<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Intelligenz Welt</title><link>https://intelligenz-welt.de</link><description>KI- und LLM-News</description><language>de</language><item><title>Musk’s AI chatbot Grok reportedly encouraged Trump to capture Venezuela’s president</title><link>https://techcrunch.com/2026/10/01/musks-ai-chatbot-grok-reportedly-encouraged-trump-to-capture-venezuelas-president/</link><guid>https://techcrunch.com/2026/10/01/musks-ai-chatbot-grok-reportedly-encouraged-trump-to-capture-venezuelas-president/</guid><description>President Trump reportedly asked for Grok&#x27;s opinion before invading Venezuela and capturing Nicolás Maduro.</description><pubDate>Thu, 01 Oct 2026 21:08:11 +0000</pubDate><source url="https://techcrunch.com/2026/10/01/musks-ai-chatbot-grok-reportedly-encouraged-trump-to-capture-venezuelas-president/">TechCrunch AI</source></item><item><title>Google’s new Guided Vision feature can help you read the fine print</title><link>https://www.theverge.com/ai-artificial-intelligence/1003756/google-gemini-live-guided-vision</link><guid>https://www.theverge.com/ai-artificial-intelligence/1003756/google-gemini-live-guided-vision</guid><description>Guided Vision is launching in Gemini Live on compatible Android devices today to use AI to give real-time audio descriptions of anything you point your phone&#x27;s camera at. By sharing your camera in Gemini Live, you can ha</description><pubDate>Thu, 01 Oct 2026 19:47:51 +0000</pubDate><source url="https://www.theverge.com/ai-artificial-intelligence/1003756/google-gemini-live-guided-vision">The Verge AI</source></item><item><title>ChatGPT can now virtually try on clothes for you</title><link>https://techcrunch.com/2026/10/01/chatgpt-can-now-virtually-try-on-clothes-for-you/</link><guid>https://techcrunch.com/2026/10/01/chatgpt-can-now-virtually-try-on-clothes-for-you/</guid><description>OpenAI is rolling out new shopping features for ChatGPT that let users virtually try on clothing and accessories using their own photos and save products they like to a Favorites library.</description><pubDate>Thu, 01 Oct 2026 19:21:53 +0000</pubDate><source url="https://techcrunch.com/2026/10/01/chatgpt-can-now-virtually-try-on-clothes-for-you/">TechCrunch AI</source></item><item><title>Google thinks SpaceX’s Starship has to launch 1,800 times before space data centers get off the ground</title><link>https://techcrunch.com/2026/10/01/google-thinks-spacexs-starship-has-to-launch-1600-times-before-space-data-centers-get-off-the-ground/</link><guid>https://techcrunch.com/2026/10/01/google-thinks-spacexs-starship-has-to-launch-1600-times-before-space-data-centers-get-off-the-ground/</guid><description>Google launched its first advanced chip into orbit to pave the way for space data centers.</description><pubDate>Thu, 01 Oct 2026 19:18:03 +0000</pubDate><source url="https://techcrunch.com/2026/10/01/google-thinks-spacexs-starship-has-to-launch-1600-times-before-space-data-centers-get-off-the-ground/">TechCrunch AI</source></item><item><title>OpenAI cuts ties with 3 safety researchers, WSJ reports</title><link>https://techcrunch.com/2026/10/01/openai-cuts-ties-with-three-safety-researchers-wsj-reports/</link><guid>https://techcrunch.com/2026/10/01/openai-cuts-ties-with-three-safety-researchers-wsj-reports/</guid><description>OpenAI has parted ways with three safety researchers after an internal investigation found they mishandled sensitive company information, report says.</description><pubDate>Thu, 01 Oct 2026 18:14:42 +0000</pubDate><source url="https://techcrunch.com/2026/10/01/openai-cuts-ties-with-three-safety-researchers-wsj-reports/">TechCrunch AI</source></item><item><title>Tavus stellt mit Griffin ein KI-Modell vor, das Videogespräche fast so menschlich führt wie ein Mensch</title><link>https://the-decoder.de/tavus-stellt-mit-griffin-ein-ki-modell-vor-das-videogespraeche-fast-so-menschlich-fuehrt-wie-ein-mensch/</link><guid>https://the-decoder.de/tavus-stellt-mit-griffin-ein-ki-modell-vor-das-videogespraeche-fast-so-menschlich-fuehrt-wie-ein-mensch/</guid><description>Tavus stellt mit Griffin das nach eigenen Angaben erste &quot;Human Interaction Model&quot; vor: eine KI, die Videogespräche in Echtzeit führt und dabei Mimik, Tonfall und Gesten verarbeitet. In einer Studie hielten 48 Prozent der</description><pubDate>Thu, 01 Oct 2026 17:53:40 +0000</pubDate><source url="https://the-decoder.de/tavus-stellt-mit-griffin-ein-ki-modell-vor-das-videogespraeche-fast-so-menschlich-fuehrt-wie-ein-mensch/">THE DECODER</source></item><item><title>Opus 5.5 loves to tell you ‘this matters’ (and other AI writing tells)</title><link>https://techcrunch.com/2026/10/01/opus-5-5-loves-to-tell-you-this-matters-and-other-ai-writing-tells/</link><guid>https://techcrunch.com/2026/10/01/opus-5-5-loves-to-tell-you-this-matters-and-other-ai-writing-tells/</guid><description>Opus 5.5’s biggest tell is the word “dependable,” which pops up 23 times more often than in human samples.