Why the Tech Hiring Market Is Broken

The tech hiring market is fundamentally broken.

Not in a “could be better” way. Much more in a “this is wasting everyone’s time and sanity” way.

If you are hiring, it feels like you are drowning in applications and still not finding the right person. If you are job searching, it feels like you are shouting into an empty room of companies that are apparently hiring and repeating the same calls and coding tests over and over.

Here’s the uncomfortable truth: both sides are reacting rationally to a system that has become outdated.

The real problem: noise (too many applications, too little signal)

Application volume has exploded, and the signal has not kept up.

Business Insider recently described this as “congestion,” and cited data from Greenhouse showing an average of 242 applications per job, up around 3x since 2017, plus recruiter workloads that can look like hundreds of applications per recruiter.

At the same time, applying got easier. Much easier. One widely cited metric is that applications on LinkedIn surged more than 45% year over year, and LinkedIn has been reported as averaging about 11,000 applications per minute.

This is the core failure mode: when the funnel gets flooded, everyone reaches for shortcuts.

Why it feels terrible for companies

If one role pulls in hundreds of applicants, you do what any overloaded system does: you drop packets.

Recruiters screen faster. Hiring managers rely more on referrals. Teams add extra steps because they are scared of a bad hire. Interview loops stretch out, calendars fill up, and time-to-hire drifts from weeks toward months (It was bad before, but it has gotten even worse!).

Ashby’s talent trends reporting shows that interview time has increased materially, with technical roles taking more interview hours per hire than before.

Then you add AI into the chaos. Not “AI recruiting software” on the company side, but the reality that candidates can generate tailored CVs and cover letters quickly, and even apply at scale. Bright Network found GenAI usage for applications rising fast among students and graduates (from 38% in 2024 to over 50% in 2025).

When volume goes up, quality assessment becomes harder, not easier. That is why hiring teams feel like they are working more and learning less.

Why it feels terrible for candidates

Candidates are also reacting rationally.

If you believe each application has a tiny chance of being seen, you apply to more roles. If you expect ghosting, you stop investing emotionally. If every company asks for a different version of the same story, you start copy-pasting.

Greenhouse’s 2025 Workforce and Hiring reporting captures the mood: only 7% of candidates felt the market favored them, and a large share reported that the job they applied for ended up being different from what was offered.

And then there’s trust. “Ghost jobs” and unfilled postings add another layer of frustration, and major outlets have reported this as a real and growing problem.

Also, remote work intensified the bottleneck. LinkedIn Economic Graph research shows demand for remote roles staying high while supply fell, including “more than 1 in 5” job seekers in the U.S. applying exclusively to remote opportunities. In practice, this concentrates applications into a smaller set of roles, which makes the flood worse.

So candidates feel ignored. Companies feel overwhelmed. Both sides blame each other. The system is broken.

The doom loop that keeps it broken

This is the loop:

1. Candidates get low response rates, so they apply to more jobs.
2. Companies get flooded, so they screen faster and rely on shortcuts.
3. Good candidates get filtered out, so companies add more steps and take longer.
4. Candidates lose patience and trust, so they apply even more broadly.

Nothing improves because the system rewards volume over clarity.

What companies can do

You do not need a brand-new hiring philosophy. You need a process that creates signal early and treats time like a scarce resource. More on that in our previous article on AI interviews

Start with the job post. If your role description is broad, your applicant pool will be broad. Write down the real constraints (must-have skills, must-have outcomes, real scope) and remove the vanity requirements. This alone reduces noise (In our own real world experience).

Add a “proof of fit” step that takes 15 minutes. Ask for one or two concrete artifacts that match the role:

• a short write-up: “Here is a system I built, here’s what I shipped, here’s what I’d do differently” (This is difficult, because it'll likely get generated again)
• a link to a repo, architecture note, postmortem, or technical blog post
• a small “debug this failing test” project

The goal is not to create homework. The goal is to give good candidates a way to stand out without playing the volume game. Adding a little bit of friction, to find candidates that are real and invested.

Make interviews smaller, faster, and more structured. Unstructured interviews do not scale in high volume markets. Use a rubric. Decide what “good” looks like before you meet the candidate. Keep the loop tight. Ashby’s data shows interview time is rising, so treating interviews as "free" is a trap. Especially because it compromises your engineers (interviewers) time that they could work on your actual product.

Be explicit about AI. If you allow AI in take-homes or live exercises, say so. If you do not, say so. Ambiguity creates bad incentives and worse comparisons. Some will secretly use it, skew your assessments and create awkward situations.

Close the loop with candidates. You cannot give bespoke feedback to 1,000 people, but you can do better than silence. Even a short rejection reason category helps candidates improve and reduces resentment.

What candidates can do (without applying to 200 roles)

You can’t fix the market alone, but you can stop playing the game that hurts you.

Pick fewer roles and go deeper. Choose 10 to 20 roles where you actually match the constraints. Your goal is not “more applications.” Your goal is “clear signal.”

Lead with proof, not adjectives. Replace “experienced” with:

• a shipped project
• a measurable outcome
• a hard problem you solved
• a short technical explanation of a decision you made

Treat AI as an editor, not a mask. Use AI to tighten wording, check clarity, or find gaps. Do not let it erase your voice. Hiring teams are reacting to generic, templated submissions because they have to.

Ask for the process upfront. A simple question like “What does success look like in the first 90 days?” filters out vague roles quickly. It also pushes the company toward clarity.

Use the one channel that still cuts through noise: warm intros. This is not fair, but it is real. In a congested system, human trust signals matter much more.

The bottom line

The market is not broken because engineers got worse or companies got careless. There is no malicious intent.

It’s broken because the hiring funnel is flooded, the signal is weak, and the process was never designed to handle this level of volume, especially with AI accelerating application spam.

Fixing it starts with: Stop optimizing for throughput. Start optimizing for signal.

That is how you can reduce noise, speed up hiring, and make the whole thing feel human again.