#arxiv — Public Fediverse posts

Today on the #arXiv:

Batygin et al. 2026, "Terminal instability of the Solar System triggered by stochastic solar mass loss" - https://arxiv.org/abs/2609.12494.

We know that the inner planets will eventually fall into the red giant Sun, unless they scatter first. The gas giants likely scatter afterwards.

The Malicious Use of Artificial Intelligence

https://arxiv.org/abs/1802.07228

#arxiv

Will there be a 7G?

https://arxiv.org/abs/2609.01877

#arxiv

Will there be a 7G?

https://arxiv.org/abs/2609.01877

#arxiv

Will there be a 7G?

https://arxiv.org/abs/2609.01877

#arxiv

Will there be a 7G?

https://arxiv.org/abs/2609.01877

#arxiv

Will there be a 7G?

https://arxiv.org/abs/2609.01877

#arxiv

RAG, поиск и LongContext: почему сложные пайплайны не всегда нужны

В 2023 году RAG был единственным способом засунуть знания в LLM — контекстное окно было маленьким, часто были галлюцинации. RAG постепенно стал одной из самых популярных технологий, чтобы получить базу знаний, по которой можно искать данные через натуральный язык. Но с другой стороны — RAG-пайплайн тяжёлый и трудозатратный. Компания хочет чат по своей документации. Разработчик говорит: нужен пайплайн — чанкинг, эмбеддинги, векторная база, реранкер. Два-три месяца работы плюс сервис, который надо вечно поддерживать, ради 400-страничной документации. И всё это занимает месяцы, тратит ресурсы ради не такой уж и большой выгоды. Сегодня многие модели держат 1M токенов, а то и больше, а в это окно контекста чаще всего спокойно влезает вся документация. Но если не влезет — то обязательно ли сразу строить весь RAG самому? А как понять, когда есть альтернатива RAG, а когда нет? В этой статье мы разберём, почему RAG стал выбором по умолчанию (и почему это было оправдано), посмотрим, как падает качество при Long Context, сравним подходы и узнаем, что, как и когда использовать.

https://habr.com/ru/companies/bothub/articles/1081080/

#документация #техническая_документация #RAG #LLM #longcontext #AI #ИИ #arxiv #эмбеддинги #поиск

Rxiv mechanobio
@Rxiv_mechanobio • Joined: Nov 05, 2024

I toot papers from #PubMed and preprints from #bioRxiv & #arXiv that seem to pertain to mechanobiology.
My filter can be tested there on PubMed: https://link.infini.fr/mechanobiology_2022_11_07
Improvement suggestions in this filter welcome, both to add or remove spurious results: contact @jocelyn_etienne

#bot #BotsOfMastodon

How many times I'll need to explain to our dear #Reviewer2 that submitting my own #arxiv paper to a conference is not #selfplagiarism 🤔🤷

There is a lot of discussion on the use of #AI for #reviewing but, for sure, I'd use #AI to automatically #flag bad reviews (according to a number of criteria, including the above but also based on the lenght of the review, use of toxic words, or even asking for validation on a new ideas paper 😁, another one of my pet peeves)

Looks like Flatpak discovery will move from one mostly generally accepted hub to a gazillion of repos since the performative LLM complaints have befallen the possible single hub. (Partially half-understandable arguments, partially performative fluff for my taste)

Which is ok as it triggers competition that might lead to better outcomes. (bazaar vs cathedral and all, rite? 😏 )
And removes a possible single point of failure.

Now what I'd like to see is an approach that does discovery of those repos and enables search across those repos without a single point of failure. As e.g. a single, centrally hosted index for repos to register to would be.

Flatpak already uses a reverse-dns-namespace structure internally. Anything that would leverage that? Wouldn't help with enumeration though.

Something something DHT/Gossip/...? (as the cool thing from the past)

Clearly a problem that will arise with more of the expectable decentralized systems that need uniform discovery and search.

What's the Fedis wisdom on current academic approaches that might be interesting around this?
Or bastardized routing tech, yet another idea on how to twist DNS into a pretzel, etc? Dare I say: Blockchain (as the unpopular thing before the LLMs 🤣 )

Throw you best Arxiv posts etc. at me. I'm just curious about the state-of-the-art techn for now, all smirk and sarcasm aside.

