Hacker News Reader

But if that's the case why put so much focus on and effort into the peer review system?

If you ask people funding research I'm pretty sure they'd prefer to fund influential ideas than non-influential "high-quality" paper production.

View on HN

Even if you were to take the extreme position that influence or citation counts are all that matter, the problem is that 'future influence' is hard if not impossible to judge well in the present. (Of course it's easy to point to cases where it's possible to make an educated or confident guess.)

Also, an emphasis on clarity and intellectual depth/rigor is important for the healthy development of a field. Not for nothing, the lack of this emphasis is a pretty common criticism of the whole AI field!

High-quality writing improves information dissemination. A paper like word2vec has probably been skimmed by 10's of thousands, perhaps 100's of thousands people.

One extra day of revising is nothing, comparatively.

Reply on news.ycombinator.com