Hiring in the AI Era: Why Recruitment Methods Always Lag Behind Technology, and What We Can Do Differently

A few weeks ago, I sat with a team I'd been working closely with. A team that had made a complete transformation - they work AI-first, barely write a line of code manually, ship features at a pace that used to take months.

I asked them how they hire new employees.

"Still with leetcode."

Wait, what?

"Yeah, we know it doesn't make sense. But we don't have an alternative."

That moment stuck with me. A team that crossed the Rubicon, that proved you can work completely differently - still stuck with hiring methods from 2015.

But the more I dug, the more I realized it's not their fault. This is a pattern that repeats itself in every technological revolution.

Most people don't know this, but the modern technical interview was invented at Microsoft in the 1990s. According to Wikipedia, Microsoft's style was unique in that it focused on technical knowledge, problem-solving, and creativity - unlike the "goals and weaknesses" interviews most companies used at the time.

The reason? Bill Gates loved puzzles. Simple as that.

An article on HackerRank describes how "brain teaser" questions became popular in technical interviews during the 90s and 2000s. Thomas Edison, by the way, did something similar long before - he asked candidates random trivia questions like "Where is the Sargasso Sea?" to test their knowledge.

In the early 2000s, Google took this approach a step further and pioneered the widespread use of algorithmic puzzles. According to research, this marked a shift from resume-based hiring to performance-based evaluations.

And that's how the world of leetcode was born.

When Apple's App Store launched in 2008, it had only 500 apps. By 2018, there were 20 million iOS developers serving 500 million weekly visitors to the App Store.

But how did they hire the first ones?

An interesting answer on Quora explains: "The existing mobile talent pool was totally unprepared for the iPhone. Prior to 2007/2008, mobile development meant using J2ME or BREW or writing real-time systems."

In other words - there were no iPhone developers before there was an iPhone. And recruiters? They had no idea how to evaluate skills that didn't exist two years earlier.

Recruiting firm Randstad Technologies reported a 104% year-over-year increase in demand for developers with mobile skills like iOS, Android, HTML5, Angular, Java, and JavaScript.

A Gartner survey found that demand for enterprise mobile apps would outstrip IT's capacity to deliver them by a factor of five to one.

A Deloitte study found that 90% of companies struggle to hire cloud talent, especially in multi-cloud, security, and AI-driven automation.

Roles like DevOps Engineer, Cloud Architect, SRE (Site Reliability Engineer) - didn't exist a decade ago. And recruiters? According to a report, 37% of IT leaders identify a lack of skills in DevOps and DevSecOps as the primary technical skills gap within their teams.

AWS research describes the situation clearly: "Managers expecting to hire their way out of skills problems face an increasingly scarce candidate pool. There are simply not enough tenured candidates out there to fill all open roles."

The pattern is clear: Technology moves fast. Hiring methods lag by 3-5 years. Everyone feels lost in the middle.

But this era also brings something that didn't exist before.

In startups it's most visible - problems that once seemed impossible are suddenly within reach. Code is no longer the bottleneck, customer feedback is. Development cycles that took months are shrinking to days.

And this opens up a crazy opportunity in hiring:

You can sit a candidate down with a real production ticket.

Not a theoretical puzzle - a real problem. With today's tools, a good candidate can within an hour:

• Understand the architecture
• Identify the bug or feature
• Propose solutions
• Maybe even write working code

But the real magic? We have a window into the entire thinking process.

Not just the final result - but:

• How they asked questions
• How they progressed when stuck
• How they created safeguards for the code
• How they tested the solution
• The level of precision and detail they achieved

Things that even take-home assignments never revealed - because there you only see results, not process.

An article on InformIT quotes a manager who said something that stuck with me: "We don't hire people for what they know, but what they can learn."

And someone else added: "As a programmer, you don't have a single 30-year career. You have 30 one-year careers."

This is a completely new hiring tool. And it still doesn't exist in most companies.

Now for the other side.

The question I get most often: "Is everything I learned about backend going to waste?"

Honestly, I don't have good answers. I don't pretend to understand what the future will look like.

But I do know one thing:

Me and many others - from the community and from companies I've worked with - can testify: even if the models plateau here and these are the strongest ones that will ever come out, there's no reason to go back to writing code manually. There's no reason not to work in parallel.

The transition happened.

So what do we do?

When Steve Jobs killed Flash in 2010 (with his famous open letter), researchers tracked what happened to developers whose skills suddenly became irrelevant.

The research that was published revealed fascinating insights:

"Those who adapted were thinking about the long term and were capable of learning by doing. They were quick to abandon skills with no perceived future."

What was their main strategy?

