ferricoxide<p><a href="https://techhub.social/@hosford42" class="u-url mention" rel="nofollow noopener" target="_blank">@hosford42@techhub.social</a><span> <br><br>Still: doesn't bode well for </span><a href="https://evil.social/tags/agenticAI" rel="nofollow noopener" target="_blank">#agenticAI</a><span> solutions.</span></p>
lingo.lol is one of the many independent Mastodon servers you can use to participate in the fediverse.
A place for linguists, philologists, and other lovers of languages.
Administered by:
Server stats:
59active users
lingo.lol: About · Profiles directory · Privacy policy
Mastodon: About · Get the app · Keyboard shortcuts · View source code · v4.4.4
#agenticai
1 post · 1 participant · 0 posts today
Miguel Afonso Caetano<p>"To put it bluntly, the path currently being taken towards agentic AI leads to an elimination of privacy and security at the application layer. It will not be possible for apps like Signal—the messaging app whose foundation I run—to continue to provide strong privacy guarantees, built on robust and openly validated encryption, if device-makers and OS developers insist on puncturing the metaphoric blood-brain barrier between apps and the OS. Feeding your sensitive Signal messages into an undifferentiated data slurry connected to cloud servers in service of their AI-agent aspirations is a dangerous abdication of responsibility.</p><p>Happily, it’s not too late. There is much that can still be done, particularly when it comes to protecting the sanctity of private data. What’s needed is a fundamental shift in how we approach the development and deployment of AI agents. First, privacy must be the default, and control must remain in the hands of application developers exercising agency on behalf of their users. Developers need the ability to designate applications as “sensitive” and mark them as off-limits to agents, at the OS level and otherwise. This cannot be a convoluted workaround buried in settings; it must be a straightforward, well-documented mechanism (similar to Global Privacy Control) that blocks an agent from accessing our data or taking actions within an app.<br>Second, radical transparency must be the norm. Vague assurances and marketing-speak are no longer acceptable. OS vendors have an obligation to be clear and precise about their architecture and what data their AI agents are accessing, how it is being used and the measures in place to protect it."</p><p><a href="https://www.economist.com/by-invitation/2025/09/09/ai-agents-are-coming-for-your-privacy-warns-meredith-whittaker" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">economist.com/by-invitation/20</span><span class="invisible">25/09/09/ai-agents-are-coming-for-your-privacy-warns-meredith-whittaker</span></a></p><p><a href="https://tldr.nettime.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://tldr.nettime.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://tldr.nettime.org/tags/AgenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticAI</span></a> <a href="https://tldr.nettime.org/tags/AIAgents" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAgents</span></a> <a href="https://tldr.nettime.org/tags/Privacy" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Privacy</span></a> <a href="https://tldr.nettime.org/tags/CyberSecurity" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CyberSecurity</span></a> <a href="https://tldr.nettime.org/tags/Surveillance" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Surveillance</span></a></p>
Benjamin Han<p>(Coding with AI: 2/4)</p><p>Remember driving a stick-shift car? Your hand clenched the gearstick while your foot danced on the clutch, trying to nail that gear change. Looking back it's clear that the driver was put to doing a machine's job, but transitioning from stick to automatic is a paradigm shift that has elicited a lot of emotional responses…</p><p>(read Part 2: <a href="https://www.linkedin.com/posts/benjaminhan_ai-genai-coding-activity-7370142721270931456-BwoW" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">linkedin.com/posts/benjaminhan</span><span class="invisible">_ai-genai-coding-activity-7370142721270931456-BwoW</span></a>)</p><p>(part 1: <a href="https://www.linkedin.com/posts/benjaminhan_running-ai-genai-activity-7369569003293372416-F3eV" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">linkedin.com/posts/benjaminhan</span><span class="invisible">_running-ai-genai-activity-7369569003293372416-F3eV</span></a>)</p><p><a href="https://sigmoid.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://sigmoid.social/tags/genAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>genAI</span></a> <a href="https://sigmoid.social/tags/coding" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>coding</span></a> <a href="https://sigmoid.social/tags/programming" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>programming</span></a> <a href="https://sigmoid.social/tags/agenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>agenticAI</span></a> <a href="https://sigmoid.social/tags/softwareEngineering" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>softwareEngineering</span></a> <a href="https://sigmoid.social/tags/futureOfWork" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>futureOfWork</span></a></p>
