What AI actually changes about hiring

Most AI-in-hiring coverage is about tools: a model that writes your CV, a model that screens it, a chatbot that schedules the interview the other two models arranged. Tool stories miss the economic event underneath, which is simple: the cost of producing a plausible claim has gone to zero.

Every hiring signal that is cheap to produce is now worthless, all at once. The polished CV, the keyword-perfect profile, the articulate cover letter, the confident written answer. None of these distinguish anyone anymore, because the marginal cost of all of them is a prompt.

Scarcity moved

When claims are free, verification is the product. The signals that still carry information are the ones a model cannot conjure: work that happened, in public or with witnesses, with outcomes someone else can check. A repository with four years of commits. An automation a business ran. A colleague who will say yes, they built that, it worked.

This is why we are unbothered about being an AI-era company that refuses to let AI assign a single capability. It is not technophobia. It is pricing. Letting a model write capabilities into profiles would mean manufacturing the exact asset whose value just collapsed, inside the one place trying to sell its opposite.

Where we spend the technology

We spend AI where drafting is the job: turning a real repository's README into a write-up its owner edits and claims, suggesting tags its owner must tick by hand. The model does the typing. The human does the claiming. The evidence does the scoring. Everything else about this platform is arranged so that sentence stays true, because in a world of free claims, the companies and candidates who can prove things are the only ones with anything to trade.