← Writing

Not a Hiring Problem. A Context Problem. Again.

May 1, 2026

And AI is making that much harder to ignore.

Let me be clear: I don’t think AI is destroying the job market.

I think we’re in a transition.

And transitions are messy.

We’ve seen this before:

Every major shift creates uncertainty, resistance, hype, bad takes, opportunists, and real opportunities.

AI, like any leverage, multiplies what already exists.

Wrong process   × AI = -1 × 100 = -100
Neutral process × AI =  0 × 100 =    0
Good process    × AI = +1 × 100 = +100

If you’re getting worse outcomes faster, the multiplier is not necessarily the problem.

The system being multiplied probably is.

And hiring is one place where I think this is becoming painfully visible.

A Transition We Haven’t Fully Understood Yet

I’m currently in the market, talking to companies, going through processes, hearing stories, and reflecting on patterns I’ve seen before:

A company wants to hire an engineer.

Requirements are vague.

A polished job post gets generated (sometimes with AI).

Candidates use AI to optimize resumes, cover letters, portfolios, and answers — often against an already fuzzy understanding of the real need.

AI or automated filters evaluate candidates against criteria nobody fully validated in the first place.

Someone survives.

Then the ritual begins:

Eventually someone gets hired.

Then reality shows up.

Months later, the hire “fails.”

And everyone looks for someone to blame:

Sincerely, I think the issue is much simpler:

Too many hiring decisions are made without enough context from the people who will actually live with the outcome.

Again.

AI didn’t create that problem.

It accelerated it.

Oh, and I’ve seen things.

A company hired a “rockstar” architect.

Books published. Strong reputation. The kind of profile that makes everyone go: wow.

A team was built around that capability.

Deadlines slipped. Expectations weren’t met.

Then one day, after months of work, the person simply closed the laptop and disappeared.

No goodbye. No explanation. Just gone.

Easy conclusion? “Bad hire.”

Harder question?

What exactly did the hiring process assess?

AI Is a Multiplier, Not a Magic Wand

This is a broader pattern I keep seeing in AI conversations — and one I’ve talked about before.

AI is not magic.

AI is leverage.

And leverage magnifies direction, not virtue.

A broken process with AI becomes a faster broken process.

A vague hiring strategy with AI becomes a scalable vague hiring strategy.

A disconnected decision process with AI becomes a clearer-looking mess.

That doesn’t mean “don’t use AI.”

It means:

Understand what you’re trying to improve before amplifying it.

What I Learned Building Teams

Years ago, while building teams, I ran into a similar problem — long before AI entered the conversation.

Hiring decisions were happening too far away from the actual work.

The people who would work with the hire every day had limited input.

Requirements were often abstract, incomplete, or simply wrong.

And “fit” was treated as instinct instead of something intentionally discussed.

So I tried something different.

Not perfect. Not universal. But pragmatic.

The core idea was simple:

The team should help assess the hire, because the team lives with the outcome.

Yeah, sounds obvious.

But the trap then — same as now — was diffusion of ownership.

“It’s not my decision.”

“I’m just following the process.”

“I’m evaluating what I was told to evaluate.”

And suddenly nobody owns the outcome.

A Practical Team-Centered Approach

The rough structure looked like this:

1. Define requirements with the team

Not for the team.

The people doing the work help define:

Not idealized bullet points.

Reality.

2. Let the actual team interview

3 to 6 people from the actual team.

A focused interview. 30–45 minutes.

Not performance theater.

Not a memorized question sheet.

A guided but authentic conversation.

The goal is understanding:

3. Calibrate against reality, not fantasy

This was the most important — and hardest — part.

Each interviewer scored relevant dimensions from 0–10.

But with one important rule:

5 = current team average.

Not perfection.

Not “industry standard.”

Your actual team.

A 10 is nearly mythical.

A 5 is honest reality.

Then average the scores.

Interpretation was directional, not absolute:

This only works with honest calibration.

A weak team pretending to be an 8 will reproduce weak decisions faster.

Structured judgment.

Not perfect.

The Unexpected Benefit

Better hiring decisions were the immediate and obvious output.

That was measurable.

That was visible.

But the value didn’t stop there.

Because the moment you ask:

“What does average actually look like for us?”

everything changes.

Now the conversation is no longer only about the candidate.

It becomes about the team.

Are we actually strong communicators?

Do we document well?

Do we collaborate well under pressure?

Do we really own outcomes?

Are we honest about our current level?

Do we even agree on what “good” looks like?

That’s where the interesting part starts.

Because now the hiring process becomes a mirror.

Not just a filter.

And suddenly you’re not only evaluating a candidate.

You’re evaluating:

That changes the conversation.

Even mistakes become useful feedback.

Bad hire?

Maybe the candidate failed.

Or maybe the team miscalibrated.

Or maybe leadership defined the wrong problem.

That’s still useful signal.

Skills matter.

But willingness to learn, ownership, and trustworthiness often outperform static expertise in fast-changing environments.

The balance depends on context.

Do you remember the last time a new hire joined and the whole team started moving forward?

Now remember the opposite.

My Point

I think AI will absolutely improve many processes.

Hiring included.

But tools don’t replace understanding.

If AI turns your hiring process into a -100 outcome, that’s not necessarily evidence the tool failed.

It may be evidence your system needs redesign.

The opportunity in this transition isn’t to reject AI.

It’s to rethink what deserves amplification.