I had my first technical interview today, and honestly, it felt terrible.
It was a CoderPad-style interview. Not a pure algorithm test, but still a timed, live coding setup where you are expected to perform on command.
I stumbled more than I expected.
For the last nine-plus months, I have relied heavily on tools like Cursor autocomplete, Claude Code, and Codex. That is how I build now. So when I had to code in a stripped-down environment, without my normal workflow, I felt the gap immediately.
Part of that is on me.
I probably should have spent more time rebuilding keyboard reps, tightening syntax recall, and doing a few more practice problems under time pressure. I can own that.
But this experience also raised a bigger question:
Is interviewing still testing the right thing?
The best interviews I have ever done were take-home projects with a follow-up discussion. Low pressure, realistic scope, and room to explain decisions, tradeoffs, and extensions.
That format works for me because it looks like real work.
I can explore a codebase, get context, and ship something thoughtful instead of trying to survive a performance test.
I also think we need to come to terms with AI in interview rooms.
The question should not be, “Can this person solve everything from memory with no assistance?”
The better question is, “How do they use AI in a real workflow, and how do they verify what it gives them?”
Anyone can prompt a solution. Not everyone can guide it well, validate it, and integrate it responsibly.
Does using AI sometimes feel like cheating? Sure.
But calculators used to feel like cheating too. Over time, we accepted that the tool is not the problem. Understanding is.
Some of our older “show your work” expectations are fading. That does not mean skill is gone. It means the skill is shifting toward judgment, system thinking, and accountability.
There is another part of this people do not see.
Tonight I went to my nephew’s birthday party. No one in my family knows I was laid off. I usually get asked, “How’s work going?”
I kept my head down and stared at my phone so I would not have to answer.
I know the layoff was not performance-related, but it still feels heavy. Interviewing is different when your family’s stability feels tied to every single conversation.
So this is where I am:
I need to improve, and I will.
But I also think the industry needs to evolve. If interviews are supposed to predict job performance, they should look more like modern engineering, including how we use AI day to day.