Why Job Sites Are Disappearing?

I wrote an early version of this piece back in 2023, arguing that job boards were losing ground to social media, direct hiring, and applicant tracking systems (ATS). That argument still holds, but the landscape has moved fast enough since then that it’s worth updating rather than leaving as a historical snapshot.

Social Media, Consolidated

Social media as a hiring channel used to mean several platforms competing for attention. In practice, LinkedIn has absorbed most of that: it’s where employers post, where candidates network, and increasingly where the “job board” function itself has migrated to. The generalist job board’s original pitch, a wide net of listings in one place, is now something LinkedIn does natively, which is a large part of why the older, standalone boards have kept shrinking.

The knock-on effect is that a candidate’s public profile now does double duty as both resume and application. Recruiters search it directly instead of waiting for an application to land, which quietly shifts power toward people who are already visible: those with an active posting habit, a full work history, endorsements. It’s a worse deal for anyone who has spent a career heads-down rather than online.

Direct Hiring Hasn’t Slowed Down

Direct hiring, going straight to candidates through referrals, networking, or an employer’s own careers page, remains popular for the same reasons it was in 2023: it’s cheaper and gives employers more control over the process. If anything, it’s become easier to execute, since sourcing tools and CRM-style candidate pipelines have matured, letting smaller teams run a direct-hire process that used to require a dedicated recruiter.

For technical hiring specifically, this has pulled the conversation away from resumes entirely. A GitHub profile, a portfolio, or a piece of open-source work someone actually shipped now carries more weight than a bullet-pointed job history, because it’s harder to fake and faster to evaluate. That favors people with visible work to point to, and it’s a real disadvantage for candidates whose best work is locked inside a former employer’s private repo.

The AI Layer Nobody Asked For

This is the part that’s genuinely new. Generative AI has hit both sides of the hiring process at once, and the two effects compound each other.

On the candidate side, AI-written resumes and cover letters, and AI tools that auto-apply to hundreds of postings on someone’s behalf, have sharply increased the volume of applications hitting every open role. A single job posting can now draw an application flood that would have been unthinkable a few years ago, most of it tailored by a model rather than by the person applying.

On the employer side, ATS platforms have responded by leaning further into automated screening and AI-assisted ranking, since manually reading that volume of applications isn’t feasible. The predictable result is an arms race that mostly cuts humans out of the loop until very late: an AI drafts the application, another AI ranks it, and the first real human judgment often doesn’t happen until a live interview. Candidates have caught on and now optimize their resumes for the algorithm rather than the person eventually doing the hiring, stuffing in keywords a model will reward and a hiring manager would never write.

A related problem that’s gotten more attention: “ghost jobs”, postings kept live for pipeline-building, internal compliance, or to project growth to the market, with no real intent to hire against them soon. Some of it is deliberate; some is just a listing nobody remembered to take down after the role was filled or shelved. Either way, it erodes trust in job boards generally, since a growing share of what’s listed isn’t real, and a candidate has no way to tell which postings are live from the outside.

The Future Looks Narrower, Not Broader

The generalist job board, one giant list of everything, is the format under the most pressure, exactly because it’s the format the AI-driven flood hits hardest: maximum surface area for auto-apply tools on one side, maximum noise for ATS filters to wade through on the other. What’s holding up better is the narrower version: platforms built around one industry, one kind of work, or one specific audience, where the listings are curated enough, and the applicant pool self-selected enough, to still be worth the visit.

Narrower also means the ghost-job problem is easier to manage. A small, focused board can afford to actually check that a listing is live; a board with a million postings can’t.

That’s the bet behind piperpiperpiper, the job board I run on the side: global, remote-friendly work, not a firehose of everything everywhere. Reach and convenience are still real advantages for a job board, they just have to be paired with enough focus that the listings stay worth trusting, and enough scale discipline that a small team can still stand behind what’s posted.