</description><pubDate>Thu, 01 Oct 2026 17:50:19 +0000</pubDate><source url="https://techcrunch.com/2026/10/01/opus-5-5-loves-to-tell-you-this-matters-and-other-ai-writing-tells/">TechCrunch AI</source></item><item><title>Ideogram 4.5 verspricht stabile KI-Bildbearbeitung ohne ungewollte Veränderungen</title><link>https://the-decoder.de/ideogram-4-5-verspricht-stabile-ki-bildbearbeitung-ohne-ungewollte-veraenderungen/</link><guid>https://the-decoder.de/ideogram-4-5-verspricht-stabile-ki-bildbearbeitung-ohne-ungewollte-veraenderungen/</guid><description>Ideogram stellt mit Ideogram 4.5 ein Bildmodell vor, das bei der Editierung nur die gewünschten Bereiche verändert und den Rest des Bildes stabil hält. Das Modell bietet native 2K-Auflösung ab 0,8 Cent pro Bild und wird </description><pubDate>Thu, 01 Oct 2026 17:25:03 +0000</pubDate><source url="https://the-decoder.de/ideogram-4-5-verspricht-stabile-ki-bildbearbeitung-ohne-ungewollte-veraenderungen/">THE DECODER</source></item><item><title>Judge dismisses antitrust lawsuits over Google’s AI Overviews</title><link>https://www.theverge.com/tech/1003589/google-ai-overviews-chegg-penske-lawsuits-dismissed</link><guid>https://www.theverge.com/tech/1003589/google-ai-overviews-chegg-penske-lawsuits-dismissed</guid><description>A federal judge has dismissed a pair of antitrust lawsuits filed by Chegg and Rolling Stone parent company Penske Media Corporation, which accused Google of driving away web traffic with its AI-powered search features, a</description><pubDate>Thu, 01 Oct 2026 17:17:36 +0000</pubDate><source url="https://www.theverge.com/tech/1003589/google-ai-overviews-chegg-penske-lawsuits-dismissed">The Verge AI</source></item><item><title>Amazon releases its own Jev clone as decision models flood the web</title><link>https://techcrunch.com/2026/10/01/amazon-releases-its-own-jev-clone-as-decision-models-flood-the-web/</link><guid>https://techcrunch.com/2026/10/01/amazon-releases-its-own-jev-clone-as-decision-models-flood-the-web/</guid><description>Amazon Web Services&#x27; Strand Labs has released the latest Jevalike decision model, Strands Decider 2B.</description><pubDate>Thu, 01 Oct 2026 16:49:22 +0000</pubDate><source url="https://techcrunch.com/2026/10/01/amazon-releases-its-own-jev-clone-as-decision-models-flood-the-web/">TechCrunch AI</source></item><item><title>Shopify debuts Canvas, a way to build online stores by chatting with AI</title><link>https://techcrunch.com/2026/10/01/shopify-debuts-canvas-a-way-to-build-online-stores-by-chatting-with-ai/</link><guid>https://techcrunch.com/2026/10/01/shopify-debuts-canvas-a-way-to-build-online-stores-by-chatting-with-ai/</guid><description>Shopify’s new Canvas site builder lets merchants create and customize their online stores by chatting with its AI agent Sidekick, while watching the changes happen in real time.</description><pubDate>Thu, 01 Oct 2026 16:44:35 +0000</pubDate><source url="https://techcrunch.com/2026/10/01/shopify-debuts-canvas-a-way-to-build-online-stores-by-chatting-with-ai/">TechCrunch AI</source></item><item><title>KI-Übersichten: Google vergütet Publisher – Spanne offenbar enorm</title><link>https://www.heise.de/news/Google-bezahlt-offenbar-rund-100-Publisher-fuer-KI-Antworten-in-der-Suche-11473155.html?wt_mc=rss.red.ho.ho.atom.beitrag.beitrag</link><guid>https://www.heise.de/news/Google-bezahlt-offenbar-rund-100-Publisher-fuer-KI-Antworten-in-der-Suche-11473155.html?wt_mc=rss.red.ho.ho.atom.beitrag.beitrag</guid><description>Google bezahlt wohl rund 100 Publisher für Inhalte in KI-Übersichten. Die Vergütungen fallen jedoch angeblich sehr unterschiedlich aus.</description><pubDate>Thu, 01 Oct 2026 15:45:00 +0000</pubDate><source url="https://www.heise.de/news/Google-bezahlt-offenbar-rund-100-Publisher-fuer-KI-Antworten-in-der-Suche-11473155.html?wt_mc=rss.red.ho.ho.atom.beitrag.beitrag">heise online</source></item><item><title>Brian Chesky interview: AI agents need their own operating system</title><link>https://techcrunch.com/2026/10/01/brian-chesky-interview-ai-agents-need-their-own-operating-system/</link><guid>https://techcrunch.com/2026/10/01/brian-chesky-interview-ai-agents-need-their-own-operating-system/</guid><description>Brian Chesky on making Airbnb agent-friendly, the state of consumer AI, and why the world needs an AI-native operating system.