We're gonna need a lot more mathematicians: Amit Sahai's guest essay on Terence Tao's blog

On 24 September 2026, Terence Tao's blog "What's new" published a guest essay, "We're gonna need a lot more mathematicians", by Amit Sahai, a UCLA professor and cryptography researcher. The text, an editor's note says, was converted from another file format using AI; Sahai adds that "GPT 6 Astra" helped him draft it.

He opens with a memory from his undergraduate years: students who felt the pace at which the strongest students absorbed new mathematics was far too fast for them. Almost all of them eventually gave up on research mathematics. He calls it "a time for humility".

The pace of AI, and a time for humility

The AI systems he has worked with, he writes, already produce beautiful new ideas. To leave research mathematics because the machines move faster would be "a profound abdication of our responsibility to humanity".

A "deployable intellectual reserve"

His proposal: a multitude of research groups, each with sustained support, each spending a term or a year trying to understand an extraordinary set of ideas produced by an AI system. He expects this to be among the most important mathematical work of coming years, and says it will require a significant expansion in the number of mathematically sophisticated human researchers worldwide: a "deployable intellectual reserve" humanity can call upon.

The essay's central thought experiment is a one-terawatt nuclear fusion plant designed by a future AI, with robots ready to build it. Before approving construction, Sahai writes, human communities would need to understand why the design works and what justifies confidence in its safety; the novelty that makes it exciting means confidence cannot be inherited from decades of similar plants. A theorem exists only within a model, so understanding a guarantee means understanding the model and its uncertainties; letting AI decide alone would leave consequential decisions resting on reasons no human community understands.

Reactions: jobs, and a call for an operating plan

The post drew 44 comments and 25 likes. Its most-supported comment (84 votes) called it "another monologue by a tenured faculty in which no concrete policies, ideas, or ethical standards are suggested". The next (71 votes) asked how more permanent jobs for mathematicians would be created, and whether faculty hiring could become "less selective and gatekeepy". A commenter with 15 votes said the academic bargain of low pay, unstable positions and extreme competition no longer looks worth taking now that AI exists.

Sahai replied himself: he is a co-organizer of a Simons Foundation effort that will soon release a report with "many concrete action items". He describes a "fork in the road" moment for the species: either yield more and more of the effort of understanding to the machines, or treat their thinking as a gift that deepens human understanding. Other commenters said an operating plan is urgently needed: concrete steps for universities, research institutions and governments.