A fun feature of SwiftUI is the ability to create progress indicators beyond the old-school left-to-right thin rectangle. Yawn. Rectangles are so 20th-century. Users today expect something big, bold and beautiful and circular progress meters fill that UX element nicely.

I also like to practice my software skills in creating digital clocks. I guess it’s a habit born of being an hourly contractor mixed in with having a post-graduate advisor who loved to restore old clocks. Rather than talk about the next steps in my studies, he loved to show off his latest centuries-old clock that was now “purring like a kitten.”

Time to put ChatGPT 4 to the test.

My goal is to describe to ChatGPT 4, a simple application that displays the time of day down to the second, using three nested circular progress meters representing the hour, minute and second on an Apple device. As a bonus, I’d also like the time displayed in the regular hour:minutes:seconds format. ChatGPT will be my software developer and generate all the SwiftUI code the app requires. Then, in a conversational style, I will act as the product manager and describe to ChatGPT the enhancements I want it to make as we drive towards a shared vision of the application.

Note the phrase “shared vision.” Working with an AI is quite similar to how I’ve worked as a member of a team of developers. Each of us has a vision of what the app will look like and how it will perform, and through a series of conversations, we eventually arrive at a shared vision. Not my vision. Our vision. Groupthink is a critical mindset when working with an AI—one of cooperation rather than domination. As a developer working with others, I must put my ego on the back burner and prioritize the shared vision over my vision. It’s not easy, but it’s always worth it.

On August 31, 2017, while most of us were looking forward the arrival of the dog days of summer, the most fundamental advancement in Artificial Intelligence (AI) slipped onto the Internet and under the radar of nearly everyone. On that sleepy Thursday, Google published a white paper with the academic title Transformer: A Novel Neural Network Architecture for Language Understanding. A document that, in hindsight, might turn out to be the precursor of AGI – Artificial General Intelligence. The dawn of Homo Sapiens 2.0. A technology that, on November 20, 2022, would stun the world with release of ChatGPT 3.0.

ChatGPT 3.0 (the ‘T’ in the name meaning Transformer) turned out to be, pun intended, transformative regarding natural language processing (NLP). Suddenly, with a few lines of text, anyone could write a poem, complete a complex essay for high school and write reams of computer code in over a dozen languages, including my latest favourite – SwiftUI.

Four months later, ChatGPT 4.0 was released. An NLP that was an order of magnitude more powerful than its version 3.5 predecessor. A version that could pass the US bar exam and write a novels in a day. And, importantly to me, one that had the skill of an entry-level software developer to that of a seasoned veteran.

After 45 years of software development, I saw my career ending. Or was it? Did the development of the application program interface (API) design pattern end my career? No. It made it more accessible. Did the development of reactive programming languages such as SwiftUI crush my career as a UX designer? No. It made it infinitely more enjoyable. The evolution of the software development industry is no stranger to things getting more accessible and reliable. And while the arrival of ChatGPT and its code-writing skills stunned me, I knew this was just another moment of technical evolution.

It was time to see for me, and perhaps you, just how good ChatGPT 4.0 was for writing SwiftUI applications.

What follows will be a series of blog posts on my first crack at using ChatGPT 4.0 to create a simple but flashy SwiftUI application. The experience was, to be honest, stunning. In just an hour, I completed what even just a year ago would have taken me an entire week.