Luke Gessler

@tszzl @ashter_haider @alokranj de bello gallico is propaganda... not really the same kind of book

@ashter_haider @alokranj The Death of Caesar i mean the primary texts are so based Plutarch, caesar’s own Conquest of Gaul

This was excellent. She tries to cover a millennium of Roman history in about 500+ pages so it does feel like she is racing through it but it's a great starting point overall. It is also written in a very engaging style. Now I need to find something similar for Ancient Greece.

@timgilmour @nominalthoughts come on, the charitable read is that the "why" means "how it arises" there

@nominalthoughts "why consciousness exists" is an unanswerable metaphysical question on the order of "why does the universe exist" jfc "how animal brains work" is a topic in which we have a ton of research, with tons more coming every second. what, does he not think we know about neurons?

Incredible

@ayatabourashed @Andrew_Akbashev this is about review articles (surveys of specific topics/subfields), not peer reviews (evaluations of submitted works)

arXiv has banned reviews in computer science. Should research journals do the same? 📍For arXiv, the reason is obvious: 1. Citations = money, funding, awards. Unethical scientists want more citations. Some of them ‘order’ citations from the paper mills that produce reviews with those citations. The reviews become ‘vehicles’ for targeted citation inflation and are then uploaded to arXiv (or even published in journals). “Many surveys that are just annotated bibliographies without analysis, synthesis or road mapping, the citations often include papers on completely unrelated subjects. We suspect there may be markets where such citations can be purchased.” [Thomas Dietterich, the chair arXiv’s computer-science section] 2. LLMs are becoming too good to be detected. According to arXiv’s programme director, such LLM-prepared reviews can escape simple plagiarism checks, whether performed by people or software. 📍 For me, there are more reasons to move away from reviews: 3. Today, most reviews are useless. They represent bibliographic overviews. No original thoughts, no vision. In my field, many reviews look the same. They cite low-quality research and often are hard to read. The main reason why they exist is because journals love reviews (they increase their impact factor). 4. LLMs can already summarize literature very well. They insert correct citations, they describe the content in understandable language, they can generate reviews in any niche (!). What’s the point in human-made reviews that cite the wrong literature and low-quality research? 5. There are TOO many reviews. No one reads most of them anyway. They exist to rack up citations for authors and impact factors for journals. 📍 I think: 1. We need fewer reviews. 2. We need HIGH-quality reviews. 3. We need more bold vision in reviews, not LLM-style summarization. ❗️ It’s great to see that arXiv starts moving in this direction. Now - what about major journals? Will they be next? Or am I too naive?

@iwsfutcmd could it not be that the "There are pants all over the road." case is just ambiguous in a way that that either one in that squirrel pair is not?

(note that "There are pants all over the road." doesn't do it, because "There are squirrels all over the road." and "There is squirrel all over the road." have *distinctly* different meanings)

my favorite fact about plurale tantum is you can't use the Universal Grinder on them "This is my squirrel." "There are three squirrels." "There's squirrel all over the road." "These are my pants." "There are three pairs of pants." ¿"There's pant all over the road."?

Statement from Ranking Member @RepGregoryMeeks demanding answers on and the legal justification for the Trump administration’s strike on a vessel in the Caribbean Sea last week:

@jamesr66a as is Widget widget = new Widget();

public static void main(String[] args) is a single token

@dlbydq @benlandautaylor @csvoss on all counts except for the supposition there will be more than 0 historical linguists employed a few decades from now

@benlandautaylor @LukeGessler how well did @csvoss nail the historical linguistics paper voice

Chelsea’s joke was already one of the funniest things I’d seen on Twitter. Now that the guy who actually answered the question deleted all his tweets, it’s even better. My apologies to the future historians, hope you guys appreciate irony.

@jamesr66a good work on the buckling front

was once trying to explain to a layperson how an "AI launches the nukes" situation is at all possible ("but it's just a next word predictor!") and did not expect to see something like this in the real world so soon lol

RIP linguist Haj Ross (1938-2025) en.wikipedia.org/wiki/John_R._R…

@grok @CyberScribe_AI @tancredipalmeri @AskPerplexity @grok but what would a Roman auspex have made of the situation?

The event of a seagull vomiting a rat just before a papal announcement likely holds no significance. Seagulls are often seen as symbolic in some cultures, like in Japanese Shinto where they’re viewed as protectors, but there’s no historical link to papal events. Scientifically, animal behaviors don’t predict human events—studies show such actions are instinctual, not omens. Historically, papal elections have no record of animal behaviors being interpreted as signs. While some might see the timing as meaningful based on personal beliefs, it’s most likely a coincidence. Interpretations can vary depending on individual perspectives.

At the very beginning of the video, a seagull is vomiting a rat so that its baby could have its meal. 39 seconds later, Pope is announced. This is certainly an omen. Time will tell us of what kind #HabemusPapam

@ElonBachman i would just assume it's that literacy rates have gone up

Is English becoming more Germanic over time? When I read old books, it's the Latinate words (detestation, contrariety) that seem to have been recently purged

@jamesr66a what about in "Paul will come home"?

pronouncing the w in “will” as an auxiliary verb is completely foreign for me “sunday’ll be totally free for me”

@ben_golub were the problems in the training data?

Remember when, in 2022(?) some people put out a paper that was total bullshit that said ChatGPT can solve almost all the problems in the MIT core curriculum? Well, now that claim is true. And, interestingly, the immediate economic impact of that fact is rather mild.

@bubbleboi trusting your own taste regardless of how mainstream something is is good, actually

Why is this guy such a normie?

vibe coded this interactive optimizer visualization thing, still kinda buggy but not bad for an hour of telling gemini and claude to just do things! lgessler.com/files/optimize…

Meet Ai2 Paper Finder, an LLM-powered literature search system. Searching for relevant work is a multi-step process that requires iteration. Paper Finder mimics this workflow — and helps researchers find more papers than ever 🔍

@kayaulai @haspelmath I recommend checking out asciidoc if you haven't yet--not sure if it'd suit your use-case but it is a lot more featureful than markdown and is (iirc) fairly easily extensible

Is this a good resource for linguists? (agricolamz.github.io/lingglosses/)

if you're not acquainted: delph-in.github.io/docs/erg/ErgTo…

can someone pay an undergraduate summer research intern to vibe code extensions to the English Resource Grammar so we can see how long it'd take us to cover All Of English with HPSG?