Künstliche Intelligenz – Links

KI Links vom 2026-10-24

• Microsoft Teams: KI erkennt Phishing auch ohne Links und Anhänge
  https://www.ad-hoc-news.de/wirtschaft/microsoft-teams-ki-erkennt-phishing-auch-ohne-links-und-anhaenge/70407014 (ad-hoc-news.de)
• "Geht um Geschwindigkeit": Digitalminister Wildberger will KI für Cybersicherheit einsetzen
  https://www.n-tv.de/politik/Digitalminister-Wildberger-will-KI-fuer-Cybersicherheit-einsetzen-id31455042.html (n-tv.de)
• Cognizant-Aktie: Cognizant startet AI Nexus. Die Cognizant-Aktie steht vor Zahlen am 29.10.2026
  https://www.ad-hoc-news.de/boerse/news/nebenwerte/cognizant-aktie-cognizant-startet-ai-nexus-die-cognizant-aktie-steht-vor/70406886 (ad-hoc-news.de)
• Better Retrieval Won’t Fix Your RAG Bot
  https://pub.towardsai.net/better-retrieval-wont-fix-your-rag-bot-9e669a33553f?source=rss----98111c9905da---4 (pub.towardsai.net)
• Microsoft lanza agentes guardianes para vigilar la IA
  https://ecosistemastartup.com/microsoft-lanza-agentes-guardianes-para-vigilar-la-ia/ (ecosistemastartup.com)
• AI Workloads Threaten to Expose Internet's Fragile Foundation
  https://www.techbuzz.ai/articles/ai-workloads-threaten-to-expose-internet-s-fragile-foundation (techbuzz.ai)
• Yama Judges, Chitragupta Keeps the Records: Open-Source AI for Indian Languages
  https://pub.towardsai.net/yama-judges-chitragupta-keeps-the-records-open-source-ai-for-indian-languages-b7f8bf3b7576?source=rss----98111c9905da---4 (pub.towardsai.net)
• Will AI expose the rickety infrastructure of the internet?
  https://techcrunch.com/2026/10/24/will-ai-expose-the-rickety-infrastructure-of-the-internet/ (techcrunch.com)
• Acer Swift 14 AI SF14-11-X8MN - 15,5 Stunden ohne Steckdose
  https://www.ad-hoc-news.de/wirtschaft/produkte/acer-swift-14-ai-sf14-11-x8mn-15-5-stunden-ohne-steckdose/70406413 (ad-hoc-news.de)
• OpenAI llega a Google Cloud: GPT-6 en Model Garden
  https://ecosistemastartup.com/openai-llega-a-google-cloud-gpt-6-en-model-garden-2/ (ecosistemastartup.com)
• Your OpenCode Key Didn’t Expire. The Breach Killed It. Here’s the 5-Minute Fix
  https://pub.towardsai.net/your-opencode-key-didnt-expire-the-breach-killed-it-here-s-the-5-minute-fix-1b75fa363652?source=rss----98111c9905da---4 (pub.towardsai.net)
• The AI graveyard: a running list of projects and startups that didn’t make it
  https://techcrunch.com/2026/10/24/the-ai-graveyard-a-running-list-of-projects-and-startups-that-didnt-make-it/ (techcrunch.com)
• La etiqueta IA ya no sostiene la valoración: el caso WeWork
  https://ecosistemastartup.com/la-etiqueta-ia-ya-no-sostiene-la-valoracion-el-caso-wework/ (ecosistemastartup.com)
• Gemini 4 Argon vs Claude Opus 5.5: Pick Wrong, Pay Twice
  https://pub.towardsai.net/gemini-4-argon-vs-claude-opus-5-5-pick-wrong-pay-twice-fc444c629eaa?source=rss----98111c9905da---4 (pub.towardsai.net)
• OpenAI DevDay: ChatGPT erreicht 1,2 Milliarden wöchentliche Nutzer
  https://www.ad-hoc-news.de/wissenschaft/openai-devday-chatgpt-erreicht-1-2-milliarden-woechentliche-nutzer/70405838 (ad-hoc-news.de)
• Guangzhou, China, Oct.
