Nature: In 2000 the Clay Mathematics Institute announced it would give $1 million to whomever solved any of seven difficult math problems. So far, only one of the Millennium Prizes, the Poincaré Conjecture, has been cracked, although the victor, Grigori Perelman, declined the prize money. Now, as Nature‘s Geoff Brumfiel reports, Vinay Deolalikar, a researcher at Hewlett-Packard, claims he has solved another of the prize problems: Whether one kind of difficult-to-solve computation, NP (for non-polynomial), can be broken down into another, easier-to-solve kind of computation called P (for polynomial). Deolalikar’s proof says no.