Miguel Afonso Caetano<p>"The point is that with each advance in AI, new hurdles become apparent; when one missing aspect of “intelligence” is filled in, we find ourselves bumping up against another gap. When I speculated about GPT-5 last year, it didn’t occur to me to question whether it would know how to set priorities, because the models of the time weren’t even capable enough for that to be a limiting factor. In a post from November, AI is Racing Forward – on a Very Long Road, I wrote:</p><p>…the real challenges may be things that we can’t easily anticipate right now, weaknesses that we will only start to put our finger on when we observe [future models] performing astonishing feats and yet somehow still not being able to write that tightly-plotted novel.</p><p>In April 2024, it seemed like agentic AI was going to be the next big thing. The ensuing 16 months have brought enormous progress on many fronts, but very little progress on real-world agency. With projects like AI Village shining a light on the profound weakness of current AI agents, I think robust real-world capability is still years away."</p><p><a href="https://secondthoughts.ai/p/gpt-5-the-case-of-the-missing-agent" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">secondthoughts.ai/p/gpt-5-the-</span><span class="invisible">case-of-the-missing-agent</span></a></p><p><a href="https://tldr.nettime.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://tldr.nettime.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://tldr.nettime.org/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> <a href="https://tldr.nettime.org/tags/Chatbots" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Chatbots</span></a> <a href="https://tldr.nettime.org/tags/AIAgents" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAgents</span></a> <a href="https://tldr.nettime.org/tags/AgenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticAI</span></a> <a href="https://tldr.nettime.org/tags/ReasoningModels" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ReasoningModels</span></a></p>
Pyrzout :vm:<p>Stealthy attack serves poisoned web pages only to AI agents <a href="https://www.helpnetsecurity.com/2025/09/05/ai-agents-prompt-injection-poisoned-web/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">helpnetsecurity.com/2025/09/05</span><span class="invisible">/ai-agents-prompt-injection-poisoned-web/</span></a> <a href="https://social.skynetcloud.site/tags/promptinjection" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>promptinjection</span></a> <a href="https://social.skynetcloud.site/tags/Don" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Don</span></a>'tmiss <a href="https://social.skynetcloud.site/tags/agenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>agenticAI</span></a> <a href="https://social.skynetcloud.site/tags/Hotstuff" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Hotstuff</span></a> <a href="https://social.skynetcloud.site/tags/research" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>research</span></a> <a href="https://social.skynetcloud.site/tags/JFrog" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>JFrog</span></a> <a href="https://social.skynetcloud.site/tags/News" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>News</span></a> <a href="https://social.skynetcloud.site/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a></p>
DiTraRe<p>✨ Wondering how an AI research assistant could work? Join our Colloquium tomorrow at 11:00am and connect with other researchers interested in the topic of "Research Automation with Agentic LLMs". Our guest speaker will be Arman Khalatyan from Leibniz-Institut für Astrophysik Potsdam (AIP) / <a href="https://social.kit.edu/tags/PUNCH4NFDI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>PUNCH4NFDI</span></a>. 🪐 🔭<br>Register here: <a href="https://lnkd.in/eqNJBFHS" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">lnkd.in/eqNJBFHS</span><span class="invisible"></span></a><br>🔗 Zoom link: <a href="https://lnkd.in/eE-sKhDu" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">lnkd.in/eE-sKhDu</span><span class="invisible"></span></a></p><p><a href="https://social.kit.edu/tags/colloquium" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>colloquium</span></a> <a href="https://social.kit.edu/tags/physics" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>physics</span></a> <span class="h-card" translate="no"><a href="https://wisskomm.social/@fiz_karlsruhe" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>fiz_karlsruhe</span></a></span> <span