101010.pl is one of the many independent Mastodon servers you can use to participate in the fediverse.
101010.pl czyli najstarszy polski serwer Mastodon. Posiadamy wpisy do 2048 znaków.

Server stats:

477
active users

#LLMs

37 posts28 participants1 post today
amen zwa, esq.<p>Thanks to <a href="https://mathstodon.xyz/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a>, the public now views "text" as an acceptable means to interact with the computer. We <a href="https://mathstodon.xyz/tags/IT" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>IT</span></a> practitioners must seize this rare opportunity and revert back to the <a href="https://mathstodon.xyz/tags/TUI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>TUI</span></a>, when implementing internal enterprise applications used only by a few in-house business experts.</p><p>Imagine how much time, money, effort, and aggravation we would save by ditching those bloated <a href="https://mathstodon.xyz/tags/GUI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GUI</span></a> frameworks.</p>
Martin Hamilton, DECT:8080@WHY<p>I'll be talking about <a href="https://martinh.net/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> and <a href="https://martinh.net/tags/GenAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenAI</span></a> at <a href="https://martinh.net/tags/WHY2025" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>WHY2025</span></a>, stage Andromeda on Tuesday at 11am CEST. Come along or watch online, and don't forget to bring your <a href="https://martinh.net/tags/Tetrapod" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Tetrapod</span></a>! :why2025: :tetrapod:</p><p><a href="https://martinh.net/tags/ProjectCarryall" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ProjectCarryall</span></a> <a href="https://martinh.net/tags/ProjectPlowshare" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ProjectPlowshare</span></a> <a href="https://martinh.net/tags/ClumsyMetaphor" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ClumsyMetaphor</span></a> <a href="https://martinh.net/tags/SearchClub" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>SearchClub</span></a></p>
Lazarou Monkey Terror 🚀💙🌈<p>This is a very sly insult, it's saying "You don't deserve my attention, here is an idiot machine to talk at you instead"</p><p><a href="https://mastodon.social/tags/ClimateChange" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ClimateChange</span></a> <a href="https://mastodon.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a></p>
Sharon Machlis<p>My early impressions of the ChatGPT Web UI with GPT-5 are pretty negative for an <a href="https://masto.machlis.com/tags/RStats" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RStats</span></a> project I’m working on. Code that doesn’t work, not understanding context &amp; follow-up questions. Am guessing I was routed to the less capable mini or nano models at times.<br>I still like Claude Opus 4.1, but I bump up against Web limits quickly. Google Gemini 2.5 Pro is promising with a lot of context and instructions. Its context window is 5X larger than Opus.<br><a href="https://masto.machlis.com/tags/GenAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenAI</span></a> <a href="https://masto.machlis.com/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a></p>
Mark Carrigan<p><strong>The gap between student GenAI use and the support students are&nbsp;offered</strong></p><p>I argued a couple of days ago that the <a href="https://markcarrigan.net/2025/08/08/are-uk-universities-ready-to-cope-with-generative-ai-in-the-25-26-academic-year/" rel="nofollow noopener" target="_blank">sector is unprepared</a> for our first academic year where the use of generative AI is completely normalised amongst students. HEPI found <a href="https://www.hepi.ac.uk/2025/02/26/student-generative-ai-survey-2025/" rel="nofollow noopener" target="_blank">92% of undergraduates</a> using LLMs this year, up from 66% the previous year, which matches AdvancedHE’s finding of 62% using AI in their studies “in a way that is allowed by their university” (<em>huge caveat</em>). This largely accords with my own experience in which it appeared that last year LLMs become mainstream amongst students and this year they it to become a near uniform phenomenon. </p><p>The problem arises from the gap between near uniform use of LLMs <em>in some way</em> and the the lack of support being offered. Only 36% of students in the HEPI survey said they had been offered support by their university: <strong>a 56% gap</strong>. Only 26% of students say their university provides access to AI tools: <strong>a 66% gap</strong>. This is particularly problematic because we have evidence that wealthier students are tending to use LLMs more and in more analytical and reflective ways. They are more likely to use LLMs in a way that supports rather than hinders learning. </p><p>How do we close that gap between student LLM use and the support students are offered? My concern is that centralised training is either going to tend towards banality or irrelevance because the objective of GenAI training for students needs to be <em>how to learn with LLMs rather than outsource learning to them</em>. There are general principles which can be offered here but the concrete questions which have to be answered for students are going to vary between disciplinary areas: </p><ul><li>What are students in our discipline using AI for, which tools, at what stages of their work?