Almost done. Thank you to everyone who read, reacted or commented. If you're reading on a day other than this one, thank you for finding your way to the series.
This post does three things: it restates the promises I made on Day 1 so you can judge them yourself, it gives you a reading map of the whole series in case you arrived partway through, and it says what I'll do next.
The three promises from Day 1
- Only write what I actually did or believe. No invented numbers, no fake wins.
- Say what didn't work. The failures are usually the useful part.
- Keep it readable in one sitting.
I'm the worst judge of whether I kept them. If any post fell short, tell me in the comments. I'd rather be corrected than polished, and I'll update the post and say what changed.
A reading map
If you only have time for a few, here's how I'd pick by interest.
If you build things:
- Day 2: putting business rules in the database
- Day 3: the deploy that silently failed
- Day 12: Whiteboard.ai and the "Saved" promise
- Day 17: Row Level Security with one example
- Day 25: a Vercel serverless route
If you're into AI:
- Day 5: using AI tools without outsourcing your thinking
- Day 11: RAG in plain English
- Day 12 again: rules first, AI second
If you're learning:
- Day 7: tutorial hell
- Day 13: a debugging routine
- Day 23: data structures that earn their place
- Day 28: reading other people's code
If you care about the human side:
- Day 4: how a student weighs earning online
- Day 9: what I'm not good at yet
- Day 14: motivation is unreliable
- Day 26: impostor syndrome without the pep talk
What I'd change next time
Honestly, a few things I'd approach differently if I ran a series like this again:
- Plan around fewer, better topics. Thirty topics in thirty days is a lot of surface area.
- Add more diagrams and runnable code. Posts with something to copy are more useful.
- Reply to comments faster. A conversation is the point.
- Build slack into the schedule for the days life gets in the way.
What's next
- Tomorrow is the last scheduled post, a short look ahead.
- After that I'd like to slow to a weekly pace and write deeper pieces. The most obvious one is a follow-up to Day 20: the CardBridge test suite, once it exists, including what it found.
- I also want to run the Whiteboard.ai AI-review eval against a real key and report the actual catch rate, whatever it turns out to be.
- And I'd like to finally publish the template and tell you honestly how it goes.
How you can help
- Correct me. If something's wrong, say where.
- Tell me which post to expand. One topic will get a long follow-up.
- Share your own writing. Drop a link below. I'd like to read it.
- Tell me what to stop doing. Feedback on style is as welcome as feedback on facts.
A closing thought on writing in public
The series has been an experiment in a simple idea: you don't need permission or a large audience to start documenting what you learn. You need a small, honest habit and the willingness to be seen being wrong occasionally.
Where to learn each topic properly
I'm a student sharing what I've built. For authoritative detail, go to the original documentation. These are the starting points I'd suggest:
-
PostgreSQL documentation: the chapters on triggers, row security policies and
SELECT … FOR UPDATE SKIP LOCKED. - Supabase documentation: auth, Row Level Security and Edge Functions.
- Yjs documentation: shared types, providers and awareness, if you're curious about collaborative editing.
-
MDN Web Docs: for anything about browsers,
sessionStorage,matchMediaand accessibility. - Vercel documentation: functions, environment variables, build output and limits.
- Three.js documentation and examples: the best way to learn is to run the examples and change them.
- WCAG quick reference: for contrast and keyboard accessibility.
- dev.to's own editor guide: for formatting, series and scheduling posts.
Documentation is dry, but it's current and it's the source of truth when a blog post (including mine) disagrees with it.
Questions I'd love answers to
If you'd like to join in, here are starter prompts. Answer any, in any depth.
- What's the most useful thing you learned from a mistake this year?
- What's one tool or habit you'd tell your younger self to adopt sooner?
- What's a project you're proud of that nobody has seen?
- What topic do you find hard to learn from the usual resources, and why?
- If you could read one post from someone further along, what would it be about?
I'll read every answer. Some of them will turn into posts, credited to you if you'd like.
A short note on honesty in public writing
Because the series is built on a promise of honesty, here are the ways I tried to keep it, so you can check me:
- Facts come from my own project notes and case studies, where limits are written down.
- Code samples are labelled as simplified or illustrative when they aren't copied from a repository.
- Opinions are marked as opinions.
- Things I haven't done (tests, evals, a published template, a pull request) are named as not done.
If you find a place where I fell short of this, I want to know.
The habit, not the streak
A daily streak is a useful scaffold for thirty days. It's a poor way to live. After this series, the aim is a sustainable rhythm: something small and regular that keeps the learning public without eating the rest of life.
A good test of whether a habit is sustainable: could I still do this during a busy week? If not, shrink it until the answer is yes.
If there's one topic you'd like a deeper dive on, tell me below.
☕ Enjoyed this post? You can support my writing on Ko-fi.


Top comments (0)