Hard-to-read ML papers are often just badly written, researchers argue

gabriberton · x · 2026-10-14

A thread on ML paper writing culture: Jasper notes that what he once took for his own lack of intelligence was often just poor writing by ML authors. Gabriberton adds that hard-to-understand work is frequently badly written, with some authors adding complexity on purpose to sound novel—most papers could be summed up in five sentences of method plus a few pages of results.

Original post →

Hard-to-read ML papers are often just badly written, researchers argue

More from Fun