If I think of PG's accomplishments they could be characterized as "discovering things which are new" to him, but also characterized as taking ideas that were cu

If I think of PG's accomplishments they could be characterized as "discovering things which are new" to him, but also characterized as taking ideas that were cutting edge, uncommon, or out of favor and bringing them to fruition (sounds more like engineering). I don't think of the YC structure as discovering anything inherently new but taking an approach that was not being done in the VC industry and trying to scientifically iterate on it. Bayesian spam filtering seemed a discovery to him, but it wasn't for science: someone else had already published a paper even. However, the results of previous attempts weren't good enough until PG focused on the problem and used a large enough corpus of data.