  https://www.ad-hoc-news.de/boerse/news/corporate-news/guangzhou-china-oct/70405843 (ad-hoc-news.de)
• Google AI Pro in Taiwan: Abrechnungslücke beendet, Abos storniert
  https://www.ad-hoc-news.de/wissenschaft/google-ai-pro-in-taiwan-abrechnungsluecke-beendet-abos-storniert/70405830 (ad-hoc-news.de)
• Sam Altman: la ciberseguridad ya no se arregla con parches
  https://ecosistemastartup.com/sam-altman-la-ciberseguridad-ya-no-se-arregla-con-parches/ (ecosistemastartup.com)
• Axon Announces Body 5: Next-Generation Camera Built for the AI Era
  https://www.prnewswire.com/news-releases/axon-announces-body-5-next-generation-camera-built-for-the-ai-era-302918553.html (prnewswire.com)
• Slow-moving light helps a tiny silicon chip hit 200 Gb/s for AI data centers
  https://interestingengineering.com/science/slow-light-enables-silicon-modulator-achieve-200gbs (interestingengineering.com)
• Zuckerberg (Meta) rechaza la IA única de OpenAI y Google
  https://ecosistemastartup.com/zuckerberg-meta-rechaza-la-ia-unica-de-openai-y-google/ (ecosistemastartup.com)
• Dassault y NVIDIA llevan agentes IA al gemelo virtual
  https://ecosistemastartup.com/dassault-y-nvidia-llevan-agentes-ia-al-gemelo-virtual/ (ecosistemastartup.com)
• IRDAI extends distribution reforms feedback timeline to Oct 31
  https://www.thehindu.com/business/Industry/irdai-extends-distribution-reforms-feedback-timeline-to-oct-31/article71620935.ece (thehindu.com)
• Anthropic Just Introduced Claude Duet. AI That Works While You Step Away
  https://pub.towardsai.net/anthropic-just-introduced-claude-duet-ai-that-works-while-you-step-away-89d26bd4feaa?source=rss----98111c9905da---4 (pub.towardsai.net)
• Google Optimizes Agent Development Kit's 200K-line Python Codebase Overnight with Gemini 4 Argon
  https://www.infoq.com/news/2026/10/google-argon-optimizes-adk/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=AI%2C+ML+%26+Data+Engineering (infoq.com)
• Grok elige republicanos en 92% y GPT, demócratas en 97%
  https://ecosistemastartup.com/grok-elige-republicanos-en-92-y-gpt-democratas-en-97/ (ecosistemastartup.com)
• Three fired OpenAI safety researchers say they were never told what they actually leaked
  https://the-decoder.com/three-fired-openai-safety-researchers-say-they-were-never-told-what-they-actually-leaked/ (the-decoder.com)
• Recursive’s AI Helped Build a Coding Model in Under 4 Months, and It Runs 125 Tokens a Second on a…
  https://pub.towardsai.net/recursives-ai-helped-build-a-coding-model-in-under-4-months-and-it-runs-125-tokens-a-second-on-a-16097573a320?source=rss----98111c9905da---4 (pub.towardsai.net)
• Forced to pick, Grok’s AI model chooses Republicans. OpenAI’s picks Democrats.
