Nile Suan
Back to all posts

Squishy Humans Need Not Apply

|
ai burnout career

What happens when AI makes you feel like a 10x engineer, and your body reminds you that you're not a machine.


I've been in the IT industry professionally since 2010, but I've never been good at writing blogs. I'm the kind of person who can't sit still for five minutes before getting bored.

Being in the industry for as long as I have, I've had the privilege of meeting and working with some truly amazing people. We used to call them 10x developers. Mehdi Sultani, my old mentor, Tim Toohey, and John Slee, to name a few. They were the kind of people who'd have a spark of inspiration for a solution, disappear for a few hours, and come back with a fully built package. Or they'd fix a low-level problem in an open-source project and contribute it upstream. I always envied how they could do that, because I never had the mental capacity or patience. Skill issues, as they say.

The Early AI Days

From 2019 to 2021, I was fortunate enough to work during the early years of the AI hype. Few-shot classification was the bleeding edge at the time. My project was voice identification via KNN. Those were the boring days of weighing TensorFlow against PyTorch, choosing algorithms, tuning nodes, layers, and activation functions. When I left that role, I thought that was the end of my dabble with AI/ML.

A year or so later, ChatGPT came out. It was pretty ordinary at first, but it improved fast. A year ago, I was using it for nothing more than autocompletion in my IDE, like some glorified IntelliSense. Now I'm confident enough that I only review the output and sign off on it. Claude writes good-quality code at blistering speed. Granted, I feed it about 20 to 25 markdown files of engineering standards and checklists, a full workflow via custom commands and skills, and all the other fancy bits. It feels like I'm one of those 10x developers now. It's exciting and addictive.

The Crash

Today, for the first time in my working life, I felt properly burnt out. With the power of Claude, I was working on four different projects simultaneously. It felt glorious, but it also felt overwhelming. When a colleague asked me to help fix something, my brain was so overloaded from context-switching that I basically clammed up. The moment I got home, I was exhausted. The kind of exhausted where you drop on the floor the second you close the front door and just sleep. And in true software engineering fashion, I asked myself why it happened and how I could avoid doing it again. The answer was pretty simple: ease off the drug called AI.

Respecting the Limit

I used to resent the five-hour session limits on Claude, even on the highest tier. I kept thinking, why would you limit us when our creative juices are flowing and we're in a state of flow? But today I learnt the hard way. Gigawatts of wasted electricity on data centres and global warming aside, it's there for the safety of the users, to stop us from overdoing it.

Taking It Slow

So from today, I'm going to apply the philosophy I've picked up from my years of living in Oz: be laid-back and take things slow. It's easy to get excited and lost in work with all the help AI can provide, but health and family are more important. I plan to stick around a bit longer to see how this world progresses with tech.


The title of this post is a reference to a YouTube video by one of my favourite creators, CGP Grey. When I felt that exhaustion hit, the first thought that crossed my mind was: if I can work on four projects at the same time and finish them all by the end of the day, I can't imagine there'll still be a need for me in the near future.