class="h-card" translate="no"><a href="https://sigmoid.social/@AnnaJacyszyn" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>AnnaJacyszyn</span></a></span> <a href="https://social.kit.edu/tags/ditrare" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ditrare</span></a> <span class="h-card" translate="no"><a href="https://chaos.social/@Feelix" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>Feelix</span></a></span> <a href="https://social.kit.edu/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://social.kit.edu/tags/agenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>agenticAI</span></a> <a href="https://social.kit.edu/tags/research" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>research</span></a> <span class="h-card" translate="no"><a href="https://social.kit.edu/@ITAS" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>ITAS</span></a></span> <span class="h-card" translate="no"><a href="https://social.kit.edu/@KIT_Karlsruhe" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>KIT_Karlsruhe</span></a></span></p>
Benjamin Han<p>(Coding with AI: 1/4)</p><p>I love <a href="https://sigmoid.social/tags/running" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>running</span></a>, but I also love coding. I can still code for hours a day. But there’s a limit: I could only get a few hours of quality work in before feeling exhausted. That changed a month ago, when I started experimenting with coding AI. Since then, I’ve had to consciously cut down my coding hours — it’s so much fun (and addictive) that I simply can’t stop!…</p><p>(read: <a href="https://www.linkedin.com/posts/benjaminhan_running-ai-genai-activity-7369569003293372416-F3eV" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">linkedin.com/posts/benjaminhan</span><span class="invisible">_running-ai-genai-activity-7369569003293372416-F3eV</span></a>)</p><p><a href="https://sigmoid.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://sigmoid.social/tags/genAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>genAI</span></a> <a href="https://sigmoid.social/tags/coding" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>coding</span></a> <a href="https://sigmoid.social/tags/programming" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>programming</span></a> <a href="https://sigmoid.social/tags/agenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>agenticAI</span></a> <a href="https://sigmoid.social/tags/softwareEngineering" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>softwareEngineering</span></a> <a href="https://sigmoid.social/tags/futureOfWork" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>futureOfWork</span></a></p>
Benjamin Han<p>5/n</p><p>REFERENCES</p><p>[1] Orion Weller, Michael Boratko, Iftekhar Naim, and Jinhyuk Lee. 2025. On the theoretical limitations of embedding-based retrieval. <a href="https://arxiv.org/abs/2508.21038" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">arxiv.org/abs/2508.21038</span><span class="invisible"></span></a> </p><p>[2] Yifu Qiu, Varun Embar, Yizhe Zhang, Navdeep Jaitly, Shay B. Cohen, and Benjamin Han. 2025. Eliciting in-context Retrieval and reasoning for long-context large language models. <a href="https://machinelearning.apple.com/research/eliciting-in-context" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">machinelearning.apple.com/rese</span><span class="invisible">arch/eliciting-in-context</span></a> repo: <a href="https://github.com/apple/ml-icr2" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">github.com/apple/ml-icr2</span><span class="invisible"></span></a></p><p><a href="https://sigmoid.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://sigmoid.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLM</span></a> <a href="https://sigmoid.social/tags/RAG" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RAG</span></a> <a href="https://sigmoid.social/tags/Embeddings" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Embeddings</span></a> <a href="https://sigmoid.social/tags/Retrieval" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Retrieval</span></a> <a href="https://sigmoid.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> <a href="https://sigmoid.social/tags/Search" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Search</span></a> <a href="https://sigmoid.social/tags/AgenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticAI</span></a> <a href="https://sigmoid.social/tags/paper" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>paper</span></a></p>
Benjamin Han<p>4/</p><p>4. Three ways out: (1) Cross-encoders—placing docs in the prompt of a Long-Context LM; accurate but costly. (2) Multi-vector retrieval. (3) Sparse retrieval.</p><p>For cross-encoders, this links directly to our earlier work on ICR² [2], where combining training data design with model re-architecting improved retrieval performance (see picture 4). Many other paths remain open!</p><p><a href="https://sigmoid.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://sigmoid.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLM</span></a> <a href="https://sigmoid.social/tags/RAG" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RAG</span></a> <a href="https://sigmoid.social/tags/Embeddings" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Embeddings</span></a> <a href="https://sigmoid.social/tags/Retrieval" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Retrieval</span></a> <a href="https://sigmoid.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> <a href="https://sigmoid.social/tags/Search" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Search</span></a> <a href="https://sigmoid.social/tags/AgenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticAI</span></a> <a href="https://sigmoid.social/tags/paper" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>paper</span></a></p>
Benjamin Han<p>3/</p><p>3. LIMIT dataset: To stress-test real models, they construct the LIMIT dataset consisting of 50k docs and 1k queries, each with 2 relevant docs (picture1). All single-vector models fail badly, while BM25 and multi-vector methods perform much better (see picture 2).</p><p><a href="https://sigmoid.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://sigmoid.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLM</span></a> <a href="https://sigmoid.social/tags/RAG" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RAG</span></a> <a href="https://sigmoid.social/tags/Embeddings" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Embeddings</span></a> <a href="https://sigmoid.social/tags/Retrieval" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Retrieval</span></a> <a href="https://sigmoid.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> <a href="https://sigmoid.social/tags/Search" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Search</span></a> <a href="https://sigmoid.social/tags/AgenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticAI</span></a> <a href="https://sigmoid.social/tags/paper" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>paper</span></a></p>
Benjamin Han<p>2/</p><p>2. Oracle experiment: They empirically confirm this bound by directly optimizing embeddings (“free embeddings”) against the relevance matrix. The critical corpus size grows only cubically with dimension. For example, with embedding dimension d = 1024, you can only represent all possible 2-doc query combinations up to about 4 million documents — far below typical web-scale retrieval needs.</p><p><a href="https://sigmoid.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://sigmoid.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLM</span></a> <a href="https://sigmoid.social/tags/RAG" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RAG</span></a> <a href="https://sigmoid.social/tags/Embeddings" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Embeddings</span></a> <a href="https://sigmoid.social/tags/Retrieval" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Retrieval</span></a> <a href="https://sigmoid.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> <a href="https://sigmoid.social/tags/Search" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Search</span></a> <a href="https://sigmoid.social/tags/AgenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticAI</span></a> <a href="https://sigmoid.social/tags/paper" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>paper</span></a></p>
Benjamin Han<p>1/</p><p>Embeddings are the beating heart of modern AI—powering RAG and serving as memory for agentic AI. But a new paper [1] shows a ceiling:</p><p>1. Dot-product retrieval is bounded by embedding dimension d; if the relevance matrix has sign-rank r, then d >= r is required—and no amount of training can avoid it.</p><p><a href="https://sigmoid.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://sigmoid.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLM</span></a> <a href="https://sigmoid.social/tags/RAG" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RAG</span></a> <a href="https://sigmoid.social/tags/Embeddings" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Embeddings</span></a> <a href="https://sigmoid.social/tags/Retrieval" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Retrieval</span></a> <a href="https://sigmoid.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> <a href="https://sigmoid.social/tags/Search" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Search</span></a> <a href="https://sigmoid.social/tags/AgenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticAI</span></a> <a href="https://sigmoid.social/tags/paper" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>paper</span></a></p>
Quinn Comendant<p>New <a href="https://mastodon.social/tags/Claude" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Claude</span></a> T&C require toggling a checkbox to disallow sharing your chats and code: <a href="https://claude.ai/settings/data-privacy-controls" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">claude.ai/settings/data-privac</span><span class="invisible">y-controls</span></a> <a href="https://mastodon.social/tags/privacy" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>privacy</span></a> <a href="https://mastodon.social/tags/llm" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>llm</span></a> <a href="https://mastodon.social/tags/claudeCode" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>claudeCode</span></a> <a href="https://mastodon.social/tags/agenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>agenticAI</span></a></p>
Marcel Waldvogel<p>➡️ Agentenbasierte KI-Browser wie Comet von Perplexity sollen uns einen Teil unserer täglichen Bildschirmarbeit abnehmen, berichtet Futurezone. Doch dieser «Agent» klingt eher nach Doppelagent: Denn er lässt sich von Scam-Webseiten sehr einfach übertölpeln. Die Cybersicherheitsfirma Guardio fasst die Resultate ihrer Studie wie folgt zusammen:</p><p>«Ein kleiner Schritt für KI-Agenten, ein grosser Rückschritt für unsere Sicherheit.»</p><p><a href="https://waldvogel.family/tags/AgenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticAI</span></a><br><a href="https://futurezone.at/digital-life/ki-browser-comet-perplexity-scam-phishing-website-mail-betrug-guardio-scamlexity/403075829" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">futurezone.at/digital-life/ki-</span><span class="invisible">browser-comet-perplexity-scam-phishing-website-mail-betrug-guardio-scamlexity/403075829</span></a></p>
United States News Beep<p>Will Coding AI Tools Ever Reach Full Autonomy?</p><p>Artificial intelligence (AI) has transformed the coding sphere, with AI coding tools completing source code, correcting syntax errors,…<br><a href="https://newsbeep.org/tags/NewsBeep" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>NewsBeep</span></a> <a href="https://newsbeep.org/tags/News" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>News</span></a> <a href="https://newsbeep.org/tags/US" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>US</span></a> <a href="https://newsbeep.org/tags/USA" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>USA</span></a> <a href="https://newsbeep.org/tags/UnitedStates" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>UnitedStates</span></a> <a href="https://newsbeep.org/tags/UnitedStatesOfAmerica" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>UnitedStatesOfAmerica</span></a> <a href="https://newsbeep.org/tags/Artificialintelligence" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Artificialintelligence</span></a> <a href="https://newsbeep.org/tags/agenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>agenticAI</span></a> <a href="https://newsbeep.org/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://newsbeep.org/tags/ArtificialIntelligence" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ArtificialIntelligence</span></a> <a href="https://newsbeep.org/tags/coding" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>coding</span></a> <a href="https://newsbeep.org/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenerativeAI</span></a> <a href="https://newsbeep.org/tags/Programming" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Programming</span></a> <a href="https://newsbeep.org/tags/SoftwareDevelopment" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>SoftwareDevelopment</span></a> <a href="https://newsbeep.org/tags/softwareengineering" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>softwareengineering</span></a> <a href="https://newsbeep.org/tags/Technology" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Technology</span></a><br><a href="https://www.newsbeep.com/us/111212/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="">newsbeep.com/us/111212/</span><span class="invisible"></span></a></p>
Harald Sack<p>AI use cases introduced by Rob Finn from EMBL-EBI, as pointed out in his <a href="https://sigmoid.social/tags/CORDI2025" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CORDI2025</span></a> keynote "Delivering life science data resources in a world of growing data and impacts from AI" </p><p><a href="https://www.nfdi.de/cordi-2025/keynotes/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="">nfdi.de/cordi-2025/keynotes/</span><span class="invisible"></span></a></p><p><a href="https://sigmoid.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://sigmoid.social/tags/agenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>agenticAI</span></a> <a href="https://sigmoid.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> <a href="https://sigmoid.social/tags/lifesciences" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>lifesciences</span></a> <a href="https://sigmoid.social/tags/ebi" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ebi</span></a> <a href="https://sigmoid.social/tags/reseachdata" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>reseachdata</span></a> <a href="https://sigmoid.social/tags/NFDIrocks" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>NFDIrocks</span></a> <span class="h-card" translate="no"><a href="https://nfdi.social/@NFDI" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>NFDI</span></a></span> <a href="https://sigmoid.social/tags/copilot" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>copilot</span></a> <a href="https://sigmoid.social/tags/agenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>agenticAI</span></a></p>
5h15h<p>great to see many of the Microsoft for Startups partners in the list : <a href="https://www.cbinsights.com/research/ai-agent-tech-stack/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">cbinsights.com/research/ai-age</span><span class="invisible">nt-tech-stack/</span></a> <a href="https://techhub.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://techhub.social/tags/AgenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticAI</span></a> <a href="https://techhub.social/tags/AIAgents" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAgents</span></a> <a href="https://techhub.social/tags/GenAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenAI</span></a> <a href="https://techhub.social/tags/msft4startups" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>msft4startups</span></a></p>
5h15h<p>The top 3 global cloud providers — Amazon, Microsoft, and Google are expanding their <a href="https://techhub.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> agent offerings across development tooling, hosting, orchestration, and more <a href="https://www.cbinsights.com/research/ai-agent-tech-stack/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">cbinsights.com/research/ai-age</span><span class="invisible">nt-tech-stack/</span></a> </p><p><a href="https://techhub.social/tags/AIAgents" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIAgents</span></a> <a href="https://techhub.social/tags/AgenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticAI</span></a> <a href="https://techhub.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://techhub.social/tags/GenAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>GenAI</span></a></p>
Jan :rust: :ferris:<p>New <a href="https://floss.social/tags/research" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>research</span></a> by <a href="https://floss.social/tags/Salesforce" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Salesforce</span></a> about <a href="https://floss.social/tags/LLMs" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLMs</span></a> and their tool use via <a href="https://floss.social/tags/MCP" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>MCP</span></a> 🚨 </p><p>MCP-Universe: Benchmarking Large Language Models with Real-World Model Context Protocol Servers:</p><p><a href="https://arxiv.org/abs/2508.14704" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">arxiv.org/abs/2508.14704</span><span class="invisible"></span></a></p><p>Reliability of LLMs using MCP for real-world tasks:<br>- GPT-5 => 43.72%<br>- Grok-4 => 33.33%<br>- Claude-4.0-Sonnet => 29.44%</p><p>Glad there are no negative percentages, I guess!?😌 </p><p>So let me get this straight: we are burning the planet for... not even a coinflip!? 🪙 🎲 </p><p>🎈 💥 </p><p><a href="https://floss.social/tags/LLM" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>LLM</span></a> <a href="https://floss.social/tags/AI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AI</span></a> <a href="https://floss.social/tags/AIHype" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AIHype</span></a> <a href="https://floss.social/tags/AgenticAI" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>AgenticAI</span></a></p>
Wulfy<p><span class="h-card" translate="no"><a href="https://infosec.exchange/@JessTheUnstill" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>JessTheUnstill</span></a></span> <span class="h-card" translate="no"><a href="https://social.chriswb.dev/@chrisw_b" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>chrisw_b</span></a></span> </p><p>AND YOU ALL ESCHEW AI!!!</p><p>Primary use case is for an AI agent to pretend to be you, occasionally throw in "brilliant" and emote "👍".<br>Then give you a 25 word summary after 3.5h</p><p><a href="https://infosec.exchange/tags/agenticai" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>agenticai</span></a> <a href="https://infosec.exchange/tags/agentai" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>agentai</span></a></p>
TrendingLive feeds
Mastodon is the best way to keep up with what's happening.
Follow anyone across the fediverse and see it all in chronological order. No algorithms, ads, or clickbait in sight.
Create accountLoginDrag & drop to upload