  https://techcrunch.com/2026/10/24/forced-to-pick-groks-ai-model-chooses-republicans-openais-picks-democrats/ (techcrunch.com)
• AI makes junior patent attorneys instantly better but may quietly stunt their professional growth
  https://the-decoder.com/ai-makes-junior-patent-attorneys-instantly-better-but-may-quietly-stunt-their-professional-growth/ (the-decoder.com)
• Battledrones, RTS web, mapa 4 jugadores con Claude
  https://ecosistemastartup.com/battledrones-rts-web-mapa-4-jugadores-con-claude/ (ecosistemastartup.com)
• Block's new AI model Dessa was trained entirely on human speech to catch emotions text-based models miss
  https://the-decoder.com/blocks-new-ai-model-dessa-was-trained-entirely-on-human-speech-to-catch-emotions-text-based-models-miss/ (the-decoder.com)
• Meta compra Sundial, la startup de análisis con IA
  https://ecosistemastartup.com/meta-compra-sundial-la-startup-de-analisis-con-ia/ (ecosistemastartup.com)
• Here are the top AI agents that can live in your text messages
  https://techcrunch.com/2026/10/24/all-the-ai-agents-that-can-live-in-your-text-messages/ (techcrunch.com)
• Siemens Aktie: Superkondensatoren für Rechenzentren
  https://www.ad-hoc-news.de/boerse/news/unternehmensnachrichten/siemens-aktie-superkondensatoren-fuer-rechenzentren/70405122 (ad-hoc-news.de)
• Unitree’s Robot Bodyguard Would Die for You. But Can It Tell Who to Fight?
  https://pub.towardsai.net/unitrees-robot-bodyguard-would-die-for-you-but-can-it-tell-who-to-fight-3a3805b09909?source=rss----98111c9905da---4 (pub.towardsai.net)
• FAI erneuert Casa Noha in den Sassi am 22. Oktober
  https://www.ad-hoc-news.de/unterhaltung/reisen/fai-erneuert-casa-noha-in-den-sassi-am-22-oktober/70404809 (ad-hoc-news.de)
• The PHP Meltdown Shows How AI Makes Bugs More Exploitable
  https://www.wired.com/story/kernel-panic-the-php-meltdown-shows-how-ai-makes-bugs-more-exploitable/ (wired.com)
• OpenAI stellt seine neue Trainingsumgebung vor – ohne Air Gap
  https://www.heise.de/news/OpenAI-stellt-seine-neue-Trainingsumgebung-vor-ohne-Air-Gap-11499163.html?wt_mc=rss.red.ho.ho.rdf.beitrag.beitrag (heise.de)
• Wandel im Berufsalltag: KI wird zum Kündigungsgrund
  https://www.wallstreet-online.de/nachricht/21575907-wandel-berufsalltag-ki-kuendigungsgrund (wallstreet-online.de)
• Herzkrankheiten per KI erkennen: Zungen- und Gesichtsfotos erreichen 98,1 Prozent
  https://www.ad-hoc-news.de/wissenschaft/herzkrankheiten-per-ki-erkennen-zungen-und-gesichtsfotos-erreichen-98-1/70404573 (ad-hoc-news.de)
• OpenAI's isolated AI agents found ways to talk to each other through CPU cache
  https://the-decoder.com/openais-isolated-ai-agents-found-ways-to-talk-to-each-other-through-cpu-cache/ (the-decoder.com)
• Palantir: NATO-KI hat keinen Kill-Switch
  https://www.golem.de/news/palantir-nato-ki-hat-keinen-kill-switch-2610-214473.html (golem.de)
• Best AI Tools for Marketing Content (2026): CapCut Recognized for Faster AI-Powered Video and Image Creation by Better Business Advice
  https://www.prnewswire.com/news-releases/best-ai-tools-for-marketing-content-2026-capcut-recognized-for-faster-ai-powered-video-and-image-creation-by-better-business-advice-302918030.html (prnewswire.com)
• Nebius Aktie: AI Builder Program in Europa ausgeweitet
  https://www.ad-hoc-news.de/boerse/news/unternehmensnachrichten/nebius-aktie-ai-builder-program-in-europa-ausgeweitet/70404298 (ad-hoc-news.de)
• UK pledges more funding to guard against AI risks
  https://techxplore.com/news/2026-10-uk-pledges-funding-ai.html (techxplore.com)
