Working from home will cost our kids their careers - i Newspaper

TEXT ANALYSIS PROTOCOL

1. THE DISSECTION

This is an economist's comfort artifact. McRae synthesizes a real empirical study—the LSE/Warwick paper on 243 million hires—to construct a narrative that lets the existing economic order off the hook. The rhetorical maneuver is precise: decouple the symptom (youth unemployment) from the cause (AI displacement) by attributing the hiring contraction to a behavioral/policy variable (WFH) that can theoretically be reversed. This is denial dressed in data.

The article's architecture:
- Opens with the genuine crisis: rising NEETs, youth employment collapse
- Introduces the Yale/ Goldman Sachs AI narrative as the popular but wrong explanation
- Positions the LSE/Warwick study as the correct explanation—WFH, not AI
- Closes with a reassuring "ladder can be fixed" prescription

The entire piece pivots on a single sentence: "If this is right, and I think it probably is, it's actually good news for young people."

2. THE CORE FALLACY

The author is diagnosing which bullet killed the patient, while the patient is already dead.

The LSE/Warwick methodology identifies WFH as a "better predictor" of hiring decline than AI adoption because WFH is a measurable policy variable and AI adoption is diffuse and hard to isolate. This is a methodological artifact, not a causal finding. The paper itself admits that "jobs most exposed to AI automation are also the occupations most amenable to remote work." McRae ignores this qualifier entirely.

The fallacy: treating WFH as the independent variable when it's actually the delivery mechanism for AI displacement.

Why did companies embrace WFH? Because digital tools—cloud platforms, async collaboration, algorithmic management—made it viable. Those same tools are the infrastructure through which AI displaces cognitive labor. WFH and AI are not competing explanations. AI is the cause. WFH is the symptom layer in which AI displacement manifests. McRae has confused the medium with the message.

The deeper error: assuming that if WFH is the problem, mandating office returns can restore the mentoring, apprenticeship, and informal knowledge transfer that historically built human capital. This assumes the problem is organizational, when it is structural. Even if every graduate showed up at the office five days a week, AI doesn't need them to do the cognitive work they're there to learn. The ladder isn't broken because young people won't climb it. The ladder leads nowhere because the destination has been automated.

3. HIDDEN ASSUMPTIONS

• Career as apprenticeship ladder still exists at scale. The entire prescription—"interact with more experienced workers," "show up," "the ladder may be broken but we know how to fix it"—assumes a functioning career structure into which new entrants can integrate. The Discontinuity Thesis says this structure is not temporarily impaired but structurally dissolving.

• Human mentorship generates durable economic value in a cognitive automation environment. McRae writes glowingly about mentoring as if the knowledge being transferred remains economically relevant. But if the tasks junior workers are being mentored to perform are themselves subject to AI replacement, the mentorship is training people for obsolescence, not careers.

• Policy can rebalance labor markets at scale. The "carrot and stick" prescription—force older workers back to offices so they mentor, mandate returns, tax incentives—treats institutional design as if it operates in a pre-AI competitive context. These interventions assume competitive dynamics that predate AI's capacity to undercut human labor across cognitive domains.

• Organizational adaptation is the rate limiter, not AI capability. McRae treats the problem as one of corporate inertia: companies haven't figured out how to train young people while enjoying WFH flexibility. The implication is that once organizations adapt, equilibrium returns. The DT says no. The adaptation IS the displacement. Organizations adapt by replacing human labor with AI. The "adaptation" McRae wants is the mechanism of destruction, not a solution to it.

• Feedback loops are symmetric. The complaint about employers not giving feedback to rejected candidates, and McRae's suggestion that universities teach "human interaction skills," assumes that human-to-human relationship building remains the scarce resource in the future labor market. It may instead be that the scarce resource is the capacity to manage, direct, and verify AI systems—competencies that current educational institutions are not remotely structured to develop.

4. SOCIAL FUNCTION

Ideological anesthetic for the managerial class.

This article performs a critical psychological service for its target readership—educated, middle-class adults concerned about their children's futures. It tells them:

• The problem is not irreversible structural displacement
• Their work-from-home convenience (and the corporate policies they benefit from) is the problem, which is fixable
• The solution involves older workers showing up, mentoring, and behaving fairly—essentially, returning to norms they remember and prefer
• The system is still legitimate; it just needs calibration

This is a lullaby. It says: your intuitions about meritocracy are still valid; the ladder can be climbed; your children's careers are not doomed; the problem is behavioral and reversible.

What it absolutely refuses to confront: AI does not need entry-level jobs to exist in order to eliminate the need for human workers at those levels. The question is not whether AI displaces entry-level workers today. The question is whether the economic logic that created entry-level jobs—who needs to be trained, mentored, and developed—is itself structurally durable. The DT says no.

5. THE VERDICT

McRae has identified the heat signature of the fire and concluded the solution is better ventilation.

The LSE/Warwick study is methodologically serious and empirically valuable. It accurately describes a real phenomenon: remote work is associated with reduced hiring of entry-level candidates. But McRae's inference—that this means WFH is the real problem and therefore AI displacement is less threatening than believed—is the analytical equivalent of noting that car accidents spike in winter, and concluding that ice is the real danger rather than vehicles traveling at speed.

The transition this article manages is not the transition from human labor to AI-mediated labor. It is the transition from acknowledging that transition to postponing it with a comfortable diagnostic story. It is transition management theater. The kind that buys comfortable middle-class parents another few years of believing the old order is recoverable.

It is not recoverable. The ladder is not broken. The game has changed.