The Modern MBA
How to get from an idea or a problem to your first paying customers. The full playbook from the talk.
The hardest part isn’t building. It’s starting.
In three years everything about making software changed. 2023 was every line by hand. 2024, AI as autocomplete. 2025, agents doing the creative work. In 2026 the job is 90% shaping specs, 10% reviewing, and zero lines written by hand.
So the bottleneck moved. Code isn’t the moat anymore, and neither are screens. Knowing the customer and shaping the work is the moat, and that’s what a room full of product and project people already does for a living.
You become responsible for building complete products. AI fills in everything that isn’t directly in your background.Marc Andreessen, 2025
In ten years the word coder may be gone, and the word builder will be everywhere. That was the talk: what a builder actually does, and how to become one by Monday. It came in three parts.
Part one: the playbook
Four steps from an idea to your first paying customer. No fundraising, no cofounder hunt, no two-year build.
1. Don’t invent. Eavesdrop.
People are already telling you what to build. There are four places to listen.
- Communities. On Reddit, Discord and Slack groups, people are begging for software that doesn’t exist.
- Boring B2B workflows. What are people paying freelancers to do by hand, every week?
- AI trend scanning. Pipe X, YouTube and Reddit into Claude and ask which complaint repeats.
- Geo-niches. High demand, and bad or no software. Macedonia is full of them. GemFlow started with 200+ jewelry stores and no good tool.
Try searching: site:reddit.com "is there a tool for"
2. Copying is smart.
A whole YouTube genre exists about this. Search “Starter Story” and you’ll find hundreds of them: $20K a month, $35K a month, $100K in twenty-four hours. They’re not unicorns. They’re boring SaaS, copied and aimed at a new niche.
The move: find a proven model. Apply it to a niche you understand. Ship it better, smaller or local.
3. The minimum viable funnel
Not an MVP. Build the funnel before the product. If one stage is broken, there’s no revenue.
- Traffic. Content, outreach, forums, ads.
- Landing page. The offer, how it works, a demo, one call to action.
- Lead magnet. Trade value for the email, or a dollar.
- Nurture. Email and SMS. Stay close.
- Offer. The buy button. Your first dollar.
4. Sell the work, not the tool.
SaaS sells a seat, and the customer does the work. Service-as-Software sells the outcome, and the agent does the work.
Part two: the Modern MBA
A company used to be six departments, six payroll lines and two years before anything shipped. In 2026 it’s one driver, one harness and five rails. That’s the whole company, on one page.
- Your team: Claude Code. Engineering, design, writing, research. With MCPs, every tool plugs into it. It’s the agent harness.
- Your brain: Obsidian. Every customer call, every spec, every idea. The second brain that feeds the agent.
- Your stack: Rails and Hetzner. One opinionated starter kit that goes from 0 to millions. Stop debating stacks and ship.
A talk in a room only goes so far, so I recorded how this works day to day: Claude Code as the harness, with the five rails driven from it.
Part three: Monday morning
Do these three things and you’re a builder. Skip them, and this was just a talk.
- Write down your ICP. One sentence: a specific niche and a specific person.
- Open the roadmap. It walks you through the funnel, step by step.
- Talk to five people in that niche this week. Not a survey. Conversations. Buy the coffee and listen hard.
From the room
Photos from PM², Skopje, May 16, 2026.
Want me to give this talk at your event? Email me at naum@builders.mk.