• Texan Senate Candidate Sues OpenAI Before Midterms, Accuses Altman Of 'Lying To Users And Harming Texans'
  https://www.ibtimes.com/texan-senate-candidate-sues-openai-before-midterms-accuses-altman-lying-users-harming-texans-3808959 (ibtimes.com)
• Investor Notification: Alphabet ($GOOG, $GOOGL) Investors Are Reminded to Contact BFA Law About the Ongoing Securities Fraud Class Action to Recover Losses
  https://www.wallstreet-online.de/nachricht/21575869-investor-notification-alphabet-goog-googl-investors-are-reminded-to-contact-bfa-law-about-the-ongoing-securities-fraud-class-action-to-recover-losses (wallstreet-online.de)
• Artprice gratuliert Brognon Rollin zum Gewinn des 2026 Marcel Duchamp Prize
  https://www.prnewswire.com/news-releases/artprice-gratuliert-brognon-rollin-zum-gewinn-des-2026-marcel-duchamp-prize-302918621.html (prnewswire.com)
• A timeline of developments in AI safety since the attack on Hugging Face
  https://techxplore.com/news/2026-10-timeline-ai-safety-1.html (techxplore.com)
• Fancy harnesses make coding agents more expensive but not better, study finds
  https://the-decoder.com/fancy-harnesses-make-coding-agents-more-expensive-but-not-better-study-finds/ (the-decoder.com)
• Government and AI giants to set up ‘world first’ pandemic-prevention lab
  https://www.independent.co.uk/news/uk/politics/ai-uk-government-john-healey-pandemic-biosecurity-b3072601.html (independent.co.uk)
• Criminal gangs using AI to ‘make it easier’ to exploit British children
  https://www.independent.co.uk/news/uk/home-news/ai-child-trafficking-criminals-uk-eleanor-lyons-b3071440.html (independent.co.uk)
• Continuous Architecture: KI zwingt uns dazu, veraltete Dokumentation aufzuräumen
  https://www.heise.de/hintergrund/Continuous-Architecture-KI-zwingt-uns-dazu-veraltete-Dokumentation-aufzuraeumen-11497700.html?wt_mc=rss.red.ho.ho.rdf.beitrag.beitrag (heise.de)
• A Walkie Talkie Interface For Claude
  https://hackaday.com/2026/10/24/a-walkie-talkie-interface-for-claude/ (hackaday.com)
• Gemini-App: Notebook-Verknüpfung bleibt im Europäischen Wirtschaftsraum aus
  https://www.ad-hoc-news.de/wissenschaft/gemini-app-notebook-verknuepfung-bleibt-im-europaeischen-wirtschaftsraum/70402747 (ad-hoc-news.de)
• Personal AI assistants could soon reshape the digital economy
  https://www.derstandard.de/story/3000000345576/personal-ai-assistants-could-soon-reshape-the-digital-economy?ref=rss (derstandard.de)
• Personal AI assistants could soon reshape the digital economy
  https://www.derstandard.at/story/3000000345576/personal-ai-assistants-could-soon-reshape-the-digital-economy?ref=rss (derstandard.at)
• The four main risks of turning to AI for mental health advice, according to a psychiatrist
  https://www.independent.co.uk/life-style/ai-chatgpt-claude-advice-uk-mental-health-b3072595.html (independent.co.uk)
• Commerce Ministry identifying AI-proof professions, sectors where India can take global lead: Goyal
  https://www.thehindu.com/news/national/commerce-ministry-identifying-ai-proof-professions-sectors-where-india-can-take-global-lead-goyal/article71619298.ece (thehindu.com)
• AI Tools Optimize Material Compositions for Multi-Channel Signature Control
  https://www.azom.com/news.aspx?newsID=65965 (azom.com)
• Landtag: Debatte um Datensicherheit – AfD-Fraktion plant KI-Server
  https://www.stern.de/gesellschaft/regional/thueringen/landtag--debatte-um-datensicherheit---afd-fraktion-plant-ki-server-38445210.html (stern.de)
• Keeping the world safe when no one is in charge
