My Claude Lifeline

About six months ago, I experienced the most vicious burnout of my life. I was completely unable to continue work as a software developer because the day to day job of writing code, testing, meeting with stakeholders to plan features, investigating bugs, and everything else I do became too anxiety inducing for me. I was on the verge of crashing out of the tech industry fast. I told my boss I quit my job, and I was trying to find any kind of work that didn't stress me out the way that software development did.

This was the push I needed to finally figure out something I had been struggling with for my entire life. I got an evaluation and discovered I am autistic and have ADHD.

Finding out you are neurodivergent and understanding what that means for how you work is very helpful, but there is also something that happened around the same time that has had an even bigger impact for me. The biggest change was the proliferation of AI coding, and especially the model improvements for Claude Code late in 2025.

I don't think enough has been said yet about the benefits of AI tools like Claude Code for neurodivergence. It's completely changed how I approach building software in ways that are very specifically helpful for how my brain works.

One of the biggest struggles I have is with executive function. It's difficult to get the motivation to start a task, especially a large task, when you are very dopamine deficient. Claude significantly reduces the scale of any given task, making it a much lower barrier to start on tasks.

Is there an obscure server error that is going to be a pain to track down and debug? I just drop the error log in Claude and let it spin for a bit to see if it can find something. It doesn't always get it right, but that doesn't matter. Just getting the task started and giving me something to interact with and correct is so much easier than trying to force myself to methodically search and read through code, etc.

Is there a new edge case I need to add to an existing feature? I've created a skill for Claude that inverts the conversation, and gets Claude to ask me questions about the feature after loading the relevant code and any related Slack conversations into its context. This is so much easier than trying to think through the requirements methodically on my own and then trying to think through the code changes required.

There are a thousand more ways that AI tools are specifically helping me adapt my work to how my brain works, to the point where I wonder if neurodivergence is a tangible advantage for working with AI. Because of this, I'm excited for my work in tech for the first time in nearly a decade.