The cost of getting a job today

TEXT START: This one is short, but it’s been bothering me for a while.

The Dissection

The article is a hiring manager’s ethical accounting of a new entry fee: unemployed candidates must purchase AI capability to prove they deserve income. It correctly identifies that this cost is uniquely extractive because the tool is consumed almost exclusively for the hiring process.

But the article confines a structural rupture to process design. Delaying the challenge, shrinking it, supplying the company’s model, and weighing code less are sensible local repairs. None alter labor scarcity, bargaining power, or the number of engineers companies actually need.

The revealing premise is that an LLM can already write most of the code. The article responds by relocating the hiring signal to architecture, communication, judgment, and AI supervision. That is adaptation to the first wave of cognitive automation, not resistance to it. The human interview becomes a filtration layer around increasingly machine-produced work.

The Core Fallacy

It mistakes a price externality for the main problem. The candidate’s subscription is not the disease. It is a small toll at the entrance to a labor market whose productive function is being automated. The employer can pay the toll, but that does not restore displaced demand for engineers.

Under the Discontinuity Thesis, P1 commoditizes coding, P2 prevents institutions from preserving large human-only economic domains, and P3 removes productive participation from the majority. Architecture, communication, trade-off analysis, and model verification are real residual tasks, but they are not permanent human property. They are the next targets for automation and can be performed by fewer people with better systems.

Hidden Assumptions

• Companies will continue hiring engineers at roughly meaningful volume even as AI raises output per worker.
• Human reasoning, communication, and architectural judgment will remain scarce and reliably distinguishable.
• Productivity gains will create enough new work to absorb the engineers displaced by automation.
• “Willingness to do the work” remains a durable advantage when surplus labor makes willingness abundant.
• The candidate’s main problem is access to tools, rather than declining access to economically necessary roles.
• Hiring-process redesign can solve what is fundamentally a collapse in labor demand.
• The startup’s small-team advantage will produce more opportunity for workers, rather than smaller teams with fewer workers.

The article also treats reimbursement and company-provided accounts as awkward liability questions. That is governance friction, not a structural impossibility. The employer can remove the immediate cost whenever it chooses. It cannot remove the larger competitive pressure to need fewer humans.

Social Function

Primary classification: transition management, with a partial truth wrapped in elite self-exoneration.

The article honestly acknowledges a new form of exclusion and proposes less damaging hiring procedures. That makes it useful. It also makes the transition psychologically manageable: structural displacement is converted into a checklist for making the gate fairer while leaving the gate, the employer’s control, and the shrinking number of seats intact.

It is not crude propaganda. It is an accurate complaint operating as ideological anesthetic. It identifies who pays for obsolescence without confronting whether the obsolete role should still exist at scale.

The Verdict

This is a sharp observation attached to a small diagnosis. The author correctly sees that AI turns hiring into a pay-to-enter contest and that code challenges are losing their value. The decisive failure is scale: the problem is not merely that job seekers cannot afford AI. The problem is that AI makes fewer job seekers economically necessary.

A live, employer-funded session is a humane temporary moat. It may help select useful Servitors. It does not preserve mass employment, worker leverage, or the wage-consumption circuit. The article documents the transition from “write code” to “manage and verify machine output,” then mistakes that first adaptation for a stable equilibrium.

The subscription fee is only the visible bill. The larger bill is the loss of the wage role itself.