toad.social

Dominik Steiger"complete accuracy collapse beyond certain complexities"Interesting paper about the current limits of AI.<a href="https://swiss.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://swiss.social/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://swiss.social/tags/LRM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LRM</a> <a href="https://machinelearning.apple.com/research/illusion-of-thinking?utm_source=perplexity" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://machinelearning.apple.com/research/illusion-of-thinking?utm_source=perplexity</a>

Ω 🌍 Gus PoseyWatching people get comfortable with generative AI as a replacement for creative expression is like reaching into a sock drawer and finding something warm and wet.<a href="https://mas.to/tags/BingImageCreator" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#BingImageCreator</a> <a href="https://mas.to/tags/ChatGPT" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ChatGPT</a> <a href="https://mas.to/tags/DallE" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#DallE</a> <a href="https://mas.to/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://mas.to/tags/LRM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LRM</a> <a href="https://mas.to/tags/Midjourney" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Midjourney</a> <a href="https://mas.to/tags/StableDiffusion" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#StableDiffusion</a>

Deborah Hartmann Preuss, pccWe've forgotten what intelligence is … <a href="https://www.linkedin.com/posts/stuart-winter-tear_in-the-1960s-we-mistook-imitation-for-understanding-activity-7336987685774876673-EAdG" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://www.linkedin.com/posts/stuart-winter-tear_in-the-1960s-we-mistook-imitation-for-understanding-activity-7336987685774876673-EAdG</a> <a href="https://cosocial.ca/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://cosocial.ca/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://cosocial.ca/tags/dystopia" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#dystopia</a>

Matt FranzWow, bloated <a href="https://infosec.exchange/tags/llm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#llm</a> <a href="https://infosec.exchange/tags/agent" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#agent</a>

Jack Yan (甄爵恩)In my blog The Barnum effect of chat bots<a href="https://jackyan.com/blog/2025/06/the-barnum-effect-of-chat-bots/" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://jackyan.com/blog/2025/06/the-barnum-effect-of-chat-bots/</a> <a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://mastodon.social/tags/chatbot" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#chatbot</a> <a href="https://mastodon.social/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> Hat tip to <a href="https://infosec.exchange/@briankrebs" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@briankrebs</a> for his post:<a href="https://infosec.exchange/@briankrebs/114636857985994147" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://infosec.exchange/@briankrebs/114636857985994147</a>

JeromeAI non-use in UX is "not about rejecting technology, but about negotiating its limits, values, and contexts." —I. Cha, R. Wong, 2025<a href="https://ethicsinfrastructures.lmc.gatech.edu/2025/05/26/paper-summary-understanding-socio-technical-factors-configuring-ai-non-use-in-ux-work-practices/" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://ethicsinfrastructures.lmc.gatech.edu/2025/05/26/paper-summary-understanding-socio-technical-factors-configuring-ai-non-use-in-ux-work-practices/</a><a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://mastodon.social/tags/ArtificialIntelligence" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ArtificialIntelligence</a> <a href="https://mastodon.social/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://mastodon.social/tags/UX" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#UX</a> <a href="https://mastodon.social/tags/UserExperience" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#UserExperience</a> <a href="https://mastodon.social/tags/BestPractices" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#BestPractices</a> <a href="https://mastodon.social/tags/research" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#research</a>

Jan Wildeboer 😷:krulorange:On the limits of LLMs (Large Language models) and LRMs (Large Reasoning Models). The TL;DR: "Our findings reveal fundamental limitations in current models: despite sophisticated self-reflection mechanisms, these models fail to develop generalizable reasoning capabilities beyond certain complexity thresholds." Meaning: accuracy collapse.Interesting paper from Apple. <a href="https://ml-site.cdn-apple.com/papers/the-illusion-of-thinking.pdf" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://ml-site.cdn-apple.com/papers/the-illusion-of-thinking.pdf</a><a href="https://social.wildeboer.net/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://social.wildeboer.net/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://social.wildeboer.net/tags/LRM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LRM</a>

ResearchBuzz: FirehoseTechCrunch: EleutherAI releases massive AI training dataset of licensed and open domain text. “The dataset, called the Common Pile v0.1, took around two years to complete in collaboration with AI startups Poolside, Hugging Face, and others, along with several academic institutions. Weighing in at 8 terabytes in size, the Common Pile v0.1 was used to train two new AI models from EleutherAI, […]<a href="https://rbfirehose.com/2025/06/07/techcrunch-eleutherai-releases-massive-ai-training-dataset-of-licensed-and-open-domain-text/" class="" rel="nofollow noopener noreferrer" target="_blank">https://rbfirehose.com/2025/06/07/techcrunch-eleutherai-releases-massive-ai-training-dataset-of-licensed-and-open-domain-text/</a>

