I once met a TV repair man in the 90s sitting idly in his shop. He wasn’t going out of business dramatically — just quietly winding down, the way trades do when the world stops needing them at the pace it used to. He didn’t seem angry about it. Just tired in a way that had nothing to do with his age. I didn’t think much of it at the time. I think about it a lot now.
The recent iterations of LLMs have led to impressive artworks and software being created by ordinary people who have not spent a lot of time in the trade. A few hours of prompting, and these things existed. The world welcomed this new convenient ability for this is not far off from a sci-fi miracle.
As for me, and probably millions of senior software engineers out there, this could be the ultimate change we’ve been bracing for. A total shake-up of the entire industry where all the rules are now being cast aside, and AI not just sets a new standard, but invites the whole world into the software department.
Speaking with some of my engineer peers, I sense the collective excitement for this new AI era, but at the same time, I do percieve a certain disbelief and grief that resonated with me. Psychologists have a name for this: the five stages of grief, the process people move through when they’re facing loss, even the loss of a way of doing things.
Denial
The first stage of grief is denial.
It can’t be true. They’re just showing off what works, hiding all the failed attempts. AI is just autocomplete with delusions of grandeur. Sure, it can write a function, but it doesn’t understand anything.
I said versions of this longer than I’d like to admit. Every dismissal was technically true and completely beside the point. I wasn’t evaluating the tool. I was protecting something.
Anger
Then my graphic designer — not a programmer, never claimed to be one — showed me something she’d built with AI. Working code. Solving a real problem.
My first reaction wasn’t curiosity. It was something smaller than I want to admit in writing. That’s supposed to be the hard part. I wasn’t mad at her. I was mad the ground had moved without asking me first.
Bargaining
This is the stage where I got busy and prove that the traditional way is more reliable, even if it means sacrificing sleep and time with family. This isn’t the first time I’ve speed-coded out of a deadline and saved the day. I remembered a wise man omce said: Never give up, never surrender!
None of that was wrong, exactly. But it was still me negotiating my way back to the old feeling of being indispensable — instead of asking whether that feeling was ever something worth hanging onto.
Depression
This one didn’t feel dramatic enough to have a name. Just a quiet, ongoing unease. If the thing I was good at wasn’t rare anymore, what was actually mine? I found myself questioning things I hadn’t questioned in years — not just my job, but a lot of what I’d been measuring myself by.
I’m not going to wrap this stage up with a bow. It wasn’t fun. Some days it still isn’t.
Acceptance
For me, acceptance has been smaller and more practical than “AI is great now.” It’s faster at the mechanical parts of coding – writing, debugging, iterating. That leaves room for what always mattered and never got enough hours: making things usable, secure, documented, actually explained to the people who’ll use them.
It also left room for a bigger realization: a lot of what I called “intelligence” – fast recall, clever pattern-matching – turned out to be more replaceable than I thought. Wisdom, the slower thing, is holding up better.
There’s a verse I keep returning to, from Isaiah: God saying not to dwell on former things, because “I am doing a new thing.” Present tense. Still happening. Life is always moving, and AI is just a part of it.