#discovery #search #decentralized #routing #dns #fedimeta #flathub #dht #cooltechfromthepast #llm #selfhosting #fossdrama #spof #singlepointoffailure #arxiv #academicchatter #academic #computerscience

The Burau representation of the braid group is faithful for n = 4

https://arxiv.org/abs/2607.05283

Comments: https://news.ycombinator.com/item?id=49077209

#HackerNews #BurauRepresentation #braidGroup #n4 #mathematics #research #arXiv

The Burau representation of the braid group is faithful for n = 4

https://arxiv.org/abs/2607.05283

Comments: https://news.ycombinator.com/item?id=49077209

#HackerNews #BurauRepresentation #braidGroup #n4 #mathematics #research #arXiv

The Burau representation of the braid group is faithful for n = 4

https://arxiv.org/abs/2607.05283

Comments: https://news.ycombinator.com/item?id=49077209

#HackerNews #BurauRepresentation #braidGroup #n4 #mathematics #research #arXiv

The Burau representation of the braid group is faithful for n = 4

https://arxiv.org/abs/2607.05283

Comments: https://news.ycombinator.com/item?id=49077209

#HackerNews #BurauRepresentation #braidGroup #n4 #mathematics #research #arXiv

The Burau representation of the braid group is faithful for n = 4

https://arxiv.org/abs/2607.05283

Comments: https://news.ycombinator.com/item?id=49077209

#HackerNews #BurauRepresentation #braidGroup #n4 #mathematics #research #arXiv

🤔📉 Ah, the riveting thrill of optimizing 32-bit division on a 64-bit machine—because who wouldn't want to spend their Saturday night wrestling with... constants? Meanwhile, #arXiv slyly slips in a request for #donations, as if any of us have been moved to tears by this mathematical melodrama. 😂💸
https://arxiv.org/abs/2604.07902 #optimizingdivision #64bitmachines #mathhumor #techthrills #codingstruggles #HackerNews #ngated

@freyablekman @Thcoudreau @johncarlosbaez @pluralistic It's surely the case that scholars need to be vocal.

However, it's also related to what the kids call #terminal #capitalism; There's only so many hills you can die on *within* faculty.

If the society at large decides* there is no money for #arXiv, then arXiv will go the way of all #enshittified platforms.

* And it should be obvious that the high-level decision-making process should be questioned, not [only] scholars fighting for #budget.

Today on the #arXiv :

Deen & Lam 2026, "2024 YR4: Identification of Possible Precoveries in 2016 IPTF Data" - https://arxiv.org/abs/2603.00449

This paper presents an archival search of old telescope images of #2024YR4 . It just happened to fall in gaps of several surveys.

But Deen & Lam identify one marginal potential detection from Palomar in 2016. If real, it would exclude the potential lunar impact in 2032.

I await the results of the latest JWST observations, which should be far clearer.

Today on the #arXiv :

Micheli et al. 2026, “Astrometric follow-up of near-Earth asteroid 2024 YR4 during a Torino scale level 3 alert” - https://arxiv.org/abs/2602.06890

This paper goes through all of the observations of #2024YR4 after its discovery at the end of 2024 until it became too faint to observe some months into 2025.

The next observations of the asteroid will be happening with JWST over the next couple of weeks.

"Until now, someone wanting to submit to arXiv for the first time only needed an email address affiliated with a reputable academic or research institution, such as a university. But a rule instituted on 21 January now requires first-time posters to be endorsed by an established arXiv author in their own field. People who have previously posted in the same disciplinary section of arXiv do not need an endorsement.
The move is an attempt to clamp down on a rising tide of fraudulent submissions, says University of Amsterdam astronomer Ralph Wijers, chair of the arXiv editorial council. A large fraction, he says, are generated with artificial intelligence (AI). The new rule is “mostly to try and discourage very junior, unskilled people from trying to get something started by sending some rubbish to arXiv,” he says."

https://www.science.org/content/article/arxiv-preprint-server-clamps-down-ai-slop

#arXiv #AIslop #PublicationEthics #PrePrints #AIinAcademia

Today on the #arXiv:

Devogèle et al. 2025, "Rapid-response characterization of near-Earth asteroid 2024 YR4 during a Torino Scale 3 alert" - https://arxiv.org/abs/2511.09405

Reviewing the case of #2024YR4 as we approach the first anniversary of its discovery.

And, to emphasize, it will miss the Earth.

Grammars of Formal Uncertainty

https://arxiv.org/abs/2505.20047

#HackerNews #Grammars #of #Formal #Uncertainty #arxiv #FormalLanguages #UncertaintyTheory #HackerNews

Oh, the irony! 🤖 An article that promises to demystify #Transformers with N-grams but really just masquerades as a job listing for a #DevOps Engineer at #arXiv. 📜 Because nothing says "deep understanding" like *skipping to the main content* for career opportunities. 🌟
https://arxiv.org/abs/2407.12034 #Ngrams #irony #careeropportunities #HackerNews #ngated

Someone's got a crystal ball, and it predicts... wait for it... images from text! 🖼️💬 Shocker! Meanwhile, #arXiv is #hiring, but who needs a job when you can just #dream about #AI wonders? 😴🔮
https://arxiv.org/abs/2504.06256 #Wonders #Image #Generation #Tech #News #Jobs #HackerNews #ngated

A new era: Computer science overtakes math+physics on arXiv.
#arXiv #Math #Physics #MathPhysics #CS #ComputerScience #NewEra