Lets admit it, this is not best time to be a software engineer or for anyone whose job is basically sitting in front of a computer. AI is moving so fast these days, and it’s so overwhelming to a lot of us.
Earlier this year, I buried my head in the sand.
I convinced myself it wouldn’t affect me that much. I will just sit tight and see it bust just like what happened to Internet around 2000. I’d just keep doing what I was doing and wait for things to settle.
How wrong was I.
I am no programmer. The first of my first 10 years of careers mostly invovled knowing which button to click on a screen. Its only started with when I realize how powerful it is to use PowerShell to manage Exchange server. The snow ball started rolling. In a few years, I was writing Golang to muck around with Kubernetes controllers. But I was never under any illusion that I was a real coder. Most of my work are the end result of hours of googling and copy/paste from Stackoverflow.
Ironically enough, at one point I was asked to conduct coding interviews. All that taught me was I am fxxxked! All I could think was what if I were the one sitting on the other side of the table?. There is no way I would solve those problems under that kind of stressful setup. I walked away convinced that if I ever lost my job, I’d be in serious trouble.
I started spending extra hours on HackerRank, trying to build muscle memory of solving those nonsense problems. Honestly I didn’t find that helped much for my daily job. As for AI, I just use Copilot to spill out some changes for me everytime I was stuck. Things are good, but not amazing. I get work done. Thats it.
Then something changed. I decided to change my atitude towards AI. I need to spend some serious hours on this thing.
After a few weeks of learning and playing with Claude Code, my way of work completely changed. I don’t spend hours searching for the right code snippet anymore. I don’t bounce between Google, ChatGPT, Stack Overflow, and documentation trying to glue together a solution. I don’t sit there approving every single Copilot suggestion one line at a time. Instead, I spend my time thinking about problems worth solving. What can I automate? What can I improve? What cool thing can I show the team next week?
My bottleneck is no longer my ability to write code, it’s my ability to come up with ideas.
That doesn’t mean churning out slops without proper scruitny. Code review still matters.
Even if I only understand half of the code Claude has written, it is still worth a while to go through the code yourself. Sometime I fix it, Sometime I simplify it, or sometime I just throw it away entirely. Not only it ensures the slop is not just pile of shxt, but also teach myself new things which I will never be able to have exposure to.
I am not one of those 10 x time programmers, which I always admired those peope, who can just build tools/apps from scratch. But with AI, I feel my time has come. It has removed ceiling that my coding ability placed on my ideas. Now I have this wonderful partner that always available to help me explore new ideas, build a PoC, explain narly code base or debug a weird network issue. Fortunately, my day job gives me no shortage of all these.
So how to survive this crazy AI world? I’d say just stop worrying just start using it (very cliche I know).
Don’t understand tech problem, ask AI, including AI itself.
Don’t know how to use AI? Ask ChatGPT, Gemini, Claude, whatever, even Copilot. Don’t know which model to use, try each one, pick the one you like. Don’t have a subscription? Try out the free plans. Almost all those AI agents have free plan, use them, abuse them and once you reach the limit, switch to a different one. If your company gives you acces to AI tools, use them.
Use them for work.
Use them to plan dinner.
Use them to help your kids with homework.
Use them to learn something new every day.
The biggest difference isn’t how you use AI.
It’s how often you use it.
My own strategy is simple.
I don’t try to learn everything.
I learn one small thing every day.
Just one.
It doesn’t feel like much, but after a couple of weeks you suddenly realise you’ve picked up a whole collection of new ideas without ever feeling overwhelmed.
I don’t think AI means programming knowledge no longer matters.
You should still understand what a Go routine is. You should still understand callbacks, concurrency, networking, and how software works. Those fundamentals help you recognise when AI gets things wrong. But don’t wait until you feel “ready” before building something. Because the truth is, AI will probably write code faster than any of us ever will.That’s no longer the competition. The competition is finding interesting problems to solve.
For the first time in my career, I don’t feel limited by how quickly I can write code. I feel limited only by my imagination. And that’s a much more exciting problem to have.
Last modified on 2026-07-25