</description><pubDate>Thu, 01 Oct 2026 15:12:00 +0000</pubDate><source url="https://techcrunch.com/2026/10/01/brian-chesky-interview-ai-agents-need-their-own-operating-system/">TechCrunch AI</source></item><item><title>Introducing Olmo-core 3: Open, scalable training infrastructure for large MoEs</title><link>https://huggingface.co/blog/allenai/olmocore3</link><guid>https://huggingface.co/blog/allenai/olmocore3</guid><description></description><pubDate>Thu, 01 Oct 2026 15:01:43 +0000</pubDate><source url="https://huggingface.co/blog/allenai/olmocore3">Hugging Face Blog</source></item><item><title>Inside our months-long investigation into Kevin O’Leary’s Utah data center debacle</title><link>https://www.theverge.com/podcast/1002851/utah-ai-data-center-stratos-kevin-oleary-investigation-backlash</link><guid>https://www.theverge.com/podcast/1002851/utah-ai-data-center-stratos-kevin-oleary-investigation-backlash</guid><description>Today I’m talking with Josh Dzieza, a longtime features writer here at The Verge, about Kevin O’Leary’s plans to build a massive data center in Utah.  The idea was to build the world’s biggest data center — a 40,000-acre</description><pubDate>Thu, 01 Oct 2026 14:39:48 +0000</pubDate><source url="https://www.theverge.com/podcast/1002851/utah-ai-data-center-stratos-kevin-oleary-investigation-backlash">The Verge AI</source></item><item><title>OpenAI’s new agent is a shot at Meta — but can it compete with free?</title><link>https://www.theverge.com/ai-artificial-intelligence/1003399/meta-openai-ai-agents-muse-dots-battle</link><guid>https://www.theverge.com/ai-artificial-intelligence/1003399/meta-openai-ai-agents-muse-dots-battle</guid><description>At OpenAI&#x27;s annual DevDay conference, the company pulled out all the stops to compete with its rivals - primarily Meta, whose Muse AI agent platform has seen early runaway success. CEO Sam Altman walked onstage to cheers</description><pubDate>Thu, 01 Oct 2026 14:38:29 +0000</pubDate><source url="https://www.theverge.com/ai-artificial-intelligence/1003399/meta-openai-ai-agents-muse-dots-battle">The Verge AI</source></item><item><title>Claude for Government: Anthropic bringt KI-Agenten in US-Bundesbehörden</title><link>https://the-decoder.de/claude-for-government-anthropic-bringt-ki-agenten-in-us-bundesbehoerden/</link><guid>https://the-decoder.de/claude-for-government-anthropic-bringt-ki-agenten-in-us-bundesbehoerden/</guid><description>Anthropic stellt Claude for Government ab sofort für US-Bundes- und Staatsbehörden bereit. Die Plattform läuft in einer FedRAMP-High-Umgebung, der strengsten US-Sicherheitsstufe für Cloud-Dienste. Behörden erhalten diese</description><pubDate>Thu, 01 Oct 2026 14:10:53 +0000</pubDate><source url="https://the-decoder.de/claude-for-government-anthropic-bringt-ki-agenten-in-us-bundesbehoerden/">THE DECODER</source></item><item><title>Photon held a funeral for mobile apps. Now it has $4.5M to help replace them with agents.</title><link>https://techcrunch.com/2026/10/01/photon-held-a-funeral-for-mobile-apps-now-it-has-4-5m-to-help-replace-them-with-agents/</link><guid>https://techcrunch.com/2026/10/01/photon-held-a-funeral-for-mobile-apps-now-it-has-4-5m-to-help-replace-them-with-agents/</guid><description>The startup helps developers build AI agents that work over iMessage, SMS/RCS, email, and other messaging platforms. It&#x27;s a bet that consumers will increasingly use agents instead of downloading apps.</description><pubDate>Thu, 01 Oct 2026 14:00:00 +0000</pubDate><source url="https://techcrunch.com/2026/10/01/photon-held-a-funeral-for-mobile-apps-now-it-has-4-5m-to-help-replace-them-with-agents/">TechCrunch AI</source></item><item><title>WTF: KI-Regeln: Trump und Tech-Riesen unterschreiben gemeinsame Erklärung</title><link>https://www.heise.de/news/Gemeinsame-Erklaerung-Trump-und-Tech-Riesen-definieren-KI-Regeln-11472591.html?wt_mc=rss.red.ho.ho.atom.beitrag.beitrag</link><guid>https://www.heise.de/news/Gemeinsame-Erklaerung-Trump-und-Tech-Riesen-definieren-KI-Regeln-11472591.html?wt_mc=rss.red.ho.ho.atom.beitrag.beitrag</guid><description>US-Präsident Trump und führende Tech-Konzerne unterzeichnen eine Erklärung zu KI-Kontrollen. Besonders spannend war dabei die Sitzordnung.