Local Bernstein theory, and lower bounds for Lebesgue constants

Terry Tao’s new math drop sparks debate as he credits ChatGPT for a key proof idea

TLDR: Terence Tao unveiled a new paper sharpening classic math bounds and, surprisingly, credited ChatGPT with a key argument. The comments split between excitement over AI as a real research tool and anxiety about “proof by prompt,” with memes and debate turning the post into must‑read math drama.

Mathematics superstar Terence Tao just posted a new paper on “local” versions of classic bounds—think tighter rules for how fast things can change and how wobbly your connect‑the‑dots fits get. Cool math, yes, but the fandom is buzzing because Tao casually notes that ChatGPT helped with a crucial step. Cue the collective gasp.

One commenter, clearly clocking the twist, highlights the line that “this latter argument was provided to me by ChatGPT,” and the thread erupts. Another points to Tao’s own words: he “gave it to ChatGPT Pro,” which recognized the angle and “gave me a duality-based proof.” Translation: the bot didn’t just vibe, it contributed a legit route to the result. Meanwhile, a third commenter dubs Tao’s approach “an algebraic jeweler’s loupe,” praising the zoom‑in, local focus—fans are swooning over the elegance.

But the split is real: purists worry about “proof by prompt,” asking if attribution lines are getting blurry. Pragmatists cheer that this is the new calculator—powerful, transparent, and wildly useful when guided by an expert. The meme brigade shows up with “co‑author GPT,” “LLM = Lagrange’s Little Machine,” and “Fourier? I hardly know her.” Love it or fear it, the vibe is the same: math just got a plot twist, and everyone’s refreshing the comments for the sequel.

Key Points

  • Terence Tao uploaded an arXiv paper titled “Local Bernstein theory, and lower bounds for Lebesgue constants.”
  • The work was motivated by an Erdős problem on Lagrange interpolation and modifies classical arguments to obtain local Bernstein-type inequalities.
  • The post recalls Bernstein’s classical derivative inequality for polynomials and related results for trigonometric polynomials and entire functions of exponential type.
  • Trigonometric polynomials of degree n are identified as functions of exponential type n, with growth O(exp(n|z|)) in the complex plane.
  • Local versions of Bernstein inequalities are used to address lower bounds for Lebesgue constants in interpolation settings.

Hottest takes

“this latter argument was provided to me by ChatGPT” — tensegrist
“try my luck giving it to ChatGPT Pro, which recognized it as an {L^1} approximation problem and gave me a duality-based proof” — _alternator_
“an algebraic jewelers loupe” — adampunk
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.