Learning new skills through projects they wanted to learn.

Not courses. Not certificates. Projects.

The research also found that developers relied on information from other programmers, technical discussion boards, and signals from industry leaders at large technology companies to figure out which technology would be the new standard in their field.

But here's the interesting thing: there was no flood of unemployed candidates. Wages barely changed, mainly because workers adapted quickly and pursued other skills.

There's a book that's been guiding me for the past year - "Why Greatness Cannot Be Planned" by Kenneth O. Stanley and Joel Lehman. Both are AI researchers (Stanley worked at OpenAI).

The central idea: Great things cannot be planned.

Not because we're not smart enough - but because the stepping stones that lead to great achievements don't resemble the destination.

A quote from the book:

"Almost no prerequisite to any major invention was invented with that invention in mind."

Think about it: if someone 1000 years ago had set a goal to invent a computer, they probably wouldn't have started by researching vacuum tubes - which were a critical stepping stone to inventing the computer. Vacuum tubes were invented as part of research into electricity and radio waves, not as part of a vision of computing.

The central metaphor in the book:

Imagine you're crossing a wide river, and the far bank is covered in fog. You can't see the path ahead. All you can do is step on the next stepping stone - the closest one, the most interesting one - and see where it leads.

Another quote:

"Objectives are well and good when they are sufficiently modest, but things get a lot more complicated when they're more ambitious. In fact, objectives actually become obstacles towards more exciting achievements, like those involving discovery, creativity, invention, or innovation."

Who knew when I started, when I told myself I'm not writing lines of code anymore, that I would lead a community? That I would lecture at conferences? That I would run processes for companies and deploy agents in production?

I didn't know. I just stepped on the next stone.

Be on the edge. Be curious.

Skepticism, sarcasm, and mental rigidity - hurt only those who operate that way.

Research on skill atrophy shows that skill decay isn't limited to technical abilities alone - soft skills like communication, leadership, and adaptability are equally prone to obsolescence if not regularly honed and updated.

Another study found that:

"The more people leaned on AI tools, the less critical thinking they engaged in, making it harder to summon those skills when needed."

But that doesn't mean avoiding AI. The opposite - it means using AI to reach new places, not to stay in place.

Follow your curiosity.

We live in a confusing era.

Backend developers are building LLM pipelines. Analysts are building dashboards in React. The roles are blending.

A CIO report warns:

"AI will increasingly eliminate low-level software development jobs, and machine intelligence will become the default for writing most modular code along with documentation."

But the same report also says:

"For centuries, jobs and enterprises have evolved with technology, and preparing for change requires a forward-looking mindset that's rooted in education, literacy, and the best way we can leverage new tech tools."

This is exactly why I built a job board on the Squid Club website.

There's no shortage of job boards in the world - but there is a shortage of a place where recruiters and candidates are looking for the same thing in an era where everything is still unclear.

Recruiters - feel free to post positions.
Job seekers - feel free to check out the board.

The board is updated frequently.

1. The pattern repeats itself - In every technological revolution (internet, mobile, cloud, AI), hiring methods lag 3-5 years behind the reality of work.

2. There's a new opportunity - Today you can see a candidate's entire thinking process, not just results. This is a hiring tool that didn't exist before.

3. The Rubicon has been crossed - Even if models plateau, there's no reason to go back.

4. The lesson from Flash developers - Those who succeeded learned through projects, not courses. They were quick to abandon skills without a future.

5. Great things cannot be planned - Just step on the next stone, the closest one, the most interesting one.

6. The main tip - Be on the edge, be curious. Skepticism and mental rigidity hurt only those who operate that way.

• The History and Future of the Coding Interview - HackerRank
• A History of Coding Interviews - Medium
• Coding Interview - Wikipedia
• 25 Years of Developer Interviews - Educative

• The Software Developer Shortage - Alpha Software
• Demand to Hire App Developers - Hyperlink InfoSystem

• Winning The Brutal Talent War - YUPRO
• Is There Still a Cloud Skills Gap? - TechTarget
• Essential DevOps Statistics - Brokee
• The Business Benefits of Hiring Early Career Cloud Talent - AWS

• What Happens to Tech Workers When Their Skills Become Obsolete? - Quartz
• Your Technical Skills Are Obsolete: Now What?

• Skill Atrophy Alert - Emeritus
• Avoiding Skill Atrophy in the Age of AI - Addy Osmani
• 5 Dead-End IT Skills - CIO

• Amazon
• Goodreads Quotes
• Book Review - LessWrong
• Review - Engineering Ideas Substack

Written by Sahar Carmel, Principal AI Engineer and founder of Squid Club community.

Want to join the community? Squid Club