Elias
Möller.
Currently shippingsites.
Twenty, based in Germany. My brother and I run two companies and ship most of what they need between us — because AI now does the work a small team used to.
- Current
- Co-founder × 2
- Stack
- AI · first
- Based
- Germany
- Status
- Shipping
I'm twenty. I run two companies. AI is how one person ships what used to take a team of five.
I started Morgentau Marketing — a marketing agency — and Morgentau Garten, a garden & landscaping business. Both are real operations with real customers in Germany, not side projects pretending to be companies.
Both companies run on an AI-native stack I built myself — the sites, the internal tools, the automations, the back-office glue. What used to need a team I now do with models that write, plan, and iterate faster than I can type.
- Age
- Twenty
- Based
- Germany
- Companies
- Two · live
- Stack
- AI · first
ignore the
existential dread
on scroll 04.
— e
A different curve.
What AI can do, 2020 → 2030.
Solid: what shipped. Dashed: what's coming.
Everything after 2026 is guessing. Safe bet — up and to the right.
Two companies. Both live, both mine.
The work below is stuff I actually run, not decks. More will land here as I build it.
The tension
I live in.
I build with AI every day. I also think about where this ends up. Two futures — both plausible, neither guaranteed.
The promise — abundance.
- 01Expert-level intelligence gets priced like electricity.
- 02The tutor only elites had → any kid, any village.
- 03Diseases we can't cure → solved problems.
- 04The ceiling on what one person can do quietly disappears.
The cost — a species without a job.
- 01Most work stops being economically necessary. Most people stop having a paycheck.
- 02Two hundred thousand years of mattering for our brains. Then we don't.
- 03We build something a thousand times smarter than us. We don't know how to turn it off.
- 04Outsource your thinking. The meaning leaks out with it.
Both are possible at once.
I'm not planning on being blind about it.
“Speed is the product. The rest is just taste and follow-through.”
AI isn't a tool I reach for — it's where every project starts. If a model can do it, a model does. What's left for me is the part only I can do: taste, direction, and the calls about what's actually worth building.
Both companies have customers who pay invoices, not users chasing a free tier. That constraint is the best filter I've found for what's actually worth building.
Tutorials didn't teach me this stack. Deadlines did. Ship the ugly version, watch where it breaks, fix that, repeat. Most "skill" is just a record of what you've already broken.
Open to interesting conversations.
Not a pitch inbox. If what I build lines up with something you're thinking about — send a note. Shortest route is Instagram.