</description><pubDate>Thu, 01 Oct 2026 14:00:00 +0000</pubDate><source url="https://www.heise.de/news/Gemeinsame-Erklaerung-Trump-und-Tech-Riesen-definieren-KI-Regeln-11472591.html?wt_mc=rss.red.ho.ho.atom.beitrag.beitrag">heise online</source></item><item><title>Sicherheits-Startup findet über 13.000 interne Firmen-Screenshots, die KI-Agenten öffentlich hochgeladen haben</title><link>https://the-decoder.de/sicherheits-startup-findet-ueber-13-000-interne-firmen-screenshots-die-ki-agenten-oeffentlich-hochgeladen-haben/</link><guid>https://the-decoder.de/sicherheits-startup-findet-ueber-13-000-interne-firmen-screenshots-die-ki-agenten-oeffentlich-hochgeladen-haben/</guid><description>KI-Agenten haben unbemerkt über 13.000 interne Screenshots von 343 Organisationen, darunter Fortune-500-Firmen, auf öffentlichen GitHub-Repos veröffentlicht. Weil die Plattform keinen geschützten Upload erlaubte, fanden </description><pubDate>Thu, 01 Oct 2026 13:14:36 +0000</pubDate><source url="https://the-decoder.de/sicherheits-startup-findet-ueber-13-000-interne-firmen-screenshots-die-ki-agenten-oeffentlich-hochgeladen-haben/">THE DECODER</source></item><item><title>Hearing tech startup Legato launches its AI hearing glasses</title><link>https://techcrunch.com/2026/10/01/hearing-tech-startup-legato-launches-its-ai-hearing-glasses/</link><guid>https://techcrunch.com/2026/10/01/hearing-tech-startup-legato-launches-its-ai-hearing-glasses/</guid><description>The glasses stem from the startup’s goal of making hearing care more accessible by addressing the cost, comfort, and stigma associated with traditional hearing aids.</description><pubDate>Thu, 01 Oct 2026 13:00:00 +0000</pubDate><source url="https://techcrunch.com/2026/10/01/hearing-tech-startup-legato-launches-its-ai-hearing-glasses/">TechCrunch AI</source></item><item><title>Satlyt, founded by a former Google and SpaceX product manager, raises $8M to run AI on satellites</title><link>https://techcrunch.com/2026/10/01/satlyt-founded-by-a-former-google-and-spacex-product-manager-raises-8m-to-run-ai-on-satellites/</link><guid>https://techcrunch.com/2026/10/01/satlyt-founded-by-a-former-google-and-spacex-product-manager-raises-8m-to-run-ai-on-satellites/</guid><description>Satlyt wants to be the Android of orbital computing, offering open software that works on many companies&#x27; satellites, versus SpaceX&#x27;s closed, all-in-one iPhone-style approach.</description><pubDate>Thu, 01 Oct 2026 12:00:00 +0000</pubDate><source url="https://techcrunch.com/2026/10/01/satlyt-founded-by-a-former-google-and-spacex-product-manager-raises-8m-to-run-ai-on-satellites/">TechCrunch AI</source></item><item><title>Ataraxos: Neue KI knackt Stratego als eines der letzten Brettspiele, in denen Menschen überlegen waren</title><link>https://the-decoder.de/ataraxos-neue-ki-knackt-stratego-als-eines-der-letzten-brettspiele-in-denen-menschen-ueberlegen-waren/</link><guid>https://the-decoder.de/ataraxos-neue-ki-knackt-stratego-als-eines-der-letzten-brettspiele-in-denen-menschen-ueberlegen-waren/</guid><description>Das KI-System Ataraxos hat den besten Stratego-Spieler aller Zeiten klar geschlagen. Das Brettspiel gilt als harte Prüfung für KI, weil beide Seiten ihre Figuren verdeckt aufstellen. Google Deepmind war daran trotz Milli</description><pubDate>Thu, 01 Oct 2026 11:29:02 +0000</pubDate><source url="https://the-decoder.de/ataraxos-neue-ki-knackt-stratego-als-eines-der-letzten-brettspiele-in-denen-menschen-ueberlegen-waren/">THE DECODER</source></item><item><title>Elon Musk’s Grokipedia has a ‘newly refreshed’ design</title><link>https://www.theverge.com/tech/1003068/elon-musk-grokipedia-v-0-3-spacexai</link><guid>https://www.theverge.com/tech/1003068/elon-musk-grokipedia-v-0-3-spacexai</guid><description>Grokipedia, SpaceXAI&#x27;s AI-powered competitor to Wikipedia, recently started incorporating edits again, and today, it got some design tweaks as part of a v0.3 update, including a new logo and refreshes to its homepage and</description><pubDate>Thu, 01 Oct 2026 10:04:13 +0000</pubDate><source url="https://www.theverge.com/tech/1003068/elon-musk-grokipedia-v-0-3-spacexai">The Verge AI</source></item><item><title>OpenAI wirft Kimi-Entwickler Moonshot AI vor, Denkprozesse seiner Modelle abgegriffen zu haben</title><link>https://the-decoder.de/openai-wirft-kimi-entwickler-moonshot-ai-vor-denkprozesse-seiner-modelle-abgegriffen-zu-haben/</link><guid>https://the-decoder.de/openai-wirft-kimi-entwickler-moonshot-ai-vor-denkprozesse-seiner-modelle-abgegriffen-zu-haben/</guid><description>OpenAI hat eine koordinierte Kampagne gestoppt, bei der über 15.000 Konten versuchten, die versteckten Denkschritte seiner Modelle zu kopieren – einen Teil der Aktivität ordnet das Unternehmen Personen im Umfeld von Moon</description><pubDate>Thu, 01 Oct 2026 09:52:49 +0000</pubDate><source url="https://the-decoder.de/openai-wirft-kimi-entwickler-moonshot-ai-vor-denkprozesse-seiner-modelle-abgegriffen-zu-haben/">THE DECODER</source></item><item><title>Google releases Gemini 4 Argon, called its most powerful model yet</title><link>https://techcrunch.com/2026/09/30/google-releases-gemini-4-argon-called-its-most-powerful-model-yet/</link><guid>https://techcrunch.com/2026/09/30/google-releases-gemini-4-argon-called-its-most-powerful-model-yet/</guid><description>Google has released its latest Gemini model, marketing it as a workhorse for coding and cybersecurity work.