"It would be very discouraging if somewhere down the line you could ask a computer if the Riemann hypothesis is correct and it said, 'Yes, it is true, but you won't be able to understand the proof'"
About this Quote
Graham is writing from inside a culture that treats understanding as the prize. Mathematicians don’t just want results; they want concepts that travel, techniques that generalize, intuitions that can be taught. A machine-generated proof that’s formally correct but cognitively opaque would sever that pipeline. It would turn the Riemann hypothesis into a black-box theorem: certified, uselessly so.
The line also anticipates an anxiety that’s now mainstream in AI: verification without comprehension. Even if we can check each step mechanically, “understand” means something else - compression into human-sized ideas, not a million-page argument or an alien formalism. Graham’s subtext is a defense of math as a communal, aesthetic enterprise. Without intelligible proofs, we don’t just lose an answer; we lose the discipline’s method of making new answers.
Quote Details
| Topic | Artificial Intelligence |
|---|---|
| Source | Verified source: The Death of Proof (Ronald Graham, 1993)
Evidence: It would be very discouraging if somewhere down the line you could ask a computer if the Riemann hypothesis is correct and it said, "Yes, it is true, but you won't be able to understand the proof." (pp. 92–103 (Vol. 269, No. 4, October 1993); quote appears within the article (exact page within 92–103 not confirmed from accessible primary text)). The earliest primary publication I could reliably trace for this wording is John Horgan’s Scientific American feature article “The Death of Proof,” published in the October 1993 issue (Vol. 269, No. 4), spanning pp. 92–103. Multiple independent quote indexes explicitly cite Horgan’s SciAm article as the source attribution for the Ronald Graham quotation. Scientific American’s own page for the article confirms the issue/date and that the article begins on p. 92; however, the full article text (and therefore the exact internal page number where the Graham quote occurs) was not accessible in the openly viewable snippet. Other candidates (1) The Riemann Hypothesis (Peter B. Borwein, 2008) compilation99.2% ... It would be very discouraging if somewhere down the line you could ask a computer if the Riemann Hypothesis is co... |
| Cite |
Citation Formats
APA Style (7th ed.)
Graham, Ronald. (2026, February 16). It would be very discouraging if somewhere down the line you could ask a computer if the Riemann hypothesis is correct and it said, 'Yes, it is true, but you won't be able to understand the proof'. FixQuotes. https://fixquotes.com/quotes/it-would-be-very-discouraging-if-somewhere-down-101914/
Chicago Style
Graham, Ronald. "It would be very discouraging if somewhere down the line you could ask a computer if the Riemann hypothesis is correct and it said, 'Yes, it is true, but you won't be able to understand the proof'." FixQuotes. February 16, 2026. https://fixquotes.com/quotes/it-would-be-very-discouraging-if-somewhere-down-101914/.
MLA Style (9th ed.)
"It would be very discouraging if somewhere down the line you could ask a computer if the Riemann hypothesis is correct and it said, 'Yes, it is true, but you won't be able to understand the proof'." FixQuotes, 16 Feb. 2026, https://fixquotes.com/quotes/it-would-be-very-discouraging-if-somewhere-down-101914/. Accessed 6 Mar. 2026.