Bob ThomsonThis week in - “Please stop fannying about with LLMs and just make basic features for users that actually bloody work”FB Messenger - Good news! You can now pin chats.(Pins chats).Sorry you can only have a maximum of 3 (!) pinned chats.lol WTF 🤡<a href="https://mastodon.social/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://mastodon.social/tags/AIslop" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AIslop</a> <a href="https://mastodon.social/tags/socialmedia" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#socialmedia</a> <a href="https://mastodon.social/tags/meta" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#meta</a> <a href="https://mastodon.social/tags/facebook" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#facebook</a>

Erik JonkerInteresting, "GPT-style models have a fixed memorization capacity of approximately 3.6 bits per parameter." <a href="https://venturebeat.com/ai/how-much-information-do-llms-really-memorize-now-we-know-thanks-to-meta-google-nvidia-and-cornell/" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://venturebeat.com/ai/how-much-information-do-llms-really-memorize-now-we-know-thanks-to-meta-google-nvidia-and-cornell/</a> <a href="https://mastodon.social/tags/ai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ai</a> <a href="https://mastodon.social/tags/memorization" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#memorization</a> <a href="https://mastodon.social/tags/llm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#llm</a>

HistoPol (#HP) 🏴 🇺🇸 🏴<a href="https://swiss.social/@evelynefoerster" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@evelynefoerster</a> <a href="https://mastodon.social/tags/K%C3%BCnstlicheIntelligenz" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#KünstlicheIntelligenz</a> und die Zukunft der Menschheit(5/n)...nicht von einer oder mehrer <a href="https://mastodon.social/tags/AGI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AGI</a>|s dominieren lassen wollen, selbst wenn der Ressourcenkonflikt durch die Verlagerung der Rechenzentren ins Orbit 1) teilweise oder größtenteils lösen lässt.Die <a href="https://mastodon.social/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a>|s wurden nach unserem Abbild,...1) <a href="https://www.scinexx.de/news/technik/china-baut-ersten-ki-supercomputer-im-all/" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://www.scinexx.de/news/technik/china-baut-ersten-ki-supercomputer-im-all/</a><a href="https://www.economist.com/science-and-technology/2025/04/09/could-data-centres-ever-be-built-in-orbit" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://www.economist.com/science-and-technology/2025/04/09/could-data-centres-ever-be-built-in-orbit</a>

Jörg Seidel<a href="https://mastodon.social/@w4w3r" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@w4w3r</a> Nein! Doch! Oh!<a href="https://det.social/tags/KI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#KI</a> <a href="https://det.social/tags/Demokratie" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Demokratie</a> <a href="https://det.social/tags/PolitikUndDigitalisierung" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#PolitikUndDigitalisierung</a> <a href="https://det.social/tags/Civolve" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Civolve</a> <a href="https://det.social/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://det.social/tags/Reasoning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Reasoning</a> <a href="https://det.social/tags/Simulation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Simulation</a> <a href="https://det.social/tags/CivicTech" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#CivicTech</a> <a href="https://det.social/tags/MachineLearning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#MachineLearning</a> <a href="https://det.social/tags/StrategischeKommunikation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#StrategischeKommunikation</a>

David Palk#RejoinEU 💖🇪🇺🤎🕊<a href="https://plush.city/@scarlet" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@scarlet</a> Re: <a href="https://mas.to/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> & <a href="https://mas.to/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> failure is inevitable i any process that doesn't allow for randomness will fall over. By it's nature randomness cannot be modelled so cannot be factored into the process. ii. all process will have it's own internal feedback giving rise to the 'law of unintended consequence' Always best to have a 'clever' Hom. Sap. on standby methinks.

IanMoore3000It is amazing that there are still grown adults out there who think generative AI is some kind of oracle and not a plausible word combiner."High court tells UK lawyers to stop misuse of AI after fake case-law citations" <a href="https://www.theguardian.com/technology/2025/jun/06/high-court-tells-uk-lawyers-to-urgently-stop-misuse-of-ai-in-legal-work" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://www.theguardian.com/technology/2025/jun/06/high-court-tells-uk-lawyers-to-urgently-stop-misuse-of-ai-in-legal-work</a><a href="https://mastodon.ie/tags/GenAI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#GenAI</a> <a href="https://mastodon.ie/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://mastodon.ie/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#GenerativeAI</a>

woozong 🅅yo fedi, how are you today?I'm trying to find the absolute best, well-argumented (!) <a href="https://post.lurk.org/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> takedown blogpost you've read lately.. preferably related to software development.Unfortunately my boss is on the hype-train (random glory posts in slack) and I'd like to counter him just a wee bit. Bring it on 😁Boosting appreciated for reach :boost_anim_vanilla:

