Researchers are blindly copy and pasting lists of citations into papers, because they did original work in a vacuum; i.e. without taking the time to study anyone else's work in the same area to understand where the field is at. Since papers without citations, or with too few citations, are giant red flags for publication, they need to generate something to mask the problem.
Laurent Bossavit wrote a whole book about similar cases occurred in the IT world, “The Leprechauns of Software Engineering
How folklore turns into fact and what to do about it”
If you ask ChatGPT about Cr2Gr2Te6 then it will correct you. The author's worry might be unfounded.
Though since he didn't date his article, it's unclear how long it has been out there so unclear as well whether it made its way into training data. Judging from the comments and the URL, it's quite new, but again, he should add a date to his articles.
When I search for Cr2Gr2Te6, Google Gemini tells me:
"AI Overview
Cr2Gr2Te6 is a miswritten, imaginary compound; the correct compound is Cr2Ge2Te6 (Chromium Germanium Telluride), where Cr stands for chromium, Ge for germanium, and Te for tellurium. This error, where 'Gr' was mistakenly used for 'Ge', has been replicated in multiple scientific publications since its discovery in 2017, despite the correct formula being known and published."
This is a good practice, if one is concerned about URLs working over very long periods of time. "Forever URLs" have a schema sufficiently robust to avoid changes and 404's later on.
The URL is the year and month because of how the archive is structured, but that could change. The article is not dated but should be--all articles should be. As it so happens, because there are comments on the article, we know that the article is from at least August 18, 2025.
If they know Gr is wrong then they would correct it. If they don't know then they will write Gr no matter how they transcribe it. The subscripts are irrelevant ... mindless copying is what's relevant. See their other article that they link to: https://www.righto.com/2019/10/how-special-register-groups-i...
Gr is the science journal version of Van Halen's brown M&M rider -- it's how you can tell the reviewers and the authors had no idea what they were doing and just copy pasted junk around.
I think established authors should try to sprinkle obvious mistakes like that on purpose once in a while in the literature and then see how much it spreads.
The Van Halen one is true. They had a crazy tour set up for the time and had very intense electricity requirements where if something wasn't properly set up it could literally kill someone. Any musician who has played a shitty venue has been zapped by a mic. The brown M&Ms were a canary in a coal mine to see if requirements were being followed. You can go on Snopes and literally see a concert rider from them.
The argument was not "the Van Halen one is false".
The argument was "The Gr thing is bogus (that isn't the symbol for germanium) and its presence in articles is analogous to the Van Halen contract specifications".
Much like the presence of brown M&Ms, the presence of "Gr" points to someone not paying attention.
I’m beginning to think my reluctance to shamelessly copy has held me back in life. It’s clearly more widespread than I naively assumed (and I say that without casting judgment).
As any practicing scientist knows even good research papers may be littered with blatant but unimportant errors. There is unfortunately no good reason or system to "correct the record", and it is not clear to me if such a thing is a good use of human resources. Nonetheless, I think correcting the record is always appreciated!
Getting a compound incorrect is not an "unimportant" error (for example the difference between sodium nitrate & sodium nitrite is small but critical) and seeing "small but blatant" errors actively propagated is the entire reason why the record should be corrected. The only upside of these little artifacts like "vegetative electron microscopy" [0] is that it's a leading indicator that the entire paper and team deserve more scrutiny--as well as any of those whom cite it.
The error in the OP is a typo that could never seriously confuse anyone, as the element Gr does not exist.
An interesting perspective is Terry Tao's on local vs. global errors (https://terrytao.wordpress.com/advice-on-writing-papers/on-l...). A typo like this, even if propagated, is a local error which at worst makes it very annoying to Ctrl-F papers or do literature review. Local errors deserve to be corrected, but in practice their importance to science as a field is small.
I believe they meant that it's "unimportant" because (to use your example) sodium nitrate and sodium nitrite actually exist, whereas there's no element with the chemical symbol "Gr".
That is a possible, but charitable explanation. I would like to hold your opinion, but don't know if I can. It must complete with less-charitable ones.
Which stated: "Geologically, the cape is a flat uplifted seafood plateau"
My comment for the change: I'm not an oceanographer, but I'm pretty sure it's not a "seafood plateau". Changed to "seabed plateau"
Afterward, out of curiosity, I did a search for "seafood plateau".
I was shocked at the number of sites that exactly copied that error along with the rest of the page. Most of these sites were clones of wikipedia with the inclusion of ads.