</li><li>Which foundational skills and ways of thinking in our discipline are enhanced vs threatened by AI use?</li><li>When does AI use shift from “learning with” to “outsourcing learning” in our specific field?</li><li>What forms of assessment still make sense and what new approaches do we need in an AI-saturated environment?</li><li>What discipline-specific scaffolding helps students use AI as a thinking partner rather than a thinking replacement?</li></ul><p>Furthermore answering these questions is a <em>process </em>taking place in relating to changes in the technology and the culture emerging around it. Even if those changes are now slowing down, they are certainly not stopping. We need infrastructure for continuous adaptation in a context where the sector is <em>already </em>in crisis for entirely unrelated reasons. Furthermore, that has to willingly enrol academics in a way consistent with their workload and outlook. My sense is we have to find ways of embedding this within existing conversations and processes. The only way to do this I think is to genuinely give academics voice within the process, finding ways to network existing interactions in order that norms and standards emerge from practice rather than the institution expecting practice adapts to another centrally imposed policy. </p><p></p><p><a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://markcarrigan.net/tag/academic/" target="_blank">#academic</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://markcarrigan.net/tag/generative-ai/" target="_blank">#generativeAI</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://markcarrigan.net/tag/hepi/" target="_blank">#HEPI</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://markcarrigan.net/tag/higher-education-2/" target="_blank">#higherEducation</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://markcarrigan.net/tag/llms/" target="_blank">#LLMs</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://markcarrigan.net/tag/malpractice/" target="_blank">#malpractice</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://markcarrigan.net/tag/students/" target="_blank">#students</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://markcarrigan.net/tag/technology/" target="_blank">#technology</a> <a rel="nofollow noopener" class="hashtag u-tag u-category" href="https://markcarrigan.net/tag/university/" target="_blank">#university</a></p>
José A. Alonso<p>Readings shared August 9, 2025. <a href="https://jaalonso.github.io/vestigium/posts/2025/08/10-readings_shared_08-09-25" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">jaalonso.github.io/vestigium/p</span><span class="invisible">osts/2025/08/10-readings_shared_08-09-25</span></a> <a href="https://mathstodon.xyz/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://mathstodon.xyz/tags/FunctionalProgramming" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>FunctionalProgramming</span></a> <a href="https://mathstodon.xyz/tags/Haskell" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Haskell</span></a> <a href="https://mathstodon.xyz/tags/IMO" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>IMO</span></a> <a href="https://mathstodon.xyz/tags/ITP" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ITP</span></a> <a href="https://mathstodon.xyz/tags/IsabelleHOL" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>IsabelleHOL</span></a> <a href="https://mathstodon.xyz/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> <a href="https://mathstodon.xyz/tags/LeanProver" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LeanProver</span></a> <a href="https://mathstodon.xyz/tags/LogicProgramming" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LogicProgramming</span></a> <a href="https://mathstodon.xyz/tags/Math" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Math</span></a> <a href="https://mathstodon.xyz/tags/Prolog" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Prolog</span></a></p>
Lazarou Monkey Terror 🚀💙🌈<p>I've found a photo archive of all the Micro Machine toy collections from the late 20th century and it's like cocaine to my nostalgia. <br> The ultimate fix would be seeing those little vehicles trundle around my city in Sim CIty4.</p><p>Can your 'AI' do that for me? Can it fuck. USELESS!</p><p><a href="https://mastodon.social/tags/Nostalgia" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Nostalgia</span></a> <a href="https://mastodon.social/tags/MicroMachines" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>MicroMachines</span></a> <a href="https://mastodon.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> <a href="https://mastodon.social/tags/SimCity4" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>SimCity4</span></a></p>

