Jonathan Newell

Weird fiction, fantasy cartography, monsters, roleplaying games. Credits include Genial Jack, work on Dolmenwood. Blog:

[Jonathan Newell @Edweirdian]

I have big news. I've launched a patreon for the Hex setting, including monthly maps, fully-illustrated installments of the Hex Guidebook detailing every single street, a playtest version of The Hex Hack system, and an archive of older high-resolution maps. Subscribe below.⬇️

[Jonathan Newell @Edweirdian]

@stanislavfort @akoustov I used credentials as part of an analogy. The point was not to demand that AI literally had credentials. It was to suggest that AI often writes like someone without them, hence the current profusion of error-filled, semi-fabricated slop flowing through academic journals.

[Jonathan Newell @Edweirdian]

@stanislavfort @akoustov In real life, an RA's credentials would signal they understand things like basic research ethics, the importance of academic integrity, not falsifying or fabricating sources, not plagiarizing. AI lacks this understanding and has regularly failed in this regard.

[Jonathan Newell @Edweirdian]

@MushtaqBilalPhD It's almost like these forms of knowledge are different!

[Jonathan Newell @Edweirdian]

@_Coati_ @akoustov That remains to be seen for me. I remain very, very resistant to its use outside of narrow circumstances in the humanities, where the entire point is to continue a rich human conversation. To some extent, naturally, the different disciplines will come to different conclusions.

[Jonathan Newell @Edweirdian]

@DavidHeadPhd @akoustov I think a lot of STEM people don't care about voice, anyway. Which, like, fair I guess?

[Jonathan Newell @Edweirdian]

@_Coati_ @akoustov If an RA turned out to be the primary cause of a series of serious academic scandals and contributed to colleagues being banned from publication and professionally embarrassed, would you really invite them to work on your paper?

[Jonathan Newell @Edweirdian]

@_Coati_ @akoustov Okay but let's not kid ourselves here, LLMs have not made "an error." They've made thousands of errors, measurably, again and again, in multiple journals and fields. If an RA makes an honest mistake, OK. This RA turns out to be a serial fabricator and plagiarizer.

[Jonathan Newell @Edweirdian]

@akoustov If these systems were perfect then a year-long ban from somewhere like arXiv would be fine and AI-boosters ought to have no complaints for even severe consequences.

[Jonathan Newell @Edweirdian]

@akoustov My understanding of how LLMs work is that fundamentally there is no substantive difference between a hallucination and a regular response; it's just the mechanism the AI uses. But the entire controversy stems from fabricated citations.

[Jonathan Newell @Edweirdian]

@akoustov Like this hypothetical RA, to follow the analogy, is a known, repeat offender in the field responsible for the professional embarrassment of numerous scientists and seemingly incapable of upholding basic standards of academic rigour. "But he can write fast" seems a little thin.

[Jonathan Newell @Edweirdian]

@akoustov I'm in the humanities but I've tried to think about the RA frame from a STEM perspective. I have to say I also find it baffling. People keep including a co-author who regularly hallucinates, has no formal academic credentials, and has been found to repeatedly fabricate sources?

[Jonathan Newell @Edweirdian]

@xriskology The condescension is really something here.

[Jonathan Newell @Edweirdian]

@MushtaqBilalPhD Yes and this is what the present technology is doing to the academy. I'm seeing the damage firsthand to student learning. "It's 2026" is not an argument.

[Jonathan Newell @Edweirdian]

@pensandpoison You must be aware that the guy who wrote 126 sonnets about a "fair youth," whose comedies are full of cross-dressing, and whose tragic characters beseech spirits to "unsex" them might be amenable to a queer theoretical approach.

[Jonathan Newell @Edweirdian]

@pensandpoison It's Open Access, you can read it: https://t.co/Q7JQhxFinR It's pretty great at a glance, seems carefully historicized and theorized. I'm not an early modernist but this seems like interesting scholarship. Genuinely have no idea what you're even trying to communicate.

[Dr. Mia Brett @QueenMab87]

Academics & writers justifying the use of AI is like an artist saying their paint by numbers picture belongs in a museum

[Jonathan Newell @Edweirdian]

His response when called out at one point was the classic "but what if it cures cancer." Sir, this is a paper on folklore.

[Mushtaq Bilal, PhD @MushtaqBilalPhD]

🚨Claude Code just gave me a complete research paper with a single prompt.

The paper has a strong argument and even beats AI-detection app, Pangram.

With a little editing, it can pass for 100% human, and can be easily submitted for peer review.

Here's the workflow I used: