The Last Technical Interview: How AI Is Rewriting the Tech Hiring Playbook

For roughly fifteen years, the technical interview has followed a familiar script. A candidate logs into a shared coding environment, an interviewer pastes in a problem about reversing a linked list or balancing a binary tree, and a ticking clock decides whether someone gets a job offer worth six figures. That ritual — equal parts hazing and meritocracy — is quietly falling apart.

The reason is obvious to anyone who has spent ten minutes with a modern AI assistant: the questions that used to separate strong engineers from weak ones can now be solved, fluently, by a chatbot. Recruiters know it. Candidates know it. And a recent scandal involving Anthropic — the AI lab behind Claude — has forced the industry to admit it out loud.

The scandal that broke the format

As Built In reported, Anthropic recently found itself at the centre of a debate after candidates were caught using AI during what were meant to be unaided coding interviews. The awkward irony — an AI company tripped up by AI-assisted cheating — has become a kind of inflection point. If even the firms building these tools can't run a clean whiteboard-style interview, the format may simply be done.

The Built In piece frames the question bluntly: is this the end of the coding interview? The answer, increasingly, looks like yes — at least in the form most engineers have known. When a candidate's webcam shows them looking down at a second monitor every few seconds, and the code arriving in the editor is suspiciously well-commented and idiomatic, the assessment isn't measuring what it claims to measure.

The deeper problem isn't cheating. It's that the line between cheating and working has dissolved. Most professional developers now use Copilot, Cursor, Claude or ChatGPT in their daily work. Asking them to solve a problem without those tools is a bit like asking an accountant to do a tax return without a spreadsheet — a test of a skill no one actually performs in isolation anymore.

The pivot: prove you're good with AI

The other half of the story is happening on the hiring manager side. Business Insider recently reported that a growing share of employers now want candidates to demonstrate AI fluency during interviews — not hide it. The framing has flipped from "don't use the tool" to "show us how you use the tool."

That shift matters. It means the technical interview is mutating from a memory-and-syntax exam into something closer to a working session. Hiring managers cited in the Business Insider piece described wanting to see how candidates prompt, how they verify AI output, how they spot hallucinations, and how they integrate generated code into a larger system. The valuable skill is no longer "can you write a binary search" — it's "can you direct a machine that writes binary searches, and catch it when it's wrong."

For Australian job seekers, this is a meaningful change. Local tech employers — from the big banks rebuilding their core platforms to Atlassian-adjacent product companies to government digital services — have been slower to formalise AI-fluent interviewing than their US counterparts. But the gap is closing fast. Expect to be asked, in the next twelve months, something like: "Here's a half-finished feature. Use whatever AI tools you like. Walk us through your thinking."

What the new interview actually looks like

If the old technical interview was a closed-book exam, the new one is more like an oral defence. A few patterns are emerging:

Notice what disappears: the LeetCode grind. The hours spent memorising sorting algorithms and dynamic programming patterns are, increasingly, hours wasted. That's bittersweet news for the thousands of junior developers who have built their entire interview strategy around it.

The credentialing problem

There's a parallel shift happening in how organisations think about hiring at all. Look at India's Tamil Nadu Public Service Commission, which recently announced 461 positions in its Combined Technical Services notification as non-interview roles, hired purely on written examination performance. It's the opposite philosophy: when you can't trust the interview, fall back on the test.

Western tech is unlikely to go that route. But you can see the same anxiety driving things like proctored online assessments, identity verification, and the resurgence of in-person final rounds. Some companies are quietly flying candidates to offices again, not because remote interviewing doesn't work, but because they want to see the candidate without a second screen.

The skills that suddenly matter more

Strip away the format changes, and a clear signal emerges about what tech hiring is actually selecting for now. It's not raw coding speed. It's the combination of skills that AI struggles to replicate: judgement, communication, and domain depth.

This echoes a theme well outside of software. New Civil Engineer recently interviewed an AtkinsRéalis technical director about a ten-year update to a highways management code — a reminder that in mature engineering disciplines, the prized expertise is contextual: regulatory knowledge, lived experience with edge cases, judgement built over decades. AI can summarise a code; it can't tell you which clause matters on a particular site in regional Victoria.

Software engineering is starting to look more like that. The most valuable engineers in the next five years will be the ones who can hold deep context about a system, a domain, or a user — the things a fresh chat session can't pick up.

What Australian candidates should do now

For anyone job hunting in the local market — Sydney, Melbourne, Brisbane, or remote-into-the-region — a few practical adjustments are worth making:

The technical interview isn't dying so much as growing up. For two decades it rewarded a narrow kind of cleverness — the ability to perform under contrived constraints. The version replacing it is messier, more conversational, and arguably fairer: it asks whether you can actually do the job, with the tools the job actually uses. That's a better question. It's also a harder one to game.

The candidates who thrive will be the ones who stop preparing for an interview that no longer exists.

Related on Bleen

Sources