A post about changing engineering interviews pulled 155 comments on Hacker News this week. On DEV this month, "We're Building AI to Replace Ourselves" passed 200. Everyone's arguing about the same thing from different angles, so let me ask it directly.
Would you hire a junior developer who's productive with AI, but stuck without it?
I keep going back and forth.
The case for yes:
- Nobody hires a developer to work without Google, Stack Overflow or an IDE. AI is the tool now, so why test without it?
- A junior who ships working features with AI on day one is more useful than one who can reverse a linked list on a whiteboard.
- Seniors use AI all day too. Holding juniors to a different standard feels a bit unfair.
The case for no:
- AI is confidently wrong sometimes. If you can't read the code it writes, you can't catch that.
- Debugging production at 2 a.m. is where the real learning happens, and it's hard to learn by pasting errors into a chat.
- Someone has to review what the AI produces. If juniors never build the basics, where do the next seniors come from?
Where I've landed for now, and I'm not confident about it: I care less about whether someone uses AI and more about whether they can explain the code they submit. If they can walk me through it and fix it when it breaks, I don't mind how it was written.
But I can see the argument that "explain it" isn't enough, and that you only really understand code you've struggled to write yourself.
So where do you stand?
- If you hire: would you take this person? What would you test them on?
- If you're a junior: do you feel you could code without AI if you had to?
- Has anyone changed their interview process because of this?
Strong opinions welcome. I'd especially like to hear from people who disagree with where I landed.
Top comments (0)