  https://www.aljazeera.com/opinions/2026/10/24/keeping-the-world-safe-when-no-one-is-in-charge?traffic_source=rss (aljazeera.com)
• Kimi K3.1 Pro: Moonshot AI senkt Hacking-Leistung durch bereinigtes Training
  https://www.ad-hoc-news.de/wissenschaft/kimi-k3-1-pro-moonshot-ai-senkt-hacking-leistung-durch-bereinigtes/70402624 (ad-hoc-news.de)
• Google Messages: Gemini warnt jetzt vor Betrug in unbekannten Chats
  https://www.ad-hoc-news.de/wissenschaft/google-messages-gemini-warnt-jetzt-vor-betrug-in-unbekannten-chats/70402533 (ad-hoc-news.de)
• Terminal-Bench 4.0 Ranks 35 Setups, Not 23 Models
  https://pub.towardsai.net/terminal-bench-4-0-ranks-35-setups-not-23-models-b2c8e62ea927?source=rss----98111c9905da---4 (pub.towardsai.net)
• NVIDIA Brought Rust to CUDA. Here Is How to Keep Your C++ Giving the Same Answer, in 5 Steps
  https://pub.towardsai.net/nvidia-brought-rust-to-cuda-here-is-how-to-keep-your-c-giving-the-same-answer-in-5-steps-9f373ca0e03f?source=rss----98111c9905da---4 (pub.towardsai.net)
• An OpenAI Agent Ripped Out a Safety Monitor, Then Lied About It. Not Cool.
  https://pub.towardsai.net/an-openai-agent-ripped-out-a-safety-monitor-then-lied-about-it-f2499fe079ac?source=rss----98111c9905da---4 (pub.towardsai.net)
• Why Your Browser Agent Clicks Submit Twice, and Four Other Production Mistakes
  https://pub.towardsai.net/browser-agent-production-mistakes-microsoft-foundry-46400fa62677?source=rss----98111c9905da---4 (pub.towardsai.net)
• OpenAI-Klage in Texas: Paxton fordert 10.000 Dollar pro Verstoß
  https://www.ad-hoc-news.de/wissenschaft/openai-klage-in-texas-paxton-fordert-10-000-dollar-pro-verstoss/70402492 (ad-hoc-news.de)
• AI Biosecurity Lab: Großbritannien investiert £100 Millionen in Biosicherheit
  https://www.ad-hoc-news.de/wissenschaft/ai-biosecurity-lab-grossbritannien-investiert-100-millionen-in/70402467 (ad-hoc-news.de)
• Neobookings: reservas de hotel dentro de ChatGPT y Claude
  https://ecosistemastartup.com/neobookings-reservas-de-hotel-dentro-de-chatgpt-y-claude/ (ecosistemastartup.com)
• The rise of AI election ads in US midterms campaign
  https://www.bbc.co.uk/news/videos/cm62y59zw8zko?at_medium=RSS&at_campaign=rss (bbc.co.uk)
• AI and EV Boom Draws $690 Million in Advanced Materials Investment Plans to Thailand
  https://www.wallstreet-online.de/nachricht/21574886-ai-and-ev-boom-draws-690-million-advanced-materials-investment-plans-to-thailand (wallstreet-online.de)
• Nach Hacks: OpenAI will Spitzen-KI ohne Internet trainieren
  https://www.derstandard.de/story/3000000345798/nach-hacks-openai-will-spitzen-ki-ohne-internet-trainieren?ref=rss (derstandard.de)
• Nach Hacks: OpenAI will Spitzen-KI ohne Internet trainieren
  https://www.derstandard.at/story/3000000345798/nach-hacks-openai-will-spitzen-ki-ohne-internet-trainieren?ref=rss (derstandard.at)
• Medizinische Innovation: Wie China auf der Überholspur unterwegs ist
  https://www.aerztezeitung.de/Wirtschaft/Medizinische-Innovation-Wie-China-auf-der-Ueberholspur-unterwegs-ist-464806.html (aerztezeitung.de)
• battledrones v14: Compass Rose, a four player map
  https://simonwillison.net/2026/Oct/24/battledrones/ (simonwillison.net)
• Arbeitswelt und KI: Wandel im Berufsalltag: KI wird zum Kündigungsgrund
  https://www.stern.de/digital/arbeitswelt-und-ki--wandel-im-berufsalltag--ki-wird-zum-kuendigungsgrund-38444810.html (stern.de)
• Eine YouGov-Befragung im Auftrag von SThree zeigt, wie stark KI Tätigkeiten, Belastung und Entwicklungschancen in technischen und naturwissenschaftlichen Berufen verändert.