</description><pubDate>Wed, 30 Sep 2026 23:43:07 +0000</pubDate><source url="https://techcrunch.com/2026/09/30/google-releases-gemini-4-argon-called-its-most-powerful-model-yet/">TechCrunch AI</source></item><item><title>Google Gemini 4 Argon schließt zur KI-Spitze auf, bleibt aber hinter Anthropics Claude Opus 5.5</title><link>https://the-decoder.de/google-gemini-4-argon-schliesst-zur-ki-spitze-auf-bleibt-aber-hinter-anthropics-claude-opus-5-5/</link><guid>https://the-decoder.de/google-gemini-4-argon-schliesst-zur-ki-spitze-auf-bleibt-aber-hinter-anthropics-claude-opus-5-5/</guid><description>Google stellt mit Gemini 4 Argon sein erstes Frontier-Modell seit mehr als sieben Monaten vor. Laut Artificial Analysis erreicht es das Niveau von GPT-6 Astra, bleibt aber hinter Anthropics Claude Opus 5.5. Der Tokenprei</description><pubDate>Wed, 30 Sep 2026 21:42:30 +0000</pubDate><source url="https://the-decoder.de/google-gemini-4-argon-schliesst-zur-ki-spitze-auf-bleibt-aber-hinter-anthropics-claude-opus-5-5/">THE DECODER</source></item><item><title>Valor, Atreides, and Sequoia back AI startup Flow Engineering at $750M valuation</title><link>https://techcrunch.com/2026/09/30/valor-atreides-and-sequoia-back-ai-startup-flow-engineering-at-750m-valuation/</link><guid>https://techcrunch.com/2026/09/30/valor-atreides-and-sequoia-back-ai-startup-flow-engineering-at-750m-valuation/</guid><description>Flow Engineering, which is bringing AI agents to hardware design, also landed Roelof Botha as an angel investor and board member.</description><pubDate>Wed, 30 Sep 2026 21:07:40 +0000</pubDate><source url="https://techcrunch.com/2026/09/30/valor-atreides-and-sequoia-back-ai-startup-flow-engineering-at-750m-valuation/">TechCrunch AI</source></item><item><title>Google announces Gemini 4 and says it’s so capable that only ‘trusted cyber defenders’ can have it right now</title><link>https://www.theverge.com/tech/1002980/google-gemini-4-argon</link><guid>https://www.theverge.com/tech/1002980/google-gemini-4-argon</guid><description>Google today revealed its next AI frontier model, which it&#x27;s calling Gemini 4 Argon. The new model delivers &quot;frontier performance in complex workflows across real-world software engineering, enterprise knowledge work lik</description><pubDate>Wed, 30 Sep 2026 20:41:41 +0000</pubDate><source url="https://www.theverge.com/tech/1002980/google-gemini-4-argon">The Verge AI</source></item><item><title>The AI Tamagotchis are coming</title><link>https://www.theverge.com/ai-artificial-intelligence/1002779/openai-dots-meta-muse-ai-agents-hardware-devices</link><guid>https://www.theverge.com/ai-artificial-intelligence/1002779/openai-dots-meta-muse-ai-agents-hardware-devices</guid><description>While AI has made plenty of inroads on people&#x27;s phones and computers, it&#x27;s largely failed in dedicated devices. But over the next year, two major AI companies, Meta and OpenAI, will attempt to change that. And they&#x27;re ap</description><pubDate>Wed, 30 Sep 2026 20:19:43 +0000</pubDate><source url="https://www.theverge.com/ai-artificial-intelligence/1002779/openai-dots-meta-muse-ai-agents-hardware-devices">The Verge AI</source></item><item><title>GPT-Synopsys: OpenAI und Chipdesign-Spezialist entwickeln Spezial-KI-Modell für Chip-Entwicklung</title><link>https://the-decoder.de/gpt-synopsys-openai-und-chipdesign-spezialist-entwickeln-spezial-ki-modell-fuer-chip-entwicklung/</link><guid>https://the-decoder.de/gpt-synopsys-openai-und-chipdesign-spezialist-entwickeln-spezial-ki-modell-fuer-chip-entwicklung/</guid><description>OpenAI und Synopsys entwickeln gemeinsam GPT-Synopsys, ein spezialisiertes KI-Modell für Chipdesign. Es soll die EDA-Werkzeuge von Synopsys wie ein erfahrener Ingenieur bedienen und Designs eigenständig optimieren. Erste</description><pubDate>Wed, 30 Sep 2026 19:08:30 +0000</pubDate><source url="https://the-decoder.de/gpt-synopsys-openai-und-chipdesign-spezialist-entwickeln-spezial-ki-modell-fuer-chip-entwicklung/">THE DECODER</source></item><item><title>OpenAI’s Jev clone could help the frontier lab stop its swarming agents</title><link>https://techcrunch.com/2026/09/30/openais-jev-clone-could-help-the-frontier-lab-stop-its-swarming-agents/</link><guid>https://techcrunch.com/2026/09/30/openais-jev-clone-could-help-the-frontier-lab-stop-its-swarming-agents/</guid><description>OpenAI&#x27;s &quot;Decisions API&quot; is a Jev clone that confirms the importance of fast, cheap intelligence.</description><pubDate>Wed, 30 Sep 2026 19:00:57 +0000</pubDate><source url="https://techcrunch.com/2026/09/30/openais-jev-clone-could-help-the-frontier-lab-stop-its-swarming-agents/">TechCrunch AI</source></item><item><title>AI voice startup ElevenLabs doubles valuation to $22B</title><link>https://techcrunch.com/2026/09/30/ai-voice-startup-elevenlabs-doubles-valuation-to-22b/</link><guid>https://techcrunch.com/2026/09/30/ai-voice-startup-elevenlabs-doubles-valuation-to-22b/</guid><description>The $300 million employee tender was co-led by Wellington and T. Rowe Price.