ResearchBuzz: FirehoseProPublica, and once again I want to assure you I did not deliberately put these two stories together: Inside the AI Prompts DOGE Used to “Munch” Contracts Related to Veterans’ Health. “When an AI script written by a Department of Government Efficiency employee came across a contract for internet service, it flagged it as cancelable. Not because it was waste, fraud or abuse — the […]<a href="https://rbfirehose.com/2025/06/06/propublica-inside-the-ai-prompts-doge-used-to-munch-contracts-related-to-veterans-health/" class="" rel="nofollow noopener noreferrer" target="_blank">https://rbfirehose.com/2025/06/06/propublica-inside-the-ai-prompts-doge-used-to-munch-contracts-related-to-veterans-health/</a>

Charlie McHenryMozilla says: <a href="https://connectop.us/tags/Meta" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Meta</a> is blurring the line between private and public — and it’s happening at the cost of our privacy. People have the right to know when they’re speaking in public, especially when they believe they’re speaking in private. If you agree, add your name to demand Meta shut down its invasive <a href="https://connectop.us/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> feed — and guarantee that no private conversations are made public without clear, explicit, and informed opt-in consent. <a href="https://connectop.us/tags/privacy" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#privacy</a> <a href="https://connectop.us/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://connectop.us/tags/ArtificialIntelligence" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ArtificialIntelligence</a> <a href="https://connectop.us/tags/petition" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#petition</a> <a href="https://www.mozillafoundation.org/en/campaigns/meta-shut-down-your-invasive-ai-discover-feed-now/" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://www.mozillafoundation.org/en/campaigns/meta-shut-down-your-invasive-ai-discover-feed-now/</a>

shentonfreude<a href="https://thepit.social/@peter" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@peter</a> Funny, these two diametrically opposed posts from folks I respect came up in my feed today, next to each other.Amin <a href="https://hachyderm.io/@mitsuhiko" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@mitsuhiko</a> "AI changes everything": <a href="https://lucumr.pocoo.org/2025/6/4/changes/" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://lucumr.pocoo.org/2025/6/4/changes/</a>Glyph <a href="https://mastodon.social/@glyph" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@glyph</a> "Done thinking about genAI" <a href="https://blog.glyph.im/2025/06/i-think-im-done-thinking-about-genai-for-now.html" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://blog.glyph.im/2025/06/i-think-im-done-thinking-about-genai-for-now.html</a>Both thoughtful pieces, worth a read. <a href="https://mastodon.online/tags/ai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ai</a> <a href="https://mastodon.online/tags/llm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#llm</a>

Fedify: an ActivityPub server frameworkDid you know? <a class="mention hashtag" rel="nofollow noopener noreferrer" href="https://hollo.social/tags/Fedify" target="_blank">#Fedify</a> provides <a class="mention hashtag" rel="nofollow noopener noreferrer" href="https://hollo.social/tags/documentation" target="_blank">#documentation</a> optimized for LLMs through the <a href="https://llmstxt.org/" rel="nofollow noopener noreferrer" target="_blank">llms.txt standard</a>.Available endpoints:<ul> <li><a href="https://fedify.dev/llms.txt" rel="nofollow noopener noreferrer" target="_blank">https://fedify.dev/llms.txt</a> — Core documentation overview</li><li><a href="https://fedify.dev/llms-full.txt" rel="nofollow noopener noreferrer" target="_blank">https://fedify.dev/llms-full.txt</a> — Complete documentation dump</li> </ul>Useful for training <a class="mention hashtag" rel="nofollow noopener noreferrer" href="https://hollo.social/tags/AI" target="_blank">#AI</a> assistants on <a class="mention hashtag" rel="nofollow noopener noreferrer" href="https://hollo.social/tags/ActivityPub" target="_blank">#ActivityPub</a>/<a class="mention hashtag" rel="nofollow noopener noreferrer" href="https://hollo.social/tags/fediverse" target="_blank">#fediverse</a> development, building documentation chatbots, or <a class="mention hashtag" rel="nofollow noopener noreferrer" href="https://hollo.social/tags/LLM" target="_blank">#LLM</a>-powered dev tools.<a class="mention hashtag" rel="nofollow noopener noreferrer" href="https://hollo.social/tags/fedidev" target="_blank">#fedidev</a>

Deborah Hartmann Preuss, pcc<a href="https://cosocial.ca/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> results are only as good (at best) as the instructions given:"When an AI script written by a DOGE employee came across a contract for internet service, it flagged it as cancelable. Not because it was waste, fraud or abuse — the Department of Veterans Affairs needs internet connectivity after all — but because the model was given unclear and conflicting instructions." 🤦‍♀️🤷‍♀️ <a href="https://cosocial.ca/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a><a href="https://mastodon.social/@grimalkina" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@grimalkina</a> <a href="https://mastodon.social/@grimalkina/114636068039946105" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://mastodon.social/@grimalkina/114636068039946105</a>

Recent searches

Search options

Administered by:

Server stats:

#llm