Rebuilding the Hiring Marketplace

01 — The Problem

Online hiring was built on the idea that lowering friction improves the market: more job boards, easier discovery, one-click applications. That held while applying still cost something. A candidate had to find the role, tailor a résumé, write something, and that small effort was itself a weak signal of real interest.

That assumption is breaking. As generative AI pushes the cost of producing a credible-looking application toward zero, the upside to applying broadly stays high while the downside of an irrelevant application stays negligible. Applying everywhere becomes the rational move.

The cost does not disappear. It moves. Employers absorb it, through recruiting software, screening headcount, assessments, interviews, and the time lost to hiring delays. Candidates capture most of the upside from each additional application; employers absorb most of the cost of processing it.

That is a classic externality, and it is quietly degrading three things an employer actually needs to know: whether a candidate is a genuine fit for the role, whether they have real intent toward this opportunity specifically, and whether the résumé in front of a recruiter reflects real competence rather than something optimized, line by line, against the job description. What a candidate claims and what an employer can trust are pulling apart.

02 — Why Now

Two forces are colliding. The first is obvious: generative AI has collapsed the cost of producing a résumé, a cover letter, a tailored application, even the judgment of which jobs to apply to in the first place. Job discovery is already shifting from a candidate actively searching to software recommending, and increasingly, to software applying on a candidate's behalf. The natural scarcity that used to cap how many jobs one person could realistically apply to is going away.

The second is less obvious, and it is what makes this an opportunity rather than only a problem: verification is getting cheaper at the same time. Maintaining a structured, persistent, evidence-backed professional identity, employment history, competency signals, identity checks, used to be operationally expensive. AI and better infrastructure are bringing that cost down too. Remote assessments are simultaneously getting cheaper to run and easier to game, which is pushing real economic value toward whichever forms of verification are harder to fake.

AI is not only the cause of this problem. It is also what makes a different kind of hiring infrastructure newly affordable to build.

03 — What We Believe

Pull away the existing product categories and the two sides of this market want something simpler than what they currently get. An employer is trying to find a handful of people worth a real conversation: capable, interested, available, and credible enough to justify the time. A candidate is trying to find the right opportunity, not to maximize how many doors they have knocked on. Somewhere along the way, the industry started treating the application itself as the goal, and built a decade of infrastructure around producing and processing more of them.

That earlier assumption, that more applications create more opportunity, is proving hard to let go of, and not for fully rational reasons. Candidates keep applying past the point of relevance because the occasional story of someone landing a role despite being underqualified makes one more application feel worth the five minutes it costs, a bet whose long tail of failures nobody hears about. Employers keep treating a large applicant pool as a sign of a healthy pipeline, even when most of it is noise. Recruiting teams, facing volumes no person could reasonably review by hand, end up trusting the ranking algorithm more than the inputs it is ranking, even as those inputs get easier to manipulate.

We believe the platforms built on top of this system have mostly been treating the symptoms. Better filters produce better-optimized applications, which justify harder assessments, which get gamed too, which justifies more verification. Each layer makes the system more expensive without making it more honest.

What a hiring market looks like if it is built around restoring signal before volume, rather than processing volume faster, is the open question underneath this thesis.

04 — Customer & Use Case

The sharpest pain sits with recruiting and talent acquisition teams at companies hiring at any real scale: the ones now receiving hundreds of applications for roles that need one or two hires, most of them a mismatch, with no reliable way to tell which ten are worth a conversation without running the same filter-and-verify gauntlet on all of them.

On the other side are the candidates for whom this is working least well: people who are a genuine fit and are not applying everywhere, whose honest, specific application is now competing for attention against a flood of near-identical, AI-assisted ones. Today, the system has no good way to tell their signal apart from the noise.

05 — Market & Business Model

We are deliberately leaving this open. Who pays, how a venture built around this reaches its first employers and candidates, and what the actual product looks like are questions for whoever takes this thesis on, not answers we are supplying in advance.

What we do believe is something about the shape of the opportunity. Existing hiring platforms are built on metrics that reward activity: more listings, more searches, more applications. A platform that deliberately produces fewer, more credible candidates is optimizing against the metric the incumbents are structurally committed to. A platform producing 30 high-confidence candidates could be worth more to an employer than one producing 3,000 applications, even though it would look smaller on every dashboard the industry currently uses to measure itself. That gap, between what the market measures and what it actually needs, is where we think the opportunity sits.

06 — Existing Alternatives

Today, employers answer this with more filtering: applicant tracking systems, résumé-ranking software, automated assessments, identity verification vendors, and recruiters doing manual triage on top of all of it. Candidates answer the same problem with their own layer: résumé optimization tools, application-automation agents, and increasingly, AI assistance during the very assessments and interviews meant to screen them out.

Each layer is a reasonable response to the layer before it, and each one makes the overall system more expensive without making the underlying information more trustworthy. The two sides are now running AI against each other before a human conversation has even happened. The arms race is the product category, not a symptom sitting outside it.

07 — What We Would Test First

Before building anything, the thesis worth testing is narrower than a product: can a form of signal, whether that is verified competence, confirmed intent, or something else entirely, be made cheaper for a genuine candidate to produce than it is for an inflated one to fake? And will employers actually change their behavior around it once it exists?

If employers keep defaulting to volume even when a higher-trust alternative is available, the thesis does not hold, regardless of how well the product is built. That is the experiment to run before anything else gets built.

08 — Founder Profile

Someone who has sat inside the real cost of this problem, on either side: a recruiting or talent-acquisition leader who has watched application volume outrun every filtering tool thrown at it, or an operator who has built identity, credentialing, or trust infrastructure in another market and recognizes the same economic pattern here. We are looking for someone with a point of view on where real signal already exists in hiring today, even in small, unscaled form, and conviction about why nobody has made it the center of a product yet.

09 — Open Questions

• Does restoring signal require a new, standalone marketplace, or could it work as a layer that sits across the hiring platforms that already have the liquidity?
• Will employers actually pay for fewer, better candidates when today's internal incentives, a visibly full pipeline, headcount-justifying activity, reward volume instead?
• Can a verification approach stay cheaper to obtain honestly than it is to fake, as the same AI that threatens it keeps improving on both sides?
• What would prove this thesis wrong? If application volume keeps rising and employers keep tolerating it rather than paying for an alternative, the premise does not hold.