</description><pubDate>Wed, 30 Sep 2026 18:23:57 +0000</pubDate><source url="https://techcrunch.com/2026/09/30/ai-voice-startup-elevenlabs-doubles-valuation-to-22b/">TechCrunch AI</source></item><item><title>Google ersetzt Gems durch Skills und folgt damit OpenAI und Anthropic beim neuen Prompt-Format</title><link>https://the-decoder.de/google-ersetzt-gems-durch-skills-und-folgt-damit-openai-und-anthropic-beim-neuen-prompt-format/</link><guid>https://the-decoder.de/google-ersetzt-gems-durch-skills-und-folgt-damit-openai-und-anthropic-beim-neuen-prompt-format/</guid><description>Google ersetzt Gems durch &quot;Skills&quot; im Gemini-Chat: detaillierte, wiederverwendbare Prompts, die Nutzer per &quot;/&quot; aufrufen oder die Gemini automatisch ausführt. Das Format geht auf einen offenen Standard von Anthropic zurüc</description><pubDate>Wed, 30 Sep 2026 18:00:04 +0000</pubDate><source url="https://the-decoder.de/google-ersetzt-gems-durch-skills-und-folgt-damit-openai-und-anthropic-beim-neuen-prompt-format/">THE DECODER</source></item><item><title>Here’s what AI leaders are saying about Trump’s new safety plan</title><link>https://www.theverge.com/ai-artificial-intelligence/1002636/ai-execs-trump-self-policing-deal-comments</link><guid>https://www.theverge.com/ai-artificial-intelligence/1002636/ai-execs-trump-self-policing-deal-comments</guid><description>After hosting a meal with Big Tech leaders on Tuesday, President Donald Trump responded to journalist questions about his artificial intelligence announcements in typical fashion. He said his previous claims regarding fe</description><pubDate>Wed, 30 Sep 2026 17:32:07 +0000</pubDate><source url="https://www.theverge.com/ai-artificial-intelligence/1002636/ai-execs-trump-self-policing-deal-comments">The Verge AI</source></item><item><title>Meta umgeht Milliarden an US-Steuern, indem es KI-Rechenzentren als Experimente deklariert</title><link>https://the-decoder.de/meta-umgeht-milliarden-an-us-steuern-indem-es-ki-rechenzentren-als-experimente-deklariert/</link><guid>https://the-decoder.de/meta-umgeht-milliarden-an-us-steuern-indem-es-ki-rechenzentren-als-experimente-deklariert/</guid><description>Meta stuft seine KI-Rechenzentren steuerlich als &quot;Pilotmodelle&quot; und Nvidia-Chips als Versuchsmaterialien ein, um Milliarden an US-Bundessteuern zu sparen. Allein 2025 waren es 3,9 Milliarden Dollar. Die genutzte Steuergu</description><pubDate>Wed, 30 Sep 2026 16:57:34 +0000</pubDate><source url="https://the-decoder.de/meta-umgeht-milliarden-an-us-steuern-indem-es-ki-rechenzentren-als-experimente-deklariert/">THE DECODER</source></item><item><title>All the latest news on Meta’s cute, creepy Muse AI agent</title><link>https://www.theverge.com/ai-artificial-intelligence/1002671/meta-muse-ai</link><guid>https://www.theverge.com/ai-artificial-intelligence/1002671/meta-muse-ai</guid><description>Meta launched a new Muse AI agent it claims can help you with everything from firing off emails to buying stuff online. Muse can be surprisingly effective at delivering on those promises — if you’re willing to trust Meta</description><pubDate>Wed, 30 Sep 2026 15:18:49 +0000</pubDate><source url="https://www.theverge.com/ai-artificial-intelligence/1002671/meta-muse-ai">The Verge AI</source></item><item><title>Google reportedly tests paying publishers for AI search results</title><link>https://www.theverge.com/tech/1002665/google-paying-publishers-ai-search-features</link><guid>https://www.theverge.com/tech/1002665/google-paying-publishers-ai-search-features</guid><description>Google has launched a pilot program that pays publishers for their contributions to its AI-powered search features, according to a report from The Information. The pilot program reportedly includes around 100 publishers </description><pubDate>Wed, 30 Sep 2026 14:55:21 +0000</pubDate><source url="https://www.theverge.com/tech/1002665/google-paying-publishers-ai-search-features">The Verge AI</source></item><item><title>Open TTS Leaderboard: Scalable Evaluation for Multilingual Text-to-Speech and Voice Cloning</title><link>https://huggingface.co/blog/open-tts-leaderboard</link><guid>https://huggingface.co/blog/open-tts-leaderboard</guid><description></description><pubDate>Wed, 30 Sep 2026 00:00:00 +0000</pubDate><source url="https://huggingface.co/blog/open-tts-leaderboard">Hugging Face Blog</source></item><item><title>NVIDIA Kumo Tabular Sets a New Accuracy-Efficiency Frontier for Tabular