Some interesting discussion here on the research of large labs overshadowing the less popular ones despite not being that different: https://www.reddit.com/r/MachineLearning/comments/hbzd5o/d_on_the_public_advertising_of_neurips/

What do you think about this? Is this a problem? My (probably very controversial) opinion:

While this sucks, I’m not sure if it can or should be fixed - researchers operate under the assumption that it’s all about merit and that research doesn’t need to be "sold" with PR - that's not only false but IMO also hurts research. 1/n

No matter what you create: A product, a painting, or a research paper, if you want other *humans* to care about it, it must be SOLD through storytelling, PR, etc. Peer review is no different. You are writing weird academic language to sell your idea to some random reviewers.

If academia moves into a direction where more emphasis is being put on selling what you've done, I don't think that's a bad thing. All markets work that way. It's fundamentally human. Trying to design merit-based systems to get around this backfires and results in people cheating

More emphasis on selling also means more accessible research for a wider audience. Good storytelling and presentation is a crucial part of selling your product.

A related question is attribution. I believe it’s strange to give credit to people who came up with an idea *first*. Who cares who came first? I want to give credit to people who put in the effort to make the idea work and present it in a beautiful way, driving adoption.

Anyone who has started a company knows that ideas are cheap and mean nothing. There are usually hundreds of other startups having the same idea. What matters is how well you execute on your ideas and how you distribute them. I don’t see why research should be different.

There are probably thousands of breakthrough “ideas” hidden somewhere on arXiv. They are not being followed up on because they were not sold and presented well enough. That we’re following certain paths now is mainly because those ideas were sold well to convince us they’re good.

TLDR; A “good” idea sold extremely well and adopted by a huge group of people has much more positive impact than an “excellent” idea that nobody understands and cares about, or the same good idea presented in an incomprehensible way 5 years earlier. Give credit to the former.