Something ridiculous like 90% of arXiv papers get lost in the sheer volume of submissions - 90% of it could be wrong (that's a made up number, sue me).

Just like flat earth or anti-vaxer theories on YouTube, incorrectly published work could be damaging and dangerous.

Getting peer-reviewed & published legitimizes the research work. It has been verified by a panel of experts.

Good for those start ups. I can't wait to attack all their Deep Learning APIs [1].

[1] https://openai.com/blog/adversarial-example-research/

Absolutely. None of the modern advancements in knowledge and understanding involve the waste of time known as "peer-review". See http://timecube.2enp.com/.

We know peer review is unnecessary for quality and progress in science because it only became a standard after WWII. Einstein published one peer reviewed paper.

There are multiple fields that have routed around peer review; everything that works on preprints and working papers. It takes more than two years to get an Economics paper published. If it took that long for things to get eyeballs progress would be exceedingly slow.

Peer review may have some good points but it certainly isn’t necessary to modern advancements in knowledge and understanding.

That one economics paper, as an example, could have made one seemingly innocuous assumption that might be overlooked by later work.

Let's say a paper that relies on the preprint take that assumption for granted. And then another paper references that one. And again. And again.

Then the fifth paper gets adopted by a government in some esoteric policy, which causes their economy to implode.

May not be necessary to just publish "stuff", but it is necessary to mitigate against the dangers of invalid assumptions.

Peer Review has not stopped the situationyou describe from occurring in multiple fields. Social Psychology is particularly bad but cancer biology isn’t great either.

——-

Why Most Published Research Findings Are False

https://journals.plos.org/plosmedicine/article?id=10.1371/jo...

Bayer and Amgen, reported dismal results when they tried to reproduce some cancer papers. Only 25 percent of the papers Bayer examined were reproduced. Amgen was able to replicate only six out of the 53 studies it examined.

https://www.npr.org/sections/health-shots/2017/01/18/5103048...

Plan to replicate 50 high-impact cancer papers shrinks to just 18

An ambitious project that set out nearly 5 years ago to replicate experiments from 50 high-impact cancer biology papers, but gradually shrank that number, now expects to complete just 18 studies.

https://www.sciencemag.org/news/2018/07/plan-replicate-50-hi...

"Proper" peer review could mean formal peer review which, indeed, has a lot of flaws. In the context of OP, however, just having any academic look at it and suggest improvements would probably already be a huge boon, regardless of potential funding or challenge: it achieves better knowledge discovery. Either form of peer review is pretty hard to obtain through arXiv.

> So, why would you want "proper" peer review nowadays?

Because I want other people to help catch my mistakes. Why wouldn't you want that?

Just send your results or ideas to someone knowledgeable about the topic and ask their opinion.

This.

If you work in the field you know the 10-100 people who provide most expertise and you can build trust relationships with them by delivering interesting thoughts and then get their feedback on These ideas. That's the healthy approach. And if you work hard and the stuff you do is reasonable you will also become known in these circles.

Also technology can help weed out the complete b.s., see HN or Stackoverflow for examples on that.

But this is what peer review is - sending it to the people in my field for their feedback.

It’s just set up formally so you don’t need to ask everyone individually and there is (an attempt at) the feedback being blind so that people who are unpopular don’t get penalised.

That’s literally what peer review is - but set up as a system to make it easier on everyone.

No it's not. Peer review is often done by people have have no clue about what you're doing. And even if they do, not many are motivated to make effort to provide a good review. Source: I received comments from 4 reviewers for my last paper, and not a single one provided any useful feedback (paper's accepted). Two of them simply copy-pasted sentences from the abstract and called it a day.

I've had a similar experience lately and it's frustrating. Literally 1 of 9 reviewers gave meaningful feedback. The rest either had no comment or such vague comments to be useless. Other published papers had comments that just regurgitated the abstract in an effort (I assume) to demonstrate they read the paper. I really wonder if many reviewers just "check the box" so they can say they review for X Journal

> Peer review is often done by people have have no clue about what you're doing.

That's not how I have ever seen peer review done by a reputable conference or journal. Why would they be reviewers if they didn't know the topic?

What do you think the word 'peer' means?

In my field the people I'd ask informally for comment are exactly the same people who would peer review my paper.

The word “peer” means a random name from the references section in my paper, picked by the editors who themselves are very unlikely to be familiar with my area. Out of 50 references, I think only 5-10 are able to review my work.