Prediction</title><link>https://huggingface.co/blog/nvidia/kumo-tabular</link><guid>https://huggingface.co/blog/nvidia/kumo-tabular</guid><description></description><pubDate>Tue, 29 Sep 2026 15:30:38 +0000</pubDate><source url="https://huggingface.co/blog/nvidia/kumo-tabular">Hugging Face Blog</source></item><item><title>Getting the Source Right, Not Just the Fact: Source-Aware Verification for MCP Agents</title><link>https://huggingface.co/blog/MultiverseComputingCAI/getting-the-source-right-not-just-the-fact-source</link><guid>https://huggingface.co/blog/MultiverseComputingCAI/getting-the-source-right-not-just-the-fact-source</guid><description></description><pubDate>Tue, 29 Sep 2026 13:07:00 +0000</pubDate><source url="https://huggingface.co/blog/MultiverseComputingCAI/getting-the-source-right-not-just-the-fact-source">Hugging Face Blog</source></item><item><title>MicroLLM Lab – Try 7 tiny LLM&#x27;s in the browser</title><link>https://stateofutopia.com/experiments/microllmlab/</link><guid>https://stateofutopia.com/experiments/microllmlab/</guid><description>Article URL: https://stateofutopia.com/experiments/microllmlab/ Comments URL: https://news.ycombinator.com/item?id=49882781 Points: 282 # Comments: 113</description><pubDate>Mon, 28 Sep 2026 18:58:53 +0000</pubDate><source url="https://stateofutopia.com/experiments/microllmlab/">Hacker News</source></item><item><title>Holo4: powering generalist computer-use agents</title><link>https://huggingface.co/blog/Hcompany/holo4</link><guid>https://huggingface.co/blog/Hcompany/holo4</guid><description></description><pubDate>Mon, 28 Sep 2026 09:44:05 +0000</pubDate><source url="https://huggingface.co/blog/Hcompany/holo4">Hugging Face Blog</source></item><item><title>&quot;As a Language Model&quot;: Chat Template Switches LLM Self-Referential Voice</title><link>https://arxiv.org/abs/2609.25021</link><guid>https://arxiv.org/abs/2609.25021</guid><description>Article URL: https://arxiv.org/abs/2609.25021 Comments URL: https://news.ycombinator.com/item?id=49865343 Points: 103 # Comments: 109</description><pubDate>Sun, 27 Sep 2026 10:26:25 +0000</pubDate><source url="https://arxiv.org/abs/2609.25021">Hacker News</source></item><item><title>Understanding the Impact of LLM Watermarking on AI Agent Behavior</title><link>https://www.lasso.security/blog/the-provenance-tax-understanding-the-impact-of-llm-watermarking-on-ai-agent-behavior</link><guid>https://www.lasso.security/blog/the-provenance-tax-understanding-the-impact-of-llm-watermarking-on-ai-agent-behavior</guid><description>Article URL: https://www.lasso.security/blog/the-provenance-tax-understanding-the-impact-of-llm-watermarking-on-ai-agent-behavior Comments URL: https://news.ycombinator.com/item?id=49856149 Points: 58 # Comments: 72</description><pubDate>Sat, 26 Sep 2026 13:05:36 +0000</pubDate><source url="https://www.lasso.security/blog/the-provenance-tax-understanding-the-impact-of-llm-watermarking-on-ai-agent-behavior">Hacker News</source></item><item><title>Generate fonts where every LLM token is the same width</title><link>https://ampdot.mesh.host/token-space-fonts.html</link><guid>https://ampdot.mesh.host/token-space-fonts.html</guid><description>Article URL: https://ampdot.mesh.host/token-space-fonts.html Comments URL: https://news.ycombinator.com/item?id=49851883 Points: 95 # Comments: 23</description><pubDate>Sat, 26 Sep 2026 00:30:03 +0000</pubDate><source url="https://ampdot.mesh.host/token-space-fonts.html">Hacker News</source></item><item><title>Best LLM for every budget, updated daily</title><link>https://bestmodelforyourbudget.terrydjony.com/</link><guid>https://bestmodelforyourbudget.terrydjony.com/</guid><description>Article URL: https://bestmodelforyourbudget.terrydjony.com/ Comments URL: https://news.ycombinator.com/item?id=49830866 Points: 184 # Comments: 113</description><pubDate>Thu, 24 Sep 2026 14:09:57 +0000</pubDate><source url="https://bestmodelforyourbudget.terrydjony.com/">Hacker News</source></item><item><title>Accelerating vision-language models with LFM2.5-VL-DSpark</title><link>https://huggingface.co/blog/LiquidAI/lfm2-5-vl-dspark</link><guid>https://huggingface.co/blog/LiquidAI/lfm2-5-vl-dspark</guid><description></description><pubDate>Thu, 24 Sep 2026 14:08:57 +0000</pubDate><source url="https://huggingface.co/blog/LiquidAI/lfm2-5-vl-dspark">Hugging Face Blog</source></item><item><title>Mercury 2.5 LLM hits 770 tokens per second</title><link>https://artificialanalysis.ai/models/mercury-2-5</link><guid>https://artificialanalysis.ai/models/mercury-2-5</guid><description>Article URL: https://artificialanalysis.ai/models/mercury-2-5 Comments URL: https://news.ycombinator.com/item?id=49823348 Points: 151 # Comments: 92</description><pubDate>Wed, 23 Sep 2026 22:16:19 +0000</pubDate><source url="https://artificialanalysis.ai/models/mercury-2-5">Hacker