[panabee]

thanks for these insights. i was unaware of the competitive and gatekeeping dynamics.

it appears that peer review may be ideal for limiting shoddy science, but at the expense of limiting breakthrough science.

can you imagine if tech operated with these dynamics?

apple, google, stripe, and many others would have never passed peer review.

apple: idea for mobile phone. peer review: please add a physical keyboard.

google: idea for search engine. peer review: the 28th search engine? please do something novel like yahoo.

stripe: idea for payments. peer review: AJAX is for kids. please be serious.

[kuhewa]

Well, that's the fundamental difference between tech and science. With tech the 'truth' is entirely instrumental - is the product useful? (Not entirely accurate, as a product could be far less useful than the alternatives but still a commercial success).

In science sometimes the goal is instrumental value, but more often it is inferential insight where there isn't a simple 'it works or it doesn't' truth value and the role of methodology and review to control for sources of false positives and false negatives, misconduct, and unwarranted interpretation of data are important.

I'd argue that peer review aids breakthrough science overall, because where shoddy but splashy research slips through review, sometimes years of research effort and funding get funneled into avenues opened by such putative breakthroughs that turn out to have been bullshit all along. The misdirection into dead ends has the opportunity cost of the potential of making real breakthroughs.

[eru]

> it appears that peer review may be ideal for limiting shoddy science, [...]

No, it's far from ideal for that. Just reread the grand-parent comment.

Btw, putting your paper on arxiv is an alternative to traditional journals.

They call them 'pre-prints' there, but it's essentially a model where you publish first and then have the peer review from any interested party.