Skip to content

Exploring the use of AI to support "web development for dummies"

My Objective:

In recent posts, I’ve been looking at how AI can help Solution and Software Architects in their day to day job. But what about software developers? Let’s narrow it down a bit: How about someone who has experience in development but who is not up on the latest languages or needs to create something in an area that is new to them? Like creating a website and deploying in a server less model to a cloud?

This is where I saw an opportunity to combine 2 of the “journey’s” I’m on in retirement. I needed a simple website for my small woodworking business to answer the question I often get: “Where can I go to see some pictures of your work?”. So, I asked myself … “Just how helpful can AI be in the creation of a website when I’ve either never done it before or very ‘rusty’?”

This post, and others that follow it, are a record of how I approached this challenge, what I learned along the way, and a summary set of observations regarding the implications and impact using AI tools as an assistant to software developers.

It should be noted that I time-boxed the initial effort. I wasn’t looking to have everything perfect. I wanted to see how much I could get in place working an hour or two each day for 2 weeks.

You can see an early beta release of this site here: nextchapterwoodworking.com .

Also.. I spent a lot of time getting help on the language, details of the deployment approaches etc… Net/Net: It’s one thing to draw architectural pictures. It’s another to do the coding and have experience in that area.

Key questions I had going into this

  1. Could I get AI to do most of the work since I wasn't an expert at this?
  2. How little information did I have to provide in the prompts to get the results I wanted?
  3. Would AI do all/most of the work for me?
  4. Would it help me come out of this smarter for the next time I needed to to this?
  5. Would I need to engage in a long dialog with the AI engine or could I provide a complete prompt/prompts, like I did in the Software Architecture exploration and get complete working code?

The Scope of the Project

My criteria for this project was:

  1. A simple, non-transactional, website to present a portfolio of work and contact information.
  2. Cloud-Native architecture using containers and server less deployment environment
  3. A content stored and managed in Cloud (Object) storage bucket.
  4. Secure (private) networking between the serverless engine (CloudRun) and the storage bucket.
  5. Use diagram driven code generation where possible.
  6. Python as the development language
  7. Configurable style sheets.
  8. Leverage Google Cloud capabilities. This includes using Google Gemini and Gemini Code Assist in Visual Studio.

The Approaches I leveraged

WIP

Observations

WORK IIN PROGRESS placeholder

Conclusions

placeholder


Last update: March 9, 2025
Back to top