AI deciphers Gödel notebooks, revealing hundreds of proved theorems
A team led by Jan von Plato at the University of Helsinki used the AI model Claude to interpret Gödel's 16 workbooks from the early 1940s — 1,009 densely written pages in an obsolete German shorthand. Within days the model produced a digest showing Gödel had proved hundreds of previously unknown theorems, including 14 unexplored themes in logic and anticipations of results from the 1970s and 80s.