What I learned by building this.

I’m learning about AI supported programming techniques. I built this site using Claude Code to help me cover the gaps in my programming expertise - specifically frontend development.

Here’s how I built this.

The most pertinent things I learned:

• Using AI assistance can rapidly escalate into a mess. It has never been more important to be clear about:
  • why you’re working on something. Clarity of purpose.
  • what you want to achieve. Be intentional.
  • how you want to go about achieving your objective. Be informed.
• This is a powerful new paradigm. It’s a really terrifying combination of power and immaturity. On top of this, it’s evolving so rapidly. It’s even more important to be discplined with your process. The quality of what you produce relies, as ever, on your skill and the points made above.
• Projects which didn’t have the right ROI (ie. too much effort/cost/time for the outcome) just became a LOT faster to build and ship. That’s pretty cool. This means we can scratch an itch on a personal project, or build prototypes and experiment / validate quicker than ever.
• Research, problem definition, ideation, prototyping, testing and convergence on the most compelling features to ship … is all still very important. It’s a lot easier to ideate and prototype with AI assistance. Getting ready to ship to production is easier too, which is why it’s so important to have the clarity I mention above.
• Non-functional aspects of any product should receive the same level of scrutiny and care, perhaps more so if you’re divesting workload to AI. I’m talking about security specifically (and performance). Maybe the old, “trust but verify” approach is best. Mad Eye Moody would say, “constant vigilance!”.

Overall, I’m quite happy with the outcome I achieved with Claude. It’s a simple start (which is the best place to begin!).