AI and the Research Team

URL SCAN: AI and the Research Team
FIRST LINE: Economics > General Economics

The Dissection

This paper turns labor displacement into a team-design problem. It shows AI first enlarges research teams by making execution cheaper, then hollows them out as codifiable tasks become automated. The result is not evidence of durable human economic necessity. It is a local measurement of the transition: output and laboratory scale can rise while the human member share collapses.

The Core Fallacy

The central error is treating team size as a proxy for human productive participation. A peak in team size is not a floor beneath employment. Once automation coverage passes the peak, the remaining researcher may be an owner, controller, verifier, or temporary bottleneck—not a broadly necessary worker. The paper’s mechanism supports P1 and P3: AI can expand the system’s productive capacity while making most human labor economically redundant.

Hidden Assumptions

• Research demand and funding remain organized around human-led institutions.
• “Irreducibly human” tasks remain irreducible instead of becoming codifiable later.
• AI capability rises smoothly enough for a quasi-concave model to describe it.
• Research-grade output implies continuing economic necessity for researchers.
• Displaced team members can access ownership, compute, data, or distribution.
• Team size, rather than wages, labor share, ownership, or control of AI capital, is the decisive variable.

Social Function

Partial truth and transition management, wrapped in prestige formalism. The paper gives institutions a clean story for contraction: teams grow first, peak, and then shrink. That is analytically useful, but it also launders a power transfer into an organizational curve. The people disappearing from the team are not necessarily disappearing from the output; they are disappearing from the claim on its value.

The Verdict

This is not a rebuttal to the Discontinuity Thesis. It is a microeconomic autopsy of one of its mechanisms. Research teams may reach their maximum human size between 2026 and 2030 while research output continues expanding. The team survives; mass participation does not. The “peak” is the last visible ridge before human execution becomes overhead.