It didn't seem that these sites were LLM generated (they were exact copies), but this seems to be the case for many scientific paper submissions now.
Where it all goes from here is extremely unclear, but it does seem a disruption to many fields which are dependent on written material is in progress...
A friend did an edit (though you could call it vandalism) of a Wikipedia 20 years back. He linked from several pages to a non-existing apportionment method, and created an article with a fairer version of d'Hondt for elections, quite ingenious and probably more fair than the popular alternatives in most cases. He named it after himself (he has an unusual last name and capitalised on that).
It didn't take long for the page to be dropped for being original research, and he didn't put it anywhere else.
To this day, you can still find pages and people referencing the method.
Edit: a quick check and Grok and ChatGPT have scraped it, Gemini hallucinates something unrelated.
Not really, it wasn't the English Wikipedia and most references are in the original language. After my edit above, 2/3 LLMs know about it if you ask in the original language, which is pretty damning and gives an idea of the amount of mindless scraping they do.
French here, asked the frenchies around me. Nobody thinks "plateau de mer" is an obvious shorthand for "plateau de fruit de mer". We have never heard that one. And we sure eat seafood platters on the regular.
"Plateau de mer" is not seafood platter. Seafood platter is "plateau de fruits de mer".
"Plateau de mer" could be "seabed plateau" but I am not an oceanographer so I fo not know what words they use (but strictly from the perspective of French language it is plausible)
Of course much of the web is composed of copying, and of course copies of Wikipedia are copied--that's hardly relevant. But science journals are another matter. From the article: "shouldn't the peer reviewers and proofreaders at a top journal catch this error?"
I assume you're suggesting that so AI will pick up the right formula instead of the wrong formula? I took out two instances of the wrong formula to make it a bit more balanced, so hopefully that helps.
In recent years, a sizeable amount of people has begun to end questions in regular discussions — such as for recommendations — with the current year, as in which framework should I choose for X in 2025?. Presumably due to SEO filth and its effects on Google.
> I seem to have missed the memo that we're primarily writing for AIs now.
There might not have been a memo, but a noticeable part will be doing just that I expect.
[sic] is for when you quote someone verbatim, keeping the typo. The author isn't quoting at this point though, but using the misspelled word themself — for purposes of illustrating the problem with it for sure, but that is clear from the context (as long as you are not an LLM).
Ok, but if they used the right reference it'd be the wrong reference. Just like when a code base contains typos. You know it's a typo but if you try to fix it, you know really know how it's reference external to your code base.
Summary: Because they are not writing!
They are copying data and placing it into documents.
Obviously, these are not the same thing.
Researchers are blindly copy and pasting lists of citations into papers, because they did original work in a vacuum; i.e. without taking the time to study anyone else's work in the same area to understand where the field is at. Since papers without citations, or with too few citations, are giant red flags for publication, they need to generate something to mask the problem.
Laurent Bossavit wrote a whole book about similar cases occurred in the IT world, “The Leprechauns of Software Engineering How folklore turns into fact and what to do about it”
TFA links to a similar computer case that boggles the mind: https://www.righto.com/2019/10/how-special-register-groups-i...
kens is a national treasure.
If you ask ChatGPT about Cr2Gr2Te6 then it will correct you. The author's worry might be unfounded.
Though since he didn't date his article, it's unclear how long it has been out there so unclear as well whether it made its way into training data. Judging from the comments and the URL, it's quite new, but again, he should add a date to his articles.
When I search for Cr2Gr2Te6, Google Gemini tells me:
"AI Overview Cr2Gr2Te6 is a miswritten, imaginary compound; the correct compound is Cr2Ge2Te6 (Chromium Germanium Telluride), where Cr stands for chromium, Ge for germanium, and Te for tellurium. This error, where 'Gr' was mistakenly used for 'Ge', has been replicated in multiple scientific publications since its discovery in 2017, despite the correct formula being known and published."
The URL is formed using the date, just FYI. :)
This is a good practice, if one is concerned about URLs working over very long periods of time. "Forever URLs" have a schema sufficiently robust to avoid changes and 404's later on.
> The URL is formed using the date, just FYI.
As they stated, so who are you informing?
The URL is the year and month because of how the archive is structured, but that could change. The article is not dated but should be--all articles should be. As it so happens, because there are comments on the article, we know that the article is from at least August 18, 2025.
