Working on a side project at this scale is uncomfortable.
Slow progress.
Limited time.
Constant trade-offs.
https://t.co/zAznwKuhNS is far from ready.
But seeing it already move between list, kanban and calendar hits different.
Because it stops being an idea.
And starts becoming a tool.
Why?
Because during MVP, your model changes every week.
Adding a new property shouldn’t require:
» Schema migrations
» ORM updates
» Refactoring half the stack
My local API manipulated JSON arrays directly.
When you use AI/vibe coding in a structured way - clear prompts + full context - you can evolve your model safely without fighting your infrastructure.
MVP is about learning velocity.
You can normalize later.
#BuildingInPublic
One of the best decisions I made in my MVP
I persisted everything in JSON instead of using a database.
It wasn’t a mock.
The CRUD was real.
All entities were related.
But I didn’t lock myself into schemas.
What was happening:
» Active sprint wasn’t visually obvious
» All cards had the same weight
» Actions had too much prominence
» Too much gray, low hierarchy
Result: every time someone opened it, they had to re-parse the screen.
That’s wasted seconds.
And in sprint planning, seconds multiply.
What I changed:
» Clear visual hierarchy
» Stronger distinction for the active sprint
» Reduced action noise
» Better spacing and grouping
No new feature.
Just less friction.
#buildinpublic
@TobyQuirk Appreciate that.
I’m new to building in public, but my edge isn’t revenue screenshots
it’s decision rationale.
I’ll document the technical trade-offs driven by business constraints, and the real problems forcing those decisions.
I’m building SaaS & AI side projects in public.
Sharing:
– Architecture decisions
– Revenue experiments
– Automation systems
– What works (and what doesn’t)
Let’s build.
Transform your course material into an AI tutor: flashcards, quizzes, and more coming soon.
✨Accelerate✨ your learnign with your personalized AI tutor Spark.E.
Get started at https://t.co/ztbSC0eDQK