ChatGPT Experiment: Building with technology we don't know

When ChatGPT disrupted our industry, after the initial shock, I decided to give it a try. I am a technical person, but I didn’t do first-person software development for a long time. And the technologies changed over time. To create a modern microservices app, one needs a lot of knowledge about different stacks, languages, and implementations.

When I saw the true power of ChatGPT4, I decided to make an experiment. My Typescript or Javascript experience is definitely not production-grade. But I wanted to see how far ChatGPT4 will take me throughout my journey to develop a standard CRUD backend service.

I started with a standard description of what needs to be done: a backend service, which provides standard CRUD operations (and maybe a bit more, we’ll see!) for two data sets: Users and Games. I wanted to have the service in Typescript, implemented via Express and MySQL, and run by Node.js. The whole service had to be packaged in a container and run in my local Docker setup (for now).

I didn’t have the chance to write any example (or production, for that matter) code. I did a few Docker courses, but I’m very far from even a junior Docker specialist (but I can read and roughly understand Dockerfile). But I wanted to see if ChatGPT4 would enable me to build all this with my current (junior level) technological knowledge and if it would walk me safely through this implementation journey.

This presentation demonstrates my findings throughout this road. We will walk that road together, demonstrating and hopefully convincing everyone that ChatGPT4 is not here to make us hunter-gatherers but to enable an even higher level of productivity. Or does it?