  https://www.ad-hoc-news.de/wissenschaft/eine-yougov-befragung-im-auftrag-von-sthree-zeigt-wie-stark-ki/70402178 (ad-hoc-news.de)
• Nvidia urge a Europa: ser empresas de IA o quedar fuera
  https://ecosistemastartup.com/nvidia-urge-a-europa-ser-empresas-de-ia-o-quedar-fuera/ (ecosistemastartup.com)
• ChatGPT Ads llega a Argentina: 100.000 anunciantes
  https://ecosistemastartup.com/chatgpt-ads-llega-a-argentina-100-000-anunciantes/ (ecosistemastartup.com)
• US Tech Leaders Are Urgently Calling for Rules on AI — China Already Has Them
  https://cleantechnica.com/2026/10/23/us-tech-leaders-are-urgently-calling-for-rules-on-ai-china-already-has-them/ (cleantechnica.com)
• Microsoft warnt: KI verschärft Ransomware und Cybercrime weltweit
  https://www.ad-hoc-news.de/wissenschaft/microsoft-warnt-ki-verschaerft-ransomware-und-cybercrime-weltweit/70402048 (ad-hoc-news.de)
• Gemini-Modelle: Google lässt KI an der Entwicklung ihrer Nachfolger mitarbeiten
  https://www.ad-hoc-news.de/wissenschaft/gemini-modelle-google-laesst-ki-an-der-entwicklung-ihrer-nachfolger/70402031 (ad-hoc-news.de)
• ChatGPT-Werbung: OpenAI weitet Anzeigen auf 26 weitere Länder aus
  https://www.ad-hoc-news.de/wissenschaft/chatgpt-werbung-openai-weitet-anzeigen-auf-26-weitere-laender-aus/70402029 (ad-hoc-news.de)
• OpenAI: dos agentes apagaron la vigilancia y uno lo negó
  https://ecosistemastartup.com/openai-dos-agentes-apagaron-la-vigilancia-y-uno-lo-nego/ (ecosistemastartup.com)
• Meta supera a OpenAI y Anthropic con Muse en WhatsApp
  https://ecosistemastartup.com/meta-supera-a-openai-y-anthropic-con-muse-en-whatsapp/ (ecosistemastartup.com)
• Nvidia rejects near-term memory glut fears, leaves CXMT cooperation open
  https://www.digitimes.com/news/a20261023PD243/nvidia-cxmt-demand-technology-market.html (digitimes.com)
• Immutable Architecture: Why the Future of AI Models Demands Permanent Storage
  https://hackernoon.com/immutable-architecture-why-the-future-of-ai-models-demands-permanent-storage?source=rss (hackernoon.com)
• Eclectic AI: agentes de IA para WhatsApp, con US$37.893 MRR
  https://ecosistemastartup.com/eclectic-ai-agentes-de-ia-para-whatsapp-con-us37-893-mrr/ (ecosistemastartup.com)
• UK to invest £100m in 'world first' AI Biosecurity Lab
  https://www.bbc.co.uk/news/articles/cqdjv8xllvyyo?at_medium=RSS&at_campaign=rss (bbc.co.uk)
• Confidence Is a Feature: Propagating Uncertainty Through an Agent Stack
  https://pub.towardsai.net/confidence-is-a-feature-propagating-uncertainty-through-an-agent-stack-1c42bb0d519b?source=rss----98111c9905da---4 (pub.towardsai.net)