Claude Is Down, and So Am I
I remember a time when life was simple. A broken build was the ultimate excuse, or being blocked by backend was a valid reason for not deploying and sitting back to wait for someone else to fix it. Productivity theatre? More like productivity destruction.
Since software development changed into an unrecognizable landscape of AI and prompts something much more concerning has happened. We simply stop. So I’ve got a question to ask, and it’s who is going to waste the earth’s natural resources if you aren’t burning tokens like a Californian forest fire.
The New Critical Dependency
AI tools have slid into the development workflow so smoothly that nobody really noticed when they became essential.
Need to write a function? AI.
Forgot a syntax detail? AI.
Debugging something you don’t quite understand? Definitely AI.
We’ve replaced “Googling it” with “asking the machine.” Faster, cleaner, less cognitive effort.
And that’s the problem.
While the normal person in the street has taken to using AI to ask “What is the capital of France” (for those unfamiliar, this is F) software engineers are doing the same with simplistic software engineering questions. Each prompt takes tokens, each prompt takes money and each pointless question burns a little bit more of the earth’s natural resources.
Why Not Waste?
At work we seem willing to burn money at the AI altar. We’re being told by management to “git gud” and become more productive at any cost to the businesses bank balance or the earth’s natural resources.
If you want to make a PR the pressure is to get Claude to write it. Instead of writing PRDs we should simply spin up some demo software and push it to production as soon as possible. Who cares if some people will not be able to access fresh water as a consequence, right? Since we are getting measured on our AI usage (and not the outputs of it) we need to get our AI usage up, whether it is required or not.
So how am I going to destroy the planet if both Claude and Cursor are down?
I think the answer is to leave the tap running if my favorite AI is not working. Leave a few lights on, I think that will help the powers that be feel satisfied that I’m truly destroying the planet for them.
The Productivity Illusion
Encouraging the wrong behaviors because you don’t know how to actually measure productivity? That tracks.
We do track the number of PRs that developers produce, but we don’t think about the quality of those PRs as they pass by.
The fact that we track behavior that destroys the world…is interesting on many levels. This isn’t the same as wasting company money by “stealing” free drinks or food at the office. This is about using the earth’s dwindling resources for results that could be better obtained through positive behaviors.
Because if you’re doing it yourself you get to retain the information from doing so, not a corporation that dreams of making you redundant. AI gives you instant answers. Instant progress. Instant gratification. When it’s not there suddenly, everything feels… harder than it should be.
You need to fill that hole with something.
The Risk Nobody Wants to Talk About
As we risk our habitual spinning ball for “ease of production” you forget what is at stake here.
If you rely on AI for:
Writing code
Explaining concepts
Solving problems
What happens to your own ability over time?
I’ve seen teams struggle with basic concepts because they’ve outsourced understanding to tools.
It works… until it doesn’t.
And when it doesn’t, the gap becomes painfully obvious. That’s when we are going to need someone to code to work out a way of saving us all. If that somehow exists in a few years time.
About The Author
Professional Software Developer “The Secret Developer” can be found on Twitter @TheSDeveloper and regularly publishes articles through Medium.com
The Secret Developer feels quite bleak today. Can you tell?