Hacker News Reader

Comment by nimish | Hacker News Reader

There's an issue this highlights and it's not that the original authors were stupid so much as there's clearly a lot of knowledge held in silos.

That's not a good thing if your goal is to advance everyone's knowledge. Whatever is going on in academia is failing relatively closely related fields which is not good.

View on HN

Is it really that siloed? The condition mentioned in the article (there being a global timelike Killing field) is discussed in all introductory texts on quantum field theory in curved spaces, it's even present in the first few paragraphs of the relevant Wikipedia article[1]. Even if it doesn't apply here, the authors ought to have mentioned why not.

I don't think they were stupid per se, nor malicious, but perhaps cavalier in pushing a result with such unexpected consequences without getting a consult.

1: https://en.wikipedia.org/wiki/Quantum_field_theory_in_curved...

It certainly wasn't in "silos", it's all on arxiv!

But yes: the world is complicated and it's easy to make mistakes outside your core field. The point of the scientific process is to get things in front of eyeballs who can spot the mistakes, c.f. the linked blog post. Then everyone fights about it or points and laughs or whatever, and the world moves on. The system worked.

What the process is not good at is filtering new ideas before people turn them into news headlines. And sure, that sucks. But it's not a problem with "academia failing", at all. The eyeballs worked!

Well, there's another aspect which is that the original authors and pop-sci journalists don't seem to be able to understand where they went wrong or how outrageous their claims are, precisely because their jobs depend on not understanding. The could have corrected it. We could not still be circling this drain 2 years later, but we are.

Kinda classic. Kinda boring.

It helps that this is a genuinely difficult process to understand and requires an enormous fluency with QFT. Most people who fit that bill have better things to do with their time than write popular science articles or correct them.

> clearly a lot of knowledge held in silos.

I don't think it's quite that, since the eventual goal is to publish, not only publicly, but as publicly as possible. More like it seems like everyone tends to hold their cards quite close to their chest until the moment of pre-print publication. Which means you can be working on something that someone could have told you months or years ago you have a problem.

The scientific equivalent of polishing a branch before making a pull request, only to be told "this has a huge memory leak and moreover what you want already works if you use this other API".

I'm not really sure there's a human-scale solution: the research landscape is so vast that you can't connect everyone to everyone else and have everyone in need of valuable input get it, and have everyone able to give it not be inundated with half-baked rubbish. Even if you assume everyone from the top to bottom has pure motivations and incentives for doing the research in the first place (in the pull request analogy CVE spammers, for example).

Perhaps not having the universities themselves so keen for PR that they'll slap a press release together about anything that looks clickable without due diligence would at least prevent making a public spectacle outside of the academic circle now and then, but it wouldn't solve the fundamental issues.

> There's an issue this highlights [...] there's clearly a lot of knowledge held in silos.

I think the real issue this highlights -- which is something everyone knows and still everyone does -- is that people love to spread and discuss sensational stories, and no one likes to hear naysayers ruining the fun.

Look the discussion of the original story here in HN[1]. There's a comment by A_D_E_P_T way down in the discussion explaining why the paper is nonsense and pointing to one of the replies objecting to it mentioned in the article from this post. That comment was downvoted by HN readers. I know because it was greyed out when I upvoted it days ago.

So there's no knowledge silo -- us simple folk just want to discuss the newest breakthrough without looking too hard, because that spoils the fun.

[1] https://news.ycombinator.com/item?id=43961226

I also think this kind of idea can be fun speculation, but I think there are better things to have fun that aren't promoting wrong ideas (like literal Science Fiction speculation!). When we can build fun on top, the physics of our reality doesn't need to be (academically) fun by itself :)

I don't think there's a lack of skepticism on HN of all places. Every article that gets posted that discusses even a mild scientific result brings at least one HN commenter out of the woodwork to dunk on it. You can bank on it--there is always That Guy who has to argue against it, whether he's right or not.

Also, the comment you reference was probably downvoted because of the tone, not because of some HN bias against naysayers. Starting out your comment with "It's nonsense." is about as conducive to a productive conversation as starting it out with "You're wrong."

The point of a statement like "it's nonsense" is to prevent a conservation that should not happen, because it will be dumb. It's the right thing to say iff it's correct.

I'd agree if it was just another arxiv draft. But, honestly I appreciate the clarity and brevity of the comment in this case. And I think that tone is warranted given the paper was published in a well known journal, lending it quite some credibility as clearly demonstrated by the high-stakes PR it received. Especially, since any retraction of that paper will likely not be followed up by the same articles.

Direct link to A_D_E_P_T's comment: https://news.ycombinator.com/item?id=43964524

It's a good comment, but too technical. It's difficult to know if it makes sense. I think it's good, but I'm used to read weird stuff in papers. Anyway, my level of general relativity is too low to understand all the details.

I skip that whole thread because I was expecting an overhyped result and I have to sleep from time to time https://xkcd.com/386/ . I'd have upvoted that comment, especially if it was gray.

The comment is like ELI35[1], but for HN it's better to write a ELI25[2] version. Or perhaps a ELI25 introduction and a second ELI35 part with even more technical details. (I never liked ELI5[3].)

[1] I just finished my postdoc in General Relativity.

[2] I just finished my major in Geology. I know atoms and calculus, but I have no idea what covariant is. Moreover, whatever gauge means is not the type of gauges I know.

[3] I just want a lollipop.

I think it's just an intractable problem at this point. There's probably millions of physicists on Earth. Everyone working in a company knows just how hard it is to get even hundreds of people to agree and read and understand the same things.

The fact is, there are just too many people doing too many things. When any technical paper sounds like gobblygook to even people in the same field but in a different specialty, it's no surprise this happens, especially when coupled with the modern pressure to scientifically publish and modern "journalism" trends.

99.999999999% of people do not have enough knowledge to even dream of beginning to understand a majority of research. Adults can barely read, much less be able to pass Calc 1.

Wher are you hiding 92 Billion people?

That percentage of the human population is everyone.

I guess they left out a bit for Einstein's brain (but not the rest of his body).

A Lot of these physics papers are interesting but ultimately just noise. An untested Theory is NOT fact, it's just someone (with or without a PhD) pulling something out of their arse that might explain things. Most of cosmology and physics is still theory (even the big bang, and string theory) and even if 90% of theory fits facts, they could still be wrong. I am seeing more and more of these un-testable theories, built on other un-testable theories, citing other un-testable theories, this is why theoretical physics is in a crisis IMHO.

MY mother and father also have an untested theory that explains all this too it's called "God", most Sci-Fi authors have plenty, and I am sure AI's will soon add to this pile.

Kudos to those scientists that create testable papers or experimentally prove stuff.

We call every model no matter how well founded a theory. An untested theory might be better called a hypothesis.

This mistake of language plays into the weirdos who opine we should disregard inconvenient ideas because they are just a theory failing to understand the difference between their own usage and the scientific usage.

Reply on news.ycombinator.com