscoverBuild. Grow. Repeat.
Build. Grow. Repeat.
Claim Ownership

Build. Grow. Repeat.

Author: Ben Sufiani

Subscribed: 4Played: 105
Share

Description

Build. Grow. Repeat. is the podcast for skilled professionals and early-stage founders who want to build a real product and grow predictable revenue — without hustle culture, or “get rich quick” fantasies.

This show helps you close the gap between building a product with vibe coding and growing with growth marketing by treating them as one craft. Every episode is designed to give you something you can actually use: a clearer decision, a tighter customer journey, a better experiment, or a concrete next step you can run this week.

What you’ll learn here:

  • Build faster (without needing to know how to code): vibe code real software and offers quickly — and validate what matters before you over-invest.
  • Grow calmly and predictably: positioning, funnels, pricing, and experiments that turn attention into revenue without persuasion theater.
  • Run on signals, not just vibes: activation, retention, and product–market-fit indicators that tell you what to double down on — and what to stop.
  • Use AI as leverage: vibe coding and AI-native workflows that speed you up while keeping human judgment in charge.
  • Commit without gambling: how to choose one direction and test it through small, reversible bets that produce proof.

You’ll get a mix of solo deep dives, builder interviews, and occasional teardowns of customer journeys and decisions that looked “logical” on paper — but fail in real life.

About the host: I’m Ben Sufiani, founder of Pirate Skills. I’ve been building and growing products for over 15 years across startups, consulting, and hands-on growth work. Pirate Skills exists to help builders turn ideas into evidence — by shipping, learning, and repeating the loop until something holds. If you want to go deeper, check out Pirate Skills at pirateskills.com.

Build. Grow. Repeat.