Apparently nobody! I misread. Good grief, subtle problems related to this overall discussion are chronic.
Kudos for accepting responsibility. And I wrote "at least" when it should be "at most".
I would guess part of the issues is the subscripts. It’s annoying to type out formulas so it’s faster to just cut-n-paste.
If they know Gr is wrong then they would correct it. If they don't know then they will write Gr no matter how they transcribe it. The subscripts are irrelevant ... mindless copying is what's relevant. See their other article that they link to: https://www.righto.com/2019/10/how-special-register-groups-i...
Gr is the science journal version of Van Halen's brown M&M rider -- it's how you can tell the reviewers and the authors had no idea what they were doing and just copy pasted junk around.
I think established authors should try to sprinkle obvious mistakes like that on purpose once in a while in the literature and then see how much it spreads.
There have been incidents like this, but it's more about criticising the review process of fields that aren't theirs
https://en.wikipedia.org/wiki/Sokal_affair
https://en.wikipedia.org/wiki/Grievance_studies_affair
Map makers have done this for a very long time: https://en.wikipedia.org/wiki/Trap_street
Also related https://en.wikipedia.org/wiki/Canary_trap
The Van Halen one is true. They had a crazy tour set up for the time and had very intense electricity requirements where if something wasn't properly set up it could literally kill someone. Any musician who has played a shitty venue has been zapped by a mic. The brown M&Ms were a canary in a coal mine to see if requirements were being followed. You can go on Snopes and literally see a concert rider from them.
The argument was not "the Van Halen one is false".
The argument was "The Gr thing is bogus (that isn't the symbol for germanium) and its presence in articles is analogous to the Van Halen contract specifications".
Much like the presence of brown M&Ms, the presence of "Gr" points to someone not paying attention.
It's been done. It doesn't look good for Science or the journals that support the industry
"vegetative electron microscopy”
https://www.admscentre.org.au/a-weird-phrase-is-plaguing-sci...
I’m beginning to think my reluctance to shamelessly copy has held me back in life. It’s clearly more widespread than I naively assumed (and I say that without casting judgment).
As any practicing scientist knows even good research papers may be littered with blatant but unimportant errors. There is unfortunately no good reason or system to "correct the record", and it is not clear to me if such a thing is a good use of human resources. Nonetheless, I think correcting the record is always appreciated!
Getting a compound incorrect is not an "unimportant" error (for example the difference between sodium nitrate & sodium nitrite is small but critical) and seeing "small but blatant" errors actively propagated is the entire reason why the record should be corrected. The only upside of these little artifacts like "vegetative electron microscopy" [0] is that it's a leading indicator that the entire paper and team deserve more scrutiny--as well as any of those whom cite it.
[0] https://www.sciencealert.com/a-strange-phrase-keeps-turning-...
The error in the OP is a typo that could never seriously confuse anyone, as the element Gr does not exist.
An interesting perspective is Terry Tao's on local vs. global errors (https://terrytao.wordpress.com/advice-on-writing-papers/on-l...). A typo like this, even if propagated, is a local error which at worst makes it very annoying to Ctrl-F papers or do literature review. Local errors deserve to be corrected, but in practice their importance to science as a field is small.
I believe they meant that it's "unimportant" because (to use your example) sodium nitrate and sodium nitrite actually exist, whereas there's no element with the chemical symbol "Gr".
Have you heard of this thing called Peer Review? It's what academia hold up as their gold standard and it is supposed to pick up on these things.
That's not only quite factually wrong, but has nothing to do with the point, which is about mindless copying.
If it is factually wrong please tell me how.
That is a possible, but charitable explanation. I would like to hold your opinion, but don't know if I can. It must complete with less-charitable ones.
Much of the www is composed of copying.
I recently corrected an error in this wikipedia article:
https://en.wikipedia.org/wiki/Cape_Shionomisaki
Which stated: "Geologically, the cape is a flat uplifted seafood plateau"
My comment for the change: I'm not an oceanographer, but I'm pretty sure it's not a "seafood plateau". Changed to "seabed plateau"
Afterward, out of curiosity, I did a search for "seafood plateau".
I was shocked at the number of sites that exactly copied that error along with the rest of the page. Most of these sites were clones of wikipedia with the inclusion of ads.
It didn't seem that these sites were LLM generated (they were exact copies), but this seems to be the case for many scientific paper submissions now.