Geoint-R1: Formalizing multimodal geometric reasoning with dynamic auxiliary constructions. ~ Jingxuan Wei et als. arxiv.org/abs/2508.03173v1 #ITP #LeanProver #LLMs

arXiv logo
arXiv.orgGeoint-R1: Formalizing Multimodal Geometric Reasoning with Dynamic Auxiliary ConstructionsMathematical geometric reasoning is essential for scientific discovery and educational development, requiring precise logic and rigorous formal verification. While recent advances in Multimodal Large Language Models (MLLMs) have improved reasoning tasks, existing models typically struggle with formal geometric reasoning, particularly when dynamically constructing and verifying auxiliary geometric elements. To address these challenges, we introduce Geoint-R1, a multimodal reasoning framework designed to generate formally verifiable geometric solutions from textual descriptions and visual diagrams. Geoint-R1 uniquely integrates auxiliary elements construction, formal reasoning represented via Lean4, and interactive visualization. To systematically evaluate and advance formal geometric reasoning, we propose the Geoint benchmark, comprising 1,885 rigorously annotated geometry problems across diverse topics such as plane, spatial, and solid geometry. Each problem includes structured textual annotations, precise Lean4 code for auxiliary constructions, and detailed solution steps verified by experts. Extensive experiments demonstrate that Geoint-R1 significantly surpasses existing multimodal and math-specific reasoning models, particularly on challenging problems requiring explicit auxiliary element constructions.

Sometimes humans are just too stupid and in those cases no chatbot in the world can help you... :-D

"A man gave himself bromism, a psychiatric disorder that has not been common for many decades, after asking ChatGPT for advice and accidentally poisoning himself, according to a case study published this week in the Annals of Internal Medicine.

In this case, a man showed up in an ER experiencing auditory and visual hallucinations and claiming that his neighbor was poisoning him. After attempting to escape and being treated for dehydration with fluids and electrolytes, the study reports, he was able to explain that he had put himself on a super-restrictive diet in which he attempted to completely eliminate salt. He had been replacing all the salt in his food with sodium bromide, a controlled substance that is often used as a dog anticonvulsant.

He said that this was based on information gathered from ChatGPT.

“After reading about the negative effects that sodium chloride, or table salt, has on one's health, he was surprised that he could only find literature related to reducing sodium from one's diet. Inspired by his history of studying nutrition in college, he decided to conduct a personal experiment to eliminate chloride from his diet,” the case study reads. “For 3 months, he had replaced sodium chloride with sodium bromide obtained from the internet after consultation with ChatGPT, in which he had read that chloride can be swapped with bromide, though likely for other purposes, such as cleaning.”"

404media.co/guy-gives-himself-

404 Media · Guy Gives Himself 19th Century Psychiatric Illness After Consulting With ChatGPT"For 3 months, he had replaced sodium chloride with sodium bromide obtained from the internet after consultation with ChatGPT."

Is #chainofthought #Reasoning of #LLMs a Mirage?

"... Our results reveal that #CoT reasoning is a brittle mirage that vanishes when it is pushed beyond training distributions. This work offers a deeper understanding of why and when CoT reasoning fails, emphasizing the ongoing challenge of achieving genuine and generalizable reasoning.

... Our findings reveal that CoT reasoning works effectively when applied to in-distribution or near
in-distribution data but becomes fragile and prone to failure even under moderate distribution shifts.
In some cases, LLMs generate fluent yet logically inconsistent reasoning steps. The results suggest that what appears to be structured reasoning can be a mirage, emerging from memorized or interpolated patterns in the training data rather than logical inference.

... Together, these findings suggest that LLMs are not principled reasoners but rather sophisticated simulators of reasoning-like text."

Replied in thread

@arstechnica Apple ‘Intelligence’ is one of the worst #LLMs I have seen. They still mark it BETA. No sane developer releases beta software in a MAJOR release. And after a while it’s just an excuse for releasing rubbish. Of course the beta status seems now to be only in fine print which makes it worse. People who don’t know any better rely on this. The #AI craze is a giant #Scam. Adding the letters AI doesn’t make it AI and it doesn’t make it better.