Cheers,

Ben

44 Episodes
Reverse
What would you build in a weekend if it never had to find a buyer? One Friday night, Claude Code interviewed me about a Latin tutor for my daughter Noëlle – and by Saturday morning it was live, for her alone.Eleven and a half hours from request to live app, and the agent worked for about two of them. Twenty questions instead of a spec, 64 artboards before a line of code – and then eight minutes with one real user that taught me more than the whole night before.This episode walks through the app – lessons, a chat with Noctua the owl, a skill tree – and the method behind it: let the agent interview you, approve and go to sleep, design first, record your one user while they test it, and make yourself the first user. Plus the parent view that shows the most common mistakes, and why every product should have one.What to build at a hackathon? Something good for someone you love.WHO THIS IS FORA parent with a weekend and a kid who is stuck on somethingAnyone heading to a hackathon and wondering what to buildA builder whose first version looked great until a real user tried itA non-developer who wants to build something personal with a coding agentIN THIS EPISODE00:00:00 – Hack something for someone you love00:01:09 – Why not just a ChatGPT session00:03:23 – A tour of the app00:05:54 – Let the agent interview you00:08:50 – Design first: 64 artboards overnight00:10:43 – The stack: Claude Code, Vercel eve, Neon00:13:37 – Watch your one user00:16:16 – Make yourself the first user00:20:51 – The parent view: where users get stuck00:22:07 – Your weekend, step by step00:23:19 – More to read, and the Vibe HackathonDEEP DIVE LINKS📖 Full Article: https://pirateskills.com/insights/hack-with-love📱 Mobile App in a Weekend: https://pirateskills.com/insights/mobile-app-in-a-weekend🔨 Watch Me Build, the live session: https://pirateskills.com/insights/feedback-to-feature🛠️ Vibe Hackathon Cologne – 20 to 22 November, €60 with the community code: https://pirateskills.com/events/vibe-hackathon-cologne/2026-11-20?code=VHC-COMMUNITY🧭 AI Compass – 25 November: https://pirateskills.com/aicompass/2026-11-25🏴‍☠️ Vibe Coding Cologne – 2 December: https://pirateskills.com/events/vibe-coding-cologne/2026-12-02🧪 Pirate Lab – 9 December: https://pirateskills.com/pirate/lab/2026-12-09⚓ Pirate Forge: https://pirateskills.com/pirate/forge🗓️ All upcoming events: https://pirateskills.com/events💬 Connect with me on LinkedIn: https://www.linkedin.com/in/bensufiani/
We all grew up on games that showed us exactly where we stand – levels, skill trees, an auction house with clear yardsticks. Then we run our businesses on numbers scattered across multiple sources, and nothing tells us the next best move.When we put our lead numbers on a board for the first time, it showed two weeks with no ads running: 11 and 26 leads. The week after: 127, our best in twelve weeks. The errors board counted 104 errors in four weeks and a fixer whose queue went from 47 items to zero – with nine decisions left waiting on me.In this episode I walk through my two live boards – lead flow and errors – each with a quest log that hands the next step to Claude Code. Then the chess position from the 2013 World Championship: once you know the better move, you never play the second-best one. In business we do it all the time. And how to draw your own board: a floor, a healthy zone, a streak, and an agent that picks the highest-value move.WHO THIS IS FORA founder who grew up on games and runs the company without a boardA founder whose ads and follow-ups slip through a busy weekAnyone with a coding agent looking for an afternoon project they will look at every dayCasual or min-maxer – the board works from either endIN THIS EPISODE00:00:00 – World of Reallife – my childhood dream00:00:21 – Two live boards: lead flow and errors00:01:15 – Why games feel better: WoW and the Air Force years00:02:46 – Designing the board with Claude00:03:56 – Two weeks without ads00:04:43 – Nobody plays the second-best move00:06:46 – Drawdown and the best opportunity00:08:44 – Min-maxer or quality of life00:11:30 – The admin boards00:12:55 – The product scout and runner00:15:03 – Draw your own board00:17:30 – Don't just be busy00:18:27 – Pirate Lab, AI Compass, Vibe Coding Cologne, Hackathon, ForgeDEEP DIVE LINKS📖 Full Article + the five steps: https://pirateskills.com/insights/world-of-reallife🧪 Pirate Lab – 11 November: https://pirateskills.com/pirate/lab/2026-11-11🛠️ Vibe Hackathon Cologne – 20 November: https://pirateskills.com/events/vibe-hackathon-cologne/2026-11-20🧭 AI Compass – 25 November: https://pirateskills.com/aicompass/2026-11-25🏴‍☠️ Vibe Coding Cologne – 2 December: https://pirateskills.com/events/vibe-coding-cologne/2026-12-02⚓ Pirate Forge: https://pirateskills.com/pirate/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/
I come from marketing and have never been a developer. Jens Dressler has written software for more than 20 years – semiconductor test equipment, then CTO of the event app Vamos, then CTO at Fanmiles after it bought Vamos. We both spent years trying to build for and sell to developers, and we both ended up somewhere else: helping people who don't code get AI to build real things for them.Jens tells what selling to CTOs was really like – "I want more control" – and what changed when he started talking to people without a development background: "It was not about the stack. … It was really about, I want to achieve this. Can you help me?" We talk about the hard parts non-developers never see coming (data privacy, auditability, race conditions, the happy path), why he thinks no-code was too restrictive, how Turbofy hides the lower layers so you only think about your business logic, and