What AI Can’t Fake in a Technical Interview

Over the last few weeks, I’ve had more conversations than I can count about how the interview process is changing. Especially in software engineering.

And here’s the reality everyone is quietly circling: It’s getting harder to tell who can actually do the job.

With new AI tools claiming to be “invisible” during interviews, candidates are literally being fed real-time answers as they’re speaking. And in some cases, it works. Until you push just a little deeper.

I’ve seen a few clips. It’s subtle, but obvious. The delays in speaking. The stiff language. The lack of connection to their own experiences. It’s not malicious. It’s just…empty.

So this week’s newsletter is about how to dig deeper. Whether you're interviewing as a hiring manager or as a candidate, these are the moments that separate the surface-level from the real deal.

Let’s say you’re interviewing someone for a frontend engineering role. You ask, “Can you explain what clean code means to you?”

Solid question. Easy to Google. Easy to recite.

Now try this follow-up:
“Can you tell me about a specific project where you applied those principles? What did the architecture look like? What challenges came up?”

This is where the truth shows up.

You’re not looking for buzzwords here. You’re looking for evidence. Context. Clarity. The little things that show they’ve been in the trenches. Real experience is textured. It comes with nuance. It comes with ownership.

You do not need to know how to code to listen for ownership.

Here are a few ways to pull that out:

Ask about their biggest technical mistake.
What was the bug? What broke? What did they learn?

Dig into project specifics.
Who made the architecture decisions? What tech was chosen and why? What tradeoffs were made?

Explore opinions with context.
Do they hate PHP? Cool. Why? Have they used it in production or are they just parroting Hacker News?

And here’s my favorite principle to remember:
The only programming language people don’t hate is the one they haven’t used.

Everyone has preferences. That’s fine. But strong opinions without experience is a red flag. It usually signals immaturity or inexperience.

The deeper goal of the interview isn’t to hear perfect answers. It’s to find out if this person knows how to think, communicate, and operate in real-world situations. It’s to see if they’ll be a calm contributor or a brittle over-optimizer.

Ask how they handled legacy code. Ask how they worked with product and design. Ask what they wish leadership understood about engineering.

That’s the stuff AI can’t fake.

Let’s be honest. Interviewing in tech today can feel like performance art.

Everyone’s trying to sound smart, seem senior, and hit the right buzzwords. Add AI tools to the mix, and now you’re also competing with people copy-pasting answers fed to them live during the interview.

It’s wild out there.

So here’s the deal: If you want to stand out, stop trying to sound perfect. Start showing what you’ve actually done.

The most impactful thing you can bring into an interview is clear, specific examples of the features you’ve built, the challenges you’ve faced, and how you’ve worked through them.

Let’s take it back to the basics.

If you’re a developer, your resume shouldn’t just say “built large-scale applications.” That’s fluff. It doesn’t mean anything.

Instead, get granular.

That’s the gold. That’s what hiring managers want to know. Not whether you can define an algorithm you’re going to Google later anyway.

And for those of you working on side projects or just getting started, don’t overthink it.

You don’t need a flashy, enterprise-scale app to prove your value.

Pick one feature. One function. Build it as well as you possibly can.

Let’s say you’re working on a grocery list app. Don’t stress about launching the full suite. Focus on one piece, like building the best search and filtering experience possible. Get curious about edge cases. Think through the data structure. Refactor when it’s messy. That shows depth. That shows care. That shows how you think, which is exactly what great interviewers are looking for.

Because here’s the truth: Most hiring managers care way less about what tech stack you used and way more about how you made decisions, where you hit friction, and how you navigated real problems.

AI can give polished answers. But it can’t talk about late nights debugging an issue that wasn’t even yours. It can’t tell stories about pushing through blockers or learning something the hard way. You can.

So if you’re prepping for an interview, here’s your new checklist:

✅ Pick 2–3 features you’ve built and break down what made them meaningful
✅ Practice telling the story, not just listing tech used
✅ Reflect on what you learned, how you grew, and what you’d do differently
✅ Be ready to talk about tradeoffs, bugs, and collaboration
✅ Don’t just prepare to impress; prepare to explain

And one more tip before we wrap:

End your interview with this question:
“While we’re talking, do you have any concerns about me that I can address?”

It does two powerful things. First, it gives you a pulse check on how things are going. Second, it gives you a chance to clear up any doubt in real time.

If they fumble or dodge? That’s telling too. Either they aren’t sure what they’re looking for, or they aren’t confident enough to say it out loud. Either way, you just learned something important about the process, and maybe the team.

At the end of the day, remember this:

The goal of the interview isn’t to get the offer.
It’s to figure out if this is the right fit for you.

Not just technically. But functionally. Culturally. Personally.

If the only way to succeed in their process is by sounding like ChatGPT…that’s a red flag. You’re not trying to win an audition. You’re trying to find the place where your brain, your experience, and your effort actually matter.

Don’t sell yourself short. Share your work. Ask the real questions. And don’t let the shiny stuff distract you from what matters most.

Whether you’re the one interviewing or being interviewed, here’s the core truth:

Great answers are everywhere. Personal experience isn’t.

Anyone can Google “What is a promise in JavaScript?” But not everyone can talk about when they used one, why it mattered, what went wrong, and what they learned from it.

That’s where the real story is.
That’s what builds trust.
That’s what makes a hire that actually sticks.

See you next Monday,
Robin

#gorogue

Enjoying The Rogue?
Subscribe, share, and forward to a friend who needs this.
Or reply and let me know what topic you want me to cover next.

📰 Newsletter Archive
🎙 Go Rogue Podcast
📧 [email protected]

High-growth startups thrive when every hour points toward product-market fit. In this month’s Go Rogue episode, CTO Jason Whitehorn unpacks how modern AI can buy those hours back without diluting vision or culture.

The question is not whether to “do AI.” It is whether you can afford to let high-leverage talent keep grinding on tasks a model can finish in minutes. Start small, codify learning, then reinvest the freed bandwidth into shipping features that set your startup apart.

Why it matters: