@AnnaBlazickova I hit the gym regularly to stay in shape, and occasionally play badminton or tennis. The gym is mostly for the physical side, while the sports are more about connecting with people over a shared interest, which is essential as a founder.
I interviewed the people, preferably in person over coffee. Listening about their pain taught me more than anything.
One I solved the basic pain points, I tried to understand whatโs absolutely dealbreaker features and only built those around the solution.
Listen about the pain, sell the solution, donโt discuss how.
Running two companies taught me a hard lesson.
I was the bottleneck.
Not because I was slow. Because every decision waited for me.
The fix wasn't more hours.
It was writing down how I decide, so the team can decide without me.
@jasonfreedman Founder side: a fast no is a gift. It frees the calendar and the head. The slow maybe is what hurts, because you keep planning around money that was never coming.
@deepakabbot Most of it is in the back office, which is where it should start. Customers notice AI only when it removes a step they hated, not when a chatbot appears. The visible wins will come from AI that finishes a task end to end, not one that just answers questions.
For modern AI native tools, the value lives in the delivery loop and presenting, not in the generator.
For Anthemius, people felt the time-saving when we took generation in the background.
@samongaro_ I talked to 100 users before making it, which took me 6 months.
Making the actual app once plan was ready took me 4-5 months. But honestly, I had some help for some part.
An architect friend's team told me generative AI was useless for their work.
Not because it was bad. Because nobody can sign off on a layout that ignores building code.
Here's what they actually do on every project:
1/ Lay out the floor in a drafting tool.
2/ Rebuild it in 3D software for the client.
3/ Render it.
4/ Count every material in a spreadsheet.
Then the client says "move this wall" and all four start over.
My take: drawing isn't the hard part. Keeping four copies in sync is.
So that's what we built Anthemius around.
Mark the zones, corridor and fixed rooms. Get a layout checked for clearances, egress and occupancy. Dress it in 3D. Walk the client through it live.
And the material list updates with every change.
#BlackFriday access is open: https://t.co/QPDVrrC1HR
Hiring doesn't fix a problem you haven't named.
Before a hire, I write one line: what will this person own that's broken today?
If I can't write it, the problem is me, not headcount.
https://t.co/BUTwgw8rd6
@haleymbryant Seedstrapping 2.0 needs a different scoreboard too. Raise $300k to reach profit and you should be judged on payback and margin, not on whether you can 100x. Founders must pick the scoreboard before the round, not after.
@bamitav Walkthroughs are when clients finally get it. Then they ask to move a wall. If layout, 3D and the material list don't update together, you're back to redrawing. That's the gap I'm building Anthemius around.
Designers redraw the same room in 4 tools each time a client changes their mind.
Drafting. 3D. Rendering. A materials spreadsheet.
I'm building Anthemius to make it one loop: lay it out, dress it in 3D, walk the client through it, get the material list.
https://t.co/VYE2bQOYaU
I'm Vedang, a founder in Mumbai.
I'm building Anthemius: say a space brief out loud, get floor layouts that meet building code. 2D, 3D and BOM, everything.
I post what I learn about AI products and building from scratch. Follow along if that's you.
https://t.co/VYE2bQOYaU
@Polycam3D Genius for existing rooms. The step before is just as broken: going from a brief to a layout that meets code. That's what I'm building with Anthemius, say the space brief out loud and get code-aware layout options. Scan + plan feels like the full loop.