What happens to the Developer?

Scott Thompson — Wed, 07 Oct 2026 21:36:15 +0000

Recently, I've been struggling more and more with the way we're using AI to automate nearly every step of software development.

At first, I thought it was fun, interesting, and exciting. We were discovering new ways to make our lives easier and remove some of the tedious parts of development. I still believe there is tremendous value in that.

But lately, I've become increasingly uncomfortable with where this is heading.

In the past, I would read articles from developers discussing their favorite framework and why they preferred it, the latest technical challenge they encountered and how they overcame it, or even something they built purely for themselves to learn and improve.

Now, I primarily see some variation of:

"I vibe coded this app in an afternoon."

And I think that shift says something important about where our priorities have gone.

We seem increasingly focused on the end result rather than the development of the individual developer along the way.

That makes me wonder:

What about the developer as an individual?

What are we actually gaining in our careers when AI completes more and more of the work for us? If a tool can generate the application, write the code, debug the problem, and explain the solution, what am I personally developing through that process?

What differentiates me from another developer if our primary contribution becomes knowing what prompts to give a computer?

I'm not saying AI doesn't have value. It clearly does. I've used it myself, and I've experienced firsthand how much it can accelerate development.

But acceleration isn't the same thing as growth.

There is something valuable about struggling with a problem for several hours. There is value in reading documentation because you don't know the answer. There is value in writing something poorly, realizing why it is poor, and rewriting it. There is value in getting stuck, asking another developer for help, and eventually understanding the problem well enough that you don't need their help the next time.

Those experiences don't just produce software.

They produce developers.

And I'm afraid we're beginning to overlook that distinction.

If AI allows a junior developer to complete a task that previously would have required a senior developer, that may be great for the immediate project. But what happens to the junior developer who never gets the opportunity to develop the skills that would have made them a senior developer?

If the machine increasingly handles the struggle, when do we get the opportunity to become better at struggling?

This is also why the current conversation around the future of software development concerns me. The industry has already experienced significant layoffs, and AI is increasingly being presented as a way to accomplish more with fewer people.

Maybe that is simply where technology is taking us. Maybe software development really will require fewer developers in the future.

But if that happens, I don't want the only thing we measure to be how much software we can produce.

I want to know what happens to the people producing it.

What skills are they developing? What are they learning? What are they becoming capable of doing that they couldn't do before?

Because at the end of the day, I'm becoming less interested in what a computer can do.

I'm much more interested in what a human being can learn to do.

I want to see developers become stronger, more capable, more curious, and more creative. I want to see people build things because they want to understand how they work, not simply because they can generate the result in an afternoon.

Maybe that's the part of software development we need to protect.

Not the code.

The developer.

Pet-Wallet: Building an Angular 20 Pet Management App with Gemini AI

Scott Thompson — Mon, 17 Aug 2026 01:23:31 +0000

This is a submission for Weekend Challenge: Dog Days Edition

What I Built

Pet-Wallet is a lightweight pet management application that allows users to add, edit, and remove their pets. The application stores pet information such as name, breed, weight, and other personal details, while also allowing users to track daily care tasks and medications for each pet.

I also integrated AI to help users collect pet information. The application can read a document containing pet data and use that information to assist with creating a new pet profile.

This project was an opportunity for me to experiment with my recent development skills, participate in one of my first coding challenges, and now make my first post in the DEV.to Community.

Demo

🐾 Try Pet-Wallet Live

Code

Griffonknox / Pet-Wallet

Pet Wallet

Pet Wallet is a pet care dashboard built with Angular that helps owners manage their pets, medication schedules, and daily care tasks in one place.

Overview

This application lets users:

create and manage multiple pets
edit pet profile details such as name, breed, age, weight, and avatar
view a pet dashboard with summary stats
add and complete care tasks for the day
add and manage medications with active/inactive status
store pet data locally in the browser
optionally import pet details from a document using the Gemini API

Features

Pet selection screen with add, edit, and delete actions
Pet dashboard with dynamic derived stats
Care task tracking and completion workflow
Medication management with status indicators
Local persistence via browser storage
Rounded, modern UI with a soft dark theme
AI-powered document extraction for pet profile setup

Tech Stack

Angular 20
TypeScript
SCSS
RxJS
Google Gemini API

Getting Started

Install dependencies:

…

View on GitHub

How I Built It

With the goal of keeping this experiment lightweight, Pet-Wallet is a front-end-only application built with Angular 20. I enjoy working with Angular because of its use of TypeScript, reusable components, and the ability to use a centralized Pet Service to manage our pet data and state.

The application is deployed using GitHub Pages, with GitHub Actions handling the deployment process automatically whenever changes are pushed to the main branch.

The final technology I wanted to experiment with was Google Gemini. I integrated Gemini to read PDF documents containing pet information and use that information to assist with creating a new pet profile.

This is where things get a little more interesting. Because I wanted to keep Pet-Wallet as a lightweight, front-end-only experiment, the Gemini API request has to happen from the browser. That means I couldn't safely include my own API key in the application or public repository.

Instead, I added a workflow where the user provides their own Gemini API key when they want to use the AI import feature. This isn't an architecture I would recommend for a production application(in a real application), I would move the Gemini request behind a backend service so the API key could remain server-side. But for this challenge, it allowed me to experiment with the Gemini API without exposing my own credentials.

I'm definitely interested in hearing how others would approach this differently, so if you have suggestions for alternative solutions, I'd love to hear them in the comments!

Prize Categories

Best Use of Google AI

Pet-Wallet uses Google Gemini to assist with creating new pet profiles. Users can provide a pet-information document, and Gemini reads the document and extracts relevant information that can then be used to populate the new pet form.

I chose this feature as a way to experiment with Google AI while making the process of creating a new pet profile easier for the user.

DEV Community: Scott Thompson