where a chat builder stops and Claude Code or Codex takes over.WHO THIS IS FORA founder or marketer without a developer background who wants to build real softwareAnyone stuck between a no-code tool and a developer quoteA technical founder who keeps getting asked "can you build my idea?"A builder deciding between a chat builder and Claude CodeIN THIS EPISODE00:00:00 – Two sides, one problem00:01:06 – From chip-test software to Berlin startups00:02:26 – Why selling to developers is so hard00:04:31 – Great ideas, 500-euro budgets00:06:23 – My side: the CTO who said "almost ready"00:08:23 – A crash at lunch, and why developers overbuild00:10:18 – Velocity of value: frame it as an experiment00:13:55 – Straight into Claude Code – every bucket at once00:15:11 – What non-developers miss: privacy, auditability, race conditions00:21:53 – How Turbofy hides the lower layers00:23:15 – Developers move up: raising the floor and the ceiling00:26:47 – Why code, not no-code00:27:52 – The December moment: 95% of the code with AI00:31:20 – A Turbofy app in three prompts00:33:05 – Where chat builders stop and Claude Code takes over00:39:14 – What to build this week00:41:17 – Where to find JensDEEP DIVE LINKS🚀 Turbofy: https://turbofy.com📚 Turbofy docs: https://docs.turbofy.com💬 Jens on LinkedIn: https://www.linkedin.com/in/jensdressler/🗓️ All upcoming events: https://pirateskills.com/events🎧 Spotify: https://open.spotify.com/show/7kFl0BJiOYih8yqQTRR9SN🍎 Apple Podcasts: https://podcasts.apple.com/us/podcast/build-grow-repeat/id1870561487💬 Connect with me on LinkedIn: https://www.linkedin.com/in/bensufiani/
At a Product Night in Cologne, people tried my app Zawq and left five post-its of feedback. Feedback is a gift and a trap: there is a gap between what you want to build and what a new user wants.In this episode I sort the notes, pick one and build it live with Claude Code: letting users choose the AI model that writes their content. You'll hear how I decide which feedback to follow, and how I apply the same filter to my agent – it looks around the code first, asks me questions, and I go my own way on most of its recommendations. Then it builds, tests itself on iPhone and iPad, and hands me a test guide.This is a real build, not a demo. My part of the two hours was about ten minutes.WHO THIS IS FORA founder who gets more feature requests than they can buildA vibe coder without a developer background who wants to see a real sessionAnyone building an AI product who is wondering whether to let users pick the modelA builder who wants to listen to users without giving up their product tasteIN THIS EPISODE00:00:00 – Five post-its from one Product Night00:03:37 – The one I picked: model choice00:04:52 – Your taste is the filter00:06:52 – Sort it before you build it00:08:27 – Briefing the feature in Claude Code00:14:31 – Use the same filter on your agent00:28:07 – Linear, and letting it grind00:29:36 – Let the agent prove it first00:35:40 – The test guide – my turn to click00:44:01 – Two hours of agent work, ten minutes of mine00:46:22 – Upcoming events and the Pirate ForgeDEEP DIVE LINKS📖 Full Article + the five steps: https://pirateskills.com/insights/feedback-to-feature🧭 AI Compass – 28 October: https://pirateskills.com/aicompass/2026-10-28🏴‍☠️ Vibe Coding Cologne – 4 November: https://pirateskills.com/events/vibe-coding-cologne/2026-11-04🛠️ Vibe Hackathon Cologne – 20 November: https://pirateskills.com/events/vibe-hackathon-cologne/2026-11-20⚓ Pirate Forge: https://pirateskills.com/pirate/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/
Your agent can build almost anything now. So why did my projects still move in bursts?Because the agents were never the slow part. I was. With several projects and a full calendar, a project only moved when I sat down with it – and sometimes that meant a whole week without progress. In this episode I show what changed it: one roadmap per project. Two jobs stay with me: deciding what gets built next, and reviewing it before a customer sees it. Everything in between is the agent's.You'll see it in practice: one idea for Customer Magic that became a roadmap item, a brief the agent wrote itself after talking to my compliance session, and a finished trust page I only had to review. Then why the deciding stays human, what a roadmap holds in four stages, and three places to keep it – an artifact in Claude Code, a page in your codebase, or a tool like Linear.This is not theory. I run four projects this way and check in about twice a day.WHO THIS IS FORA founder with several projects where one always waitsA vibe coder whose agent builds fast, in no clear directionAnyone who starts every session by explaining again what comes nextA small team that wants steady progress without a project managerIN THIS EPISODE00:00:00 – Your agent needs a roadmap from you00:00:36 – The bottleneck is you00:01:38 – Two jobs stay with you00:02:04 – One idea, from message to finished page00:05:07 – Why deciding stays human00:06:58 – What a roadmap holds00:07:51 – Three places to keep it00:11:17 – Deciding what's next, twice a day00:13:31 – Give your agent a roadmap00:14:48 – Upcoming events and the Pirate ForgeDEEP DIVE LINKS📖 Full Article + the starter prompt: https://pirateskills.com/insights/agent-roadmap🧪 Pirate Lab – 21 October: https://pirateskills.com/pirate/lab/2026-10-21🧭 AI Compass – 28 October: https://pirateskills.com/aicompass/2026-10-28🏴‍☠️ Vibe Coding Cologne – 4 November: https://pirateskills.com/events/vibe-coding-cologne/2026-11-04🛠️ Vibe Hackathon Cologne – 20 November: https://pirateskills.com/events/vibe-hackathon-cologne/2026-11-20⚓ Pirate Forge: https://pirateskills.com/pirate/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/
loading
Comments