News</source></item><item><title>LLM Ass Bench</title><link>https://www.assbench.com/</link><guid>https://www.assbench.com/</guid><description>Article URL: https://www.assbench.com/ Comments URL: https://news.ycombinator.com/item?id=49807688 Points: 170 # Comments: 46</description><pubDate>Tue, 22 Sep 2026 20:34:19 +0000</pubDate><source url="https://www.assbench.com/">Hacker News</source></item><item><title>Jev introduces a new shape of LLM</title><link>https://simonwillison.net/2026/Sep/21/jev/</link><guid>https://simonwillison.net/2026/Sep/21/jev/</guid><description>Article URL: https://simonwillison.net/2026/Sep/21/jev/ Comments URL: https://news.ycombinator.com/item?id=49796843 Points: 53 # Comments: 20</description><pubDate>Tue, 22 Sep 2026 04:38:31 +0000</pubDate><source url="https://simonwillison.net/2026/Sep/21/jev/">Hacker News</source></item><item><title>How UK AISI and EvalEval Are Making Benchmark Results Reproducible</title><link>https://huggingface.co/blog/evaleval-aisi</link><guid>https://huggingface.co/blog/evaleval-aisi</guid><description></description><pubDate>Tue, 22 Sep 2026 00:00:00 +0000</pubDate><source url="https://huggingface.co/blog/evaleval-aisi">Hugging Face Blog</source></item><item><title>Transformers now runs llama.cpp quants</title><link>https://huggingface.co/blog/transformers-llama-cpp-quants</link><guid>https://huggingface.co/blog/transformers-llama-cpp-quants</guid><description></description><pubDate>Tue, 22 Sep 2026 00:00:00 +0000</pubDate><source url="https://huggingface.co/blog/transformers-llama-cpp-quants">Hugging Face Blog</source></item><item><title>Jun Kim, oMLX creator and maintainer, joins Hugging Face to support the MLX community</title><link>https://huggingface.co/blog/omlx</link><guid>https://huggingface.co/blog/omlx</guid><description></description><pubDate>Tue, 22 Sep 2026 00:00:00 +0000</pubDate><source url="https://huggingface.co/blog/omlx">Hugging Face Blog</source></item><item><title>tokenizers v1: encode, decode and scaling, measured</title><link>https://huggingface.co/blog/tokenizers-v1</link><guid>https://huggingface.co/blog/tokenizers-v1</guid><description></description><pubDate>Mon, 21 Sep 2026 00:00:00 +0000</pubDate><source url="https://huggingface.co/blog/tokenizers-v1">Hugging Face Blog</source></item><item><title>Pirate Face Rescues LLM Models from Deletion</title><link>https://pirateface.co/</link><guid>https://pirateface.co/</guid><description>Article URL: https://pirateface.co/ Comments URL: https://news.ycombinator.com/item?id=49776699 Points: 569 # Comments: 149</description><pubDate>Sun, 20 Sep 2026 15:16:07 +0000</pubDate><source url="https://pirateface.co/">Hacker News</source></item><item><title>The Implications of Linguistic Illegibility for LLM Security</title><link>https://arxiv.org/abs/2609.02852</link><guid>https://arxiv.org/abs/2609.02852</guid><description>Article URL: https://arxiv.org/abs/2609.02852 Comments URL: https://news.ycombinator.com/item?id=49758689 Points: 79 # Comments: 29</description><pubDate>Fri, 18 Sep 2026 19:00:06 +0000</pubDate><source url="https://arxiv.org/abs/2609.02852">Hacker News</source></item><item><title>How to Write with an LLM</title><link>https://sockpuppet.org/blog/2026/09/17/how-to-write-with-an-llm/</link><guid>https://sockpuppet.org/blog/2026/09/17/how-to-write-with-an-llm/</guid><description>Article URL: https://sockpuppet.org/blog/2026/09/17/how-to-write-with-an-llm/ Comments URL: https://news.ycombinator.com/item?id=49747070 Points: 769 # Comments: 420</description><pubDate>Thu, 17 Sep 2026 21:48:38 +0000</pubDate><source url="https://sockpuppet.org/blog/2026/09/17/how-to-write-with-an-llm/">Hacker News</source></item><item><title>LLM Classification Is Feature Engineering</title><link>https://minimallysufficient.com/posts/llm-classification-is-feature-extraction/</link><guid>https://minimallysufficient.com/posts/llm-classification-is-feature-extraction/</guid><description>Article URL: https://minimallysufficient.com/posts/llm-classification-is-feature-extraction/ Comments URL: https://news.ycombinator.com/item?id=49742437 Points: 116 # Comments: 25</description><pubDate>Thu, 17 Sep 2026 15:40:17 +0000</pubDate><source url="https://minimallysufficient.com/posts/llm-classification-is-feature-extraction/">Hacker News</source></item><item><title>Your Agent Aced the Task. Will It Do It Again?</title><link>https://huggingface.co/blog/ibm-research/altk-evolve-consistency</link><guid>https://huggingface.co/blog/ibm-research/altk-evolve-consistency</guid><description></description><pubDate>Tue, 15 Sep 2026 16:00:44 +0000</pubDate><source url="https://huggingface.co/blog/ibm-research/altk-evolve-consistency">Hugging Face Blog</source></item></channel></rss>