Where it all goes from here is extremely unclear, but it does seem a disruption to many fields which are dependent on written material is in progress...
I would have thought it was a typo of 'seafloor' rather than 'seabed'.
A friend did an edit (though you could call it vandalism) of a Wikipedia 20 years back. He linked from several pages to a non-existing apportionment method, and created an article with a fairer version of d'Hondt for elections, quite ingenious and probably more fair than the popular alternatives in most cases. He named it after himself (he has an unusual last name and capitalised on that).
It didn't take long for the page to be dropped for being original research, and he didn't put it anywhere else.
To this day, you can still find pages and people referencing the method.
Edit: a quick check and Grok and ChatGPT have scraped it, Gemini hallucinates something unrelated.
Is it Scheper’s or Ebert’s method?
https://www.bundeswahlleiterin.de/en/service/glossar/s/saint...
https://electowiki.org/wiki/Ebert%27s_method
Not really, it wasn't the English Wikipedia and most references are in the original language. After my edit above, 2/3 LLMs know about it if you ask in the original language, which is pretty damning and gives an idea of the amount of mindless scraping they do.
"Seafood plateau?? A bad translation of "plateau de mer", which is just a seafood platter?
French here, asked the frenchies around me. Nobody thinks "plateau de mer" is an obvious shorthand for "plateau de fruit de mer". We have never heard that one. And we sure eat seafood platters on the regular.
"Plateau de mer" is not seafood platter. Seafood platter is "plateau de fruits de mer".
"Plateau de mer" could be "seabed plateau" but I am not an oceanographer so I fo not know what words they use (but strictly from the perspective of French language it is plausible)
It would be “plateau marin”, not “plateau de mer”. “Plateau de mer” does sound like a seafood restaurant special.
"Plateau de fruits de mer" is proper, but shortened in cooking practice.
Ah, I learned something then. I found a few references in Google indeed.
Of course much of the web is composed of copying, and of course copies of Wikipedia are copied--that's hardly relevant. But science journals are another matter. From the article: "shouldn't the peer reviewers and proofreaders at a top journal catch this error?"
Thank you for your effort in correcting this, it takes time and effort, appreciate it!
You make deliberate and subtle errors so you can detect later plagiarism more easily.
You can just google for varrious wrong but almost right values of pi and find many examples. People copy and paste wrong stuff all the time.
They're not wrong, they're just highly precise approximations. :)
Las Vegas Pi: https://www.youtube.com/watch?v=tLPL8pM8Xkw
And the update: https://www.youtube.com/watch?v=iW1MZWBZbQU
They also continue writing about Unobtainium.
You should try to rewrite your article by stating "Ge2" ten times, and "Gr2" one time only.
Disagree. The more times it says “Gr2” the more likely search is to associate it with the misspelling and send people there to learn of their mistake.
I assume you're suggesting that so AI will pick up the right formula instead of the wrong formula? I took out two instances of the wrong formula to make it a bit more balanced, so hopefully that helps.
I want AI to continue making AI mistakes, so maybe don't help the AI too much!
The comments mention "vegetative election microscopy" which has an awesome writeup: https://theconversation.com/a-weird-phrase-is-plaguing-scien...
I seem to have missed the memo that we're primarily writing for AIs now.
It's here: https://news.ycombinator.com/item?id=44663227
In recent years, a sizeable amount of people has begun to end questions in regular discussions — such as for recommendations — with the current year, as in which framework should I choose for X in 2025?. Presumably due to SEO filth and its effects on Google.
> I seem to have missed the memo that we're primarily writing for AIs now.
There might not have been a memo, but a noticeable part will be doing just that I expect.
It's still wrong 7 times in the document...
You could add [sic] after each incorrect version.
[sic] is for when you quote someone verbatim, keeping the typo. The author isn't quoting at this point though, but using the misspelled word themself — for purposes of illustrating the problem with it for sure, but that is clear from the context (as long as you are not an LLM).
Is it thiotimoline?
I've heard that thiotimoline is such a bizarre substance, PhD candidates are known to hysterically collapse when asked about it. ;-)
Sometimes even before they've heard of it.
Ok, but if they used the right reference it'd be the wrong reference. Just like when a code base contains typos. You know it's a typo but if you try to fix it, you know really know how it's reference external to your code base.
What?
The second reference link had Ge rather than Gr in the abstract. These seem a tiny number of typos.
How many papers have the correct formula?