100 SUBSCRIBERS! Celebrating While Fixing Our Website Design | Learning In Public Day 15

So here's the thing - I hit 100 subscribers on YouTube this week, which honestly happened faster than I expected. But instead of just celebrating, I spent the entire day fighting with my website's design and getting absolutely nowhere.

Quick summary

Day 15 of building my travel directory in public turned into one of those frustrating coding days where nothing goes according to plan. I wanted to make my stays page look less boring, add a rich text editor, and fix some alignment issues. What actually happened? A lot of trial and error, some honest mistakes, and a reality check about AI coding tools.

Why I'm sharing this mess

Living here in Bali and building this travel directory with my partner Grace, I've learned that the messy days are just as valuable as the breakthrough moments. Plus, if you're thinking about building something similar or working with AI coding tools, you need to know it's not always as smooth as the highlight reels make it seem.

The challenge I was facing

My stays detail page was working functionally - we had the form storing information, images displaying, location data, room details (finally fixed from last episode). But visually? It looked like a boring corporate website from 2010.

The problem was alignment and visual appeal. Our about page had these nice green and blue blocks that made it feel more alive, but the stays page was just... flat. Everything felt disconnected, and worst of all, the content wasn't properly aligned in containers like the rest of the site.

What I tried first (spoiler: it didn't work)

I opened up Cursor - my AI coding assistant - feeling confident. I mean, how hard could it be to add some colorful blocks and fix alignment, right?

My first prompt was basically: "The styling is boring, our about page has green and blue blocks, can we make this more colorful? Also make the hero image smaller and fix the alignment issues."

Cursor said it was done. I refreshed the page. Nothing. Literally nothing had changed except maybe the button color got slightly different.

So I tried again with a screenshot, thinking maybe it needed visual context. Still nothing meaningful.

The breakthrough moment (that wasn't really a breakthrough)

After several failed attempts, I realized I was approaching this wrong. Instead of trying to get AI to magically fix my design, I needed to:

1. Research proper container structures myself

2. Find design inspiration on Dribbble

3. Maybe add some doodles or visual elements manually

I started looking at travel website designs and realized what I actually wanted - those cute, approachable elements that would appeal to Grace's mostly female Instagram following.

Here's what I actually accomplished

Step 1: Implemented a rich text editor (sort of)

I decided to tackle the rich text editor using TipTap, which I've used in previous projects. The idea was to let us format the descriptions, experience notes, and pro tips properly instead of having plain text everywhere.

I gave Cursor the TipTap documentation and asked it to implement it in both the admin form and the frontend display. This time, it actually worked... mostly.

Step 2: Discovered the real problem

The rich text editor was saving bold text and formatting, but line breaks weren't displaying properly on the frontend. Classic case of the editor storing HTML but the frontend not rendering it correctly.

After some back and forth, we figured out it needed additional CSS styling to handle paragraphs and line breaks properly.

Step 3: Accepted temporary defeat on design

Instead of continuing to fight with AI prompts that weren't working, I made a strategic decision: fix this stuff offline and focus on the bigger picture.

The stuff that went wrong

Let me be brutally honest about what didn't work:

AI coding struggles: Cursor, which had been amazing in earlier episodes, seemed to struggle with these visual changes. Maybe it's because I'm on a 4K display, maybe the prompts weren't specific enough, but it kept claiming to make changes that I couldn't see.

Scope creep in my own head: I started wanting doodles, then better typography, then perfect alignment, then... you get the idea. Classic case of letting perfect become the enemy of good.

Technical debt catching up: Some of the quick fixes from earlier episodes were now causing alignment issues that weren't easy to patch.

How I'm fixing the problems

Instead of forcing everything in one session, here's my new approach:

For the design issues: I'm going to research container structures properly and implement them manually. Sometimes you just need to write the CSS yourself instead of relying on AI.

For the rich text editor: The functionality works, just needs some CSS tweaks for proper line break rendering.

For the bigger picture: Rather than perfect every detail now, I'm planning to get the basic functionality solid, then duplicate this template for dining and other categories.

What it looks like now

The stays page is functional but not pretty. We can:

• Add new stays through the admin

• Display all the information correctly

• Format text with bold, italics (line breaks still need work)

• Show images and location data

But it still looks boring and has alignment issues that bug me every time I look at it.

Lessons I'm taking from this

• AI coding tools have limits: They're amazing for functionality but sometimes struggle with nuanced design requests

• Perfect alignment obsession is real: I spent way too much mental energy on something most users probably wouldn't notice

• Building in public means showing the messy days: Not every episode can be a breakthrough moment

• Strategic retreats are okay: Sometimes you need to step back and approach problems differently

What I'm working on next

Next episode, I'm planning to:

1. Fix the alignment and design issues properly (probably offline first)

2. Get the rich text editor line breaks working

3. Add some SEO-focused heading structure

4. Maybe implement that map embedding feature

5. Start thinking about the review system structure

The bigger goal is to get this template solid enough to duplicate for dining experiences, then start having Grace add real content so we can push the site live and let Google start indexing it.

Final thoughts

You know what's funny? I used to think these coding sessions would get easier as the project progressed. But actually, now that we have a working foundation, the challenges are more about polish and user experience - which are somehow harder to nail than basic functionality.

I'm still excited about this project, even on frustrating days like this. The fact that we went from zero to 100 subscribers while building something real feels pretty good. And honestly, sharing these struggles makes the eventual wins feel more meaningful.

Plus, living in Bali has taught me that not every day needs to be productive in the traditional sense. Sometimes you just need to accept that today was a research and learning day, not a shipping day.

Following along with the 60-day challenge? I'd love to hear if you've had similar struggles with AI coding tools or design decisions. The messy days are part of the journey too.