Close Menu
The Financial News 247The Financial News 247
  • Home
  • News
  • Business
  • Finance
  • Companies
  • Investing
  • Markets
  • Lifestyle
  • Tech
  • More
    • Opinion
    • Climate
    • Web Stories
    • Spotlight
    • Press Release
What's On

Why IT Leaders Should Treat AI As A Teammate Rather Than A Tool

September 10, 2026

Wholesale inflation rises by strongest increase in 3 months

September 10, 2026

Your Cryptographic Inventory Is Lying To You

September 10, 2026

How Energy Leaders Structure Innovation

September 10, 2026

The Missing Layer In Agentic AI: Governing Autonomous Decision-Making

September 10, 2026
Facebook X (Twitter) Instagram
The Financial News 247The Financial News 247
Demo
  • Home
  • News
  • Business
  • Finance
  • Companies
  • Investing
  • Markets
  • Lifestyle
  • Tech
  • More
    • Opinion
    • Climate
    • Web Stories
    • Spotlight
    • Press Release
The Financial News 247The Financial News 247
Home » Things AI Still Can’t Screen For When You’re Hiring Engineers

Things AI Still Can’t Screen For When You’re Hiring Engineers

By News RoomAugust 14, 2026No Comments5 Mins Read
Facebook Twitter Pinterest LinkedIn WhatsApp Telegram Reddit Email Tumblr
Share
Facebook Twitter LinkedIn Pinterest Email

Elton Chan is the Co-Founder of Second Talent, a solution that connects global tech leaders with AI-native engineering talent across Asia.

The best hire I ever made would not have survived our current screening process. The worst hire I ever made cleared every stage of it.

That is not a criticism of the tools. Our screening stack works. It reads résumés faster than any human on my team, it catches inconsistencies I would have missed, and it has quietly removed a category of bias that used to live in the first 10 seconds of a résumé scan. If you are a founder hiring engineers in 2026 and you are not using AI somewhere in your funnel, you are paying a tax for no reason.

The problem is narrower and more uncomfortable. AI screens for what is legible. And the qualities that separated my good engineering hires from my expensive ones were never legible in the first place.

The Mistake I Keep Coming Back To

A few years ago, we hired a senior back-end engineer for a client platform build. Eleven years of experience, two of them at a name everyone in the room recognized. The take-home came back in the top decile of everything we had scored that quarter. Four interviewers, four clear yes votes, no dissent worth recording.

By month two, the pull requests were still arriving and the project was still slipping. The work was correct. It was also solving the problem exactly as written, including the parts of the spec that were wrong and nobody had said so out loud. We restructured the team in month five.

What I remember most is that nothing in the process was wrong. Every signal we collected was accurate. The candidate really did have the experience listed. The code sample really was clean. We had simply measured a set of things that turned out not to predict the outcome we cared about.

I have made a version of that mistake at least three times. Each time I responded by adding another stage to the process. Each time the process got longer without getting better.

What The Funnel Cannot See

Three things, specifically.

How Someone Behaves When They Are Wrong​

Every engineer interviews as though they are calibrated. You find out otherwise in week five, when a decision they pushed for does not work and you watch whether they surface it early or defend it quietly for a month. No take-home surfaces this. No AI-assisted interview transcript scores it, because it is not present in the transcript at all.

Judgment About What Not To Build​

This is the one I underweight most consistently. Screening rewards the candidate who solves the problem in front of them. It has no way to reward the candidate who would have questioned whether the problem was worth solving. In a small company, that second person is worth several of the first, and they screen worse, because the exercise you gave them was already scoped.

Whether They Raise The Standard Around Them​

Some engineers make the four people nearest them noticeably better within a quarter. Others are individually excellent and change nothing. On paper these two look identical. On a team of 12, they are not remotely the same hire.

I train jiujitsu, and there is a version of this on the mat. Plenty of people can demonstrate a technique cleanly with a cooperative partner. Far fewer can find it against someone resisting. The demonstration is legible. The other thing is what you actually wanted to know.

Why This Is Getting Harder, Not Easier​

Here is the part founders underestimate. Candidates have the same tools we do.

When every applicant can produce a clean, well-commented, well-tested take-home, the take-home stops discriminating. The distribution compresses at the top. This is not cheating, and treating it as cheating is a strategic error. It is what competent people do with available tools, which is exactly what you want them doing on the job.

But it means the signals your funnel was built on are decaying in value at the same time your funnel is getting better at reading them. You end up with a screening process that is more efficient, more consistent and less predictive than it was three years ago.

I see this at scale on the other side of my business. My organization screens roughly 830 applicants for every engineer we place across nine Asian markets, and over more than 10,000 placements the pattern is consistent: The technical assessment still sorts candidates, it just sorts them into a much tighter band than it used to. When the gap between the median submission and the top decile narrows, the stage stops doing the job you built it for, even though the scores look healthier than ever.

What I Do Differently Now

I have stopped trying to make the funnel detect these things. It cannot, and every stage I added in pursuit of that cost me good candidates who ran out of patience.

Instead: I let AI own the first half entirely. Sourcing, screening, scheduling, structured note-taking, consistency checks across interviewers. It is better than we are at all of it, and it does not get tired at candidate 40.

Then I spend the time I saved on one thing that does not scale: a working session on a real, currently ambiguous problem from our road map, where the scope is deliberately unclear and I am watching what questions get asked before any code appears.

The Honest Version

I am not confident this is solved. I think I have moved the failure rate down, and I think the failures I have now are different failures, which is progress of a kind.

What I am confident about is the direction. As AI absorbs more of the legible half of hiring, the return on founder attention shifts almost entirely to the illegible half. That is not a reason to use less AI in recruitment. It is a reason to be very clear about which half you are still personally responsible for.​​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?

Elton Chan
Share. Facebook Twitter Pinterest LinkedIn Tumblr Email

Related News

Why IT Leaders Should Treat AI As A Teammate Rather Than A Tool

September 10, 2026

Your Cryptographic Inventory Is Lying To You

September 10, 2026

How Energy Leaders Structure Innovation

September 10, 2026

The Missing Layer In Agentic AI: Governing Autonomous Decision-Making

September 10, 2026

Why Your Security Stack Was Never Built For This

September 9, 2026

Work Has A Supply Chain Problem. AI Just Made That Problem Urgent

September 9, 2026
Add A Comment
Leave A Reply Cancel Reply

Don't Miss

Wholesale inflation rises by strongest increase in 3 months

Business September 10, 2026

Wholesale inflation in August saw its strongest increase in three months, largely driven by higher…

Your Cryptographic Inventory Is Lying To You

September 10, 2026

How Energy Leaders Structure Innovation

September 10, 2026

The Missing Layer In Agentic AI: Governing Autonomous Decision-Making

September 10, 2026
Stay In Touch
  • Facebook
  • Twitter
  • Pinterest
  • Instagram
  • YouTube
  • Vimeo
Our Picks

Hunter Biden’s meme coin crushed 4 out of 5 buyers while one mystery trader raked in $1M

September 10, 2026

Hunter Biden’s $LAPTOP meme coin rockets from $3 to $300 — then crashes within minutes

September 9, 2026

Pokémon cards would be global currency during apocalypse, venture capitalist predicts

September 9, 2026

OpenAI says it cracked decades-old math problem — but professor claims they cribbed his work and then threatened him

September 9, 2026
The Financial News 247
Facebook X (Twitter) Instagram Pinterest
  • Privacy Policy
  • Terms of use
  • Advertise
  • Contact us
© 2026 The Financial 247. All Rights Reserved.

Type above and press Enter to search. Press Esc to cancel.