Post-Cutoff

Jess Werk: ‘Cognitive Sanctuaries or: Manifesto of the Department Chair’ (guest post on Terence Tao’s blog)

Jess WerkBlog

Why it matters

A department chair’s practical plan for protecting PhD training from AI-driven productivity pressure, published by Tao alongside the Harvard CMSA PhD recommendations.

Summary

Guest post by Jess Werk, professor and chair of Astronomy at the University of Washington, on Terence Tao’s blog (Sept 27, 2026). She argues that productivity-based incentives combined with generative AI’s cheap output weaken the value of the PhD and push the costs onto students and early-career researchers. Her answer is the “cognitive sanctuary”: department-level structures that protect slow, iterative thinking and failure. Proposals in three groups: PhD processes (detailed defence rubrics with independent scoring, live unassisted checkpoints, no publication requirement for candidacy, clear AI-use boundaries for advisors and students), faculty rewards (mentoring weighted heavily in promotion, service credit for attending department events, teaching relief for committee work) and community (social events, colloquia with discussion time, a standard against undisclosed AI-generated text). The only concrete step attributed to her own department is a Sept 17 faculty retreat that discussed an AI scenario and agreed on four ideas. 13 comments when read on Oct 15, 2026 (via a fetcher summary).

Archived text

Blog: summary plus short quotes only.

  • “a cognitive sanctuary protects the thought process, for everyone in the department, but above all, for Ph.D. students.”
  • “the thesis is not the point of the Ph.D. The student is.”
Source: terrytao.wordpress.com/2026/09/27/cognitive-sanctuaries-or-manifesto-of-the-department-chair

Cited in

  1. Policy & safety 87 days after the cutoff

    Harvard CMSA summit: recommendations for the maths PhD in the age of AI

People in this post

Terence Tao, Professor of mathematics, UCLA