I can guarantee he was not paraphrasing Grothendieck and, if you've seen the movie, you'd understand it's not some cynical performative humility either.
You should actually read the article. In addition to that great mask-slip moment I mentioned, you'll see yet another reason why he's malding about OpenAI's Navier-Stokes work and a few other fun Easter eggs!
It's interesting that you mention Grothendieck tho. I don't suspect Tao is much of a fan. No doubt, there are plenty of honest mathematicians of reasonable character, especially of the "problem-solver" and applied variety, that don't exactly worship at the alter of Alexander Grothendieck for reasons that are essential aesthetic or practical, but I would say he's the archetypical mathematician that people like Terence Tao specifically endeavor to gatekeep from contemporary academia and, ultimately, erase from history altogether.
Based on a very recent guest post on his blog by Noah Giansiracusa:
https://terrytao.wordpress.com/2026/09/16/let-the-diners-int...
Tao believes the works of Grothendieck (mostly as written by the mathematicians that became the Bourbaki group), the Bourbaki group, and those that followed are an active liability to their gatekeeping mediocrity cartel in the AI era. So maybe we can look forward to him making all that even more explicit in the near future!