Starting Smaller Next Time

If I started Beads of Hope T1D over again, I would not change the reason for it.

That part is still the same. My son was diagnosed with Type 1 Diabetes when he was two, and this whole project is for him. The bracelets, necklaces, badge reels, pens, the awareness piece, the giving-back part — I would keep all of that.

What I would change is how I began.

On June 2, I asked the AI for a full production-ready ecommerce and community website in one shot. Shop page. Checkout. Google login. Stripe. Database. Community pieces. The whole thing.

And because these tools are slightly magical and slightly dangerous, it actually tried.

In about 20 minutes, I had a real-looking site with 22 database tables (basically the behind-the-scenes storage for orders, products, users, and all the bits I did not know I was asking for). I remember feeling amazed.

Then the bill came due.

Not a money bill. A confusion bill.

That same evening, I spent hours fighting product images that would not show up on the published site. The space was there, but the images were invisible. When I clicked where an image should have been, the page snapped back to the top of the “all” category. Duplicate category buttons kept appearing too, just to make it extra charming.

At one point I told the AI, “images are still not visible and each time I click on the box the image should be in, the screen immediately reverts back to the top.”

That sentence probably sums up my entire mood that night.

If I were starting again, I would build the shop page first. Just the shop page. Products, photos, categories, and clicking into an item. I would not touch loyalty points, community features, emails, or checkout until I could see a bracelet on the screen and click it without wanting to throw my laptop into the yard.

Then I would add checkout.

Then shipping and tax.

Then emails.

Then loyalty points.

Then the nice-to-have things.

I also would not obsess so early over tiny design details. I spent real time moving a blue ribbon graphic down a quarter inch, getting the heart closer to the crossover point, and adding white shading behind “T1D Hope” so it stood out.

Was it wrong to care? No. It mattered to me.

But did I need to care about that before I knew the product images worked? Probably not.

Same with the hover-zoom on badge reel images. I asked for 3x zoom, then 2x, then only on badge reels and not everything else. I’m laughing at myself a little now, because 99% of visitors may never notice. But I noticed. And sometimes noticing is how you get stuck.

The other thing I would do differently is take security seriously earlier, even if I did not understand it.

Later, a scan flagged nine issues with names like “guest cart hijacking,” “client-controlled unit price,” and “RLS policies too loose.” RLS means row-level security, which is basically rules about who is allowed to see or change certain information in the database.

I understood maybe 40% of what was happening. Maybe.

The AI fixed the issues and re-ran the scan, but next time I would ask for those checks sooner, especially before anything involving money.

Because connecting Stripe was its own emotional event. Wiring it into the code was one thing. Activating a real account with bank information, business information, and two-factor login was the first moment I thought, “Am I really doing this?”

I still would do it.

I would just walk toward that moment in smaller steps.

The good news is that the project did get there. The shop images finally worked. The badge reel zoom finally behaved. Loyalty point redemption worked. Order confirmation emails showed up in my inbox looking real. The store is live-ready now, with payments, shipping, tax, and emails wired. I’m mostly waiting on the very human parts, like Stripe activation and getting brave enough for the first real order.

So no, I would not start over with less heart.

I would start over with a smaller first prompt.

What I learned: asking for everything at once can work, but it can also make every problem harder to untangle.