The world's most celebrated living mathematician says AI has changed the texture of his work without touching its substance. Terence Tao told podcaster Dwarkesh Patel that AI tools have transformed secondary tasks - literature searches, numerical plots, formatting - but the hardest part of solving a math problem still happens with pen and paper.
"The type of papers that I would write today, if I had to do them without AI assistance, would definitely take five times longer," Tao said in a conversation reported by "Hvylya", citing the Dwarkesh Podcast. "But I would not write my papers that way."
The distinction matters. Tao's papers now contain more code, more visualizations, and deeper literature surveys than before - additions that enrich the final product but do not advance the mathematical argument at its core. "Some plot which would have taken me hours to do, now I can do in minutes. But in the past, I just wouldn't have put the plot in my paper in the first place," he said.
When Patel pressed on whether AI had made him measurably more productive, Tao refused to commit to a single number. "They've really sped up lots of secondary tasks. They haven't yet sped up the core thing that I do," he said. He still uses an AI agent to reformat parentheses, run background literature scans, and handle administrative drudgery - tasks he described as freeing up time rather than replacing thought.
Tao drew a broader lesson for the field. If researchers measured papers written in 2020 against current output of the same functional depth, AI has not saved much time at all. What it has done is raise the floor of what a single researcher can include. "It's made the papers richer and broader, but not necessarily deeper," he said.
"Hvylya" previously explored how a finance graduate turned to tree pruning as white-collar jobs dried up across the United States.