From Vibe to Viable

Event recap · April 29, 2026 · The Beacon, Antwerp

You've built something. It works. Now what?

That's the question we set out to answer on April 29 at The Beacon. Three speakers, three different angles on the same wall every vibe-coded founder eventually hits.

Talk 1: We built the MVP in 8 hours. Then the real work started.

Thomas Butstraen, co-founder at Maurice & Nora

Thomas had an idea: connect students with families who need non-medical care. Gardening, school pickups, support for the elderly. The cheapest quote he got to build it? 100k. For an MVP.

So he opened Lovable instead. First prompt in March 2025. Platform live 8 hours later.

He went on to finish 4th worldwide in the Lovable Shipped hackathon out of 5,800+ entries, flew to Sweden to meet the Lovable founding team, and now runs a platform with 3,000+ students and 2,000+ clients across Flanders and Brussels.

They've graduated from Lovable. They'll never graduate from AI. Three engineers, ten-person output. Native mobile apps, WhatsApp integrations, AI voice agents for student screening.

"Two years ago this was never possible. A non-technical person building a platform, getting users, getting validation. Really, really crazy."

Three things Thomas wished he'd known earlier:

• Talk to your LLM, don't type at it. Voice input is faster and picks up context better. He uses Superwhisper.

• Validate fast on a small budget. Don't over-invest before you have proof.

• Build your own design system. AI UI looks like AI UI. Design is how you stand out.

Watch it here.

Talk 2: 180 hours, 1.300 euros, zero investors. And it works.

Sebastiaan Schillebeeckx, Partner & CPO at Dashdot · Founder of Bandsailor

Does it make business sense to build software for DIY bands with no budget? No. Sebastiaan built it anyway.

Bandsailor is a band management app: shows, finances, merch, venues. A back office for bands that never had one. Built evenings and weekends from October 2025. One person. Lovable, Cursor, Resend. 1.300 euros a year to run.

"Technology can really open up new markets. Target groups that were never interesting economically, suddenly they're viable because the cost to build dropped by 90%."

First pilots in January. Paying customers by April. 30 bands, 100 users, 250 shows in the system. Not massive numbers. But it's real, it's useful, and it cost almost nothing to find out.

He was honest about the limits too. No staging environment, everything on production, security audit largely ignored. Fine for now. Not fine when a booking agency with 100 bands wants to onboard. That's exactly where the next talk picks up.

Watch it here.

Talk 3: Four pillars to build trust around your Lovable app.

Maarten Jansen, CEO & Partner at Dashdot

Things break in ways you can't debug. Pages reload. API calls fail. You're fixing more than you're building. That's the Vibe Coding Vortex. The exit is structure.

Maarten framed it as a trust problem. You don't trust your own code, your users are getting skeptical, your investors have questions. The fix is the same for all three.

Ownership: Connect GitHub, get your code out of Lovable, do one quick security review. Most issues are cheap to fix if you catch them early.

Reliability: Split dev and production. Keep Lovable for development, run a separate environment for real users via GitHub branches and a dedicated Supabase instance. Break things in dev. Never in prod.

Control: Add automated testing and CI/CD. Every change triggers a safety check before anything hits production.

Insights: PostHog for product analytics, Sentry for error tracking. Know what's broken and how many users it's actually affecting.

You don't need all four today. Maarten's phase breakdown:

• Validation: GitHub connection and one security review. That's it.

• Early Scale: Split your environments. Stop shipping into production every time you make a change.

• Full Product: Add testing, CI/CD, and observability. Bring in engineering support.

Watch it here.

The takeaway

Vibe coding changed what's possible. A non-technical founder can go from idea to live product in a day. A solo designer can build and ship a real app after hours. That's not hype, it's what happened on this stage.

But the vibe has limits. When the stakes go up (real users, sensitive data, investors) "it mostly works" stops being enough. The answer isn't to start over. It's to know which steps to take and when.

That's what this evening was about. And we're doing it again.

Next up: agent engineering. How to build your own custom AI agents on top of what you've already got. Stay tuned.