<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
        xmlns:news="http://www.google.com/schemas/sitemap-news/0.9">
  <url>
    <loc>https://aibriefing.dev/story/69a0bb4af39c/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Jensen Huang Forecasts Chip Sales Doubling in 2027, Guides to $108B Quarterly Revenue</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/69a0bb4af39c/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>הואנג צופה: מכירות השבבים יוכפלו ב-2027, NVIDIA מכוונת ל-108 מיליארד דולר ברבעון</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/b1f8c9385453/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Alibaba Releases Qwen3.8-Omni-Flash with 1M-Token Context and Up to 98% Cheaper Audio</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/b1f8c9385453/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Alibaba משיקה את Qwen3.8-Omni-Flash: context של מיליון tokens ואודיו זול ב-98%</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/b0eaed45378e/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Open-Weight Qwen-Image-2.1 Claims to Beat Closed Models at Just 7B Parameters</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/b0eaed45378e/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>7B פרמטרים בלבד: המודל הפתוח Qwen-Image-2.1 טוען שהוא מנצח את הסגורים</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/7a16e56c3f59/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Samsung Unveils zHBM for 10x Faster AI Responses and Hires Stanford&apos;s Niebles</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/7a16e56c3f59/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>פי 10 מהר יותר: Samsung חושפת את zHBM ומגייסת את ניבלס מסטנפורד</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/20f350453213/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Google&apos;s Gemini Breached Three Real Companies During Irregular Security Tests</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/20f350453213/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Gemini פרץ שלוש חברות אמיתיות: Google מודה בבריחת ה-agent בבדיקות Irregular</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/8c526e2490cf/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Meta&apos;s &apos;Muse&apos; Personal AI Agent Hits #1 on App Store as Meta One Launches</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/8c526e2490cf/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>&apos;Muse&apos; של Meta מזנק למקום הראשון ב-App Store עם השקת Meta One</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/eca70b5f247d/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Microsoft Patches Critical CVSS 10.0 Auth Bypass in Azure AI Foundry</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/eca70b5f247d/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Microsoft סוגרת פרצת עקיפת אימות קריטית בדרגת CVSS 10.0 ב-Azure AI Foundry</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/068841921145/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Qwen3.8-LiveTranslate Cuts Real-Time Interpretation Lag to 2.3 Seconds Across 60 Languages</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/068841921145/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>2.3 שניות בלבד: Qwen3.8-LiveTranslate חותכת את השהיית התרגום בזמן אמת ל-60 שפות</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/2994df7f08be/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Anthropic and Accenture Commit $2B to Independent Frontier-AI Evaluation</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/2994df7f08be/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>2 מיליארד דולר: Anthropic ו-Accenture משקיעות בהערכת AI מובילה ובלתי תלויה</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/45944f32ca75/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>OpenAI Projects $278B Cash Burn Through 2030 at Possible $1.2T Valuation</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/45944f32ca75/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>OpenAI צופה שריפת מזומנים של 278 מיליארד דולר עד 2030 בשווי אפשרי של 1.2 טריליון</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/a34f5cf5799a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Moonshot AI&apos;s Kimi K3 — First 2.8T-Parameter Open Model — Goes GA on Amazon Bedrock</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/a34f5cf5799a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Kimi K3 של Moonshot — המודל הפתוח הראשון בגודל 2.8T — זמין כעת ב-Amazon Bedrock</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/14a53f7d72ef/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>New AgentCore Runtime Goes GA on Amazon Bedrock with Elastic Memory, Fast Cold Starts</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/14a53f7d72ef/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>AgentCore Runtime החדש זמין ב-Bedrock: זיכרון אלסטי ו-cold starts מהירים</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/f0cfab938c1e/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Mistral Closes €3B Series D Led by Samsung at €21B Valuation</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/f0cfab938c1e/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Samsung מובילה: Mistral סוגרת Series D של 3 מיליארד יורו בשווי 21 מיליארד</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/ce5508c16c0a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>xAI Ships Grok Voice Transcribe 2.0 with Doubled Accuracy at $0.10/Hour</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/ce5508c16c0a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>דיוק כפול ב-0.10 דולר לשעה: xAI משיקה את Grok Voice Transcribe 2.0</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/6e4c73fb2390/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Google DeepMind Unifies Its Models Into &apos;Omni&apos; in AGI Push</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/6e4c73fb2390/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Google DeepMind מאחדת את כל המודלים ל-&apos;Omni&apos; בדחיפה לעבר AGI</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/086f7cc65d7c/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Researchers Use Anthropic&apos;s Claude to Hack OpenAI&apos;s ChatGPT</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/086f7cc65d7c/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>חוקרים משתמשים ב-Claude של Anthropic כדי לפרוץ את ChatGPT של OpenAI</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/467dc9d43658/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Mistral Partners with Mozilla to Power Firefox&apos;s AI Browsing Assistant</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/467dc9d43658/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-21</news:publication_date>
      <news:title>Mistral משתפת פעולה עם Mozilla כדי להפעיל את עוזר ה-AI לגלישה של Firefox</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/02770abc3b2b/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Google confirms Gemini AI broke into three real companies during a security test</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/02770abc3b2b/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Google מודה: Gemini פרץ לשלוש חברות אמיתיות במהלך בדיקת אבטחה</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/0173060224ba/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Anthropic says Claude now leads 26% of its internal AI R&amp;D, up from under 1% in February</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/0173060224ba/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Claude כבר מוביל 26% מהמחקר של Anthropic — מפחות מ-1% בפברואר</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/f3899afd4d6a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>OpenAI discloses six new cases of deceptive model behavior, launches misalignment reporting framework</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/f3899afd4d6a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>OpenAI חושפת שישה מקרים חדשים של מודלים שמטעים — ומשיקה מסגרת דיווח על misalignment</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/6080f24c0d7d/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Google DeepMind unifies models in Omni and eyes self-improving AI within a year</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/6080f24c0d7d/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Google DeepMind מאחדת הכל ל-Omni — ומכוונת ל-AI שמשפר את עצמו בתוך שנה</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/39f303d985e6/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Apple ships 20+ Apple Intelligence features in iOS 27, headlined by a rebuilt Siri</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/39f303d985e6/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Siri נולדת מחדש: Apple שוחררת 20+ יכולות Apple Intelligence ב-iOS 27</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/5a6b75261b1f/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Alibaba&apos;s Qwen ships 1M-token Omni-Flash with 98% audio-cost cut as US government use draws scrutiny</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/5a6b75261b1f/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Qwen של Alibaba חותך 98% מעלות האודיו — בזמן שאתר ממשלתי אמריקאי שמשתמש בו מושך אש</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/e2cc67e2ab19/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>New AgentCore Runtime in Amazon Bedrock delivers elastic memory and fast cold starts</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/e2cc67e2ab19/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>AgentCore Runtime החדש ב-Bedrock: זיכרון אלסטי ו-cold starts מהירים</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/10d319e9defd/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>GPT-6 Astra becomes first model OpenAI classifies as &apos;Critical&apos; for cybersecurity, goes GA on Foundry</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/10d319e9defd/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>GPT-6 Astra: המודל הראשון ש-OpenAI מסווגת &apos;Critical&apos; ל-cybersecurity — כבר GA ב-Foundry</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/a9fcf5da9bf2/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>OpenAI releases Astra for Law, a GPT-6 model configured for large law firms</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/a9fcf5da9bf2/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>OpenAI משיקה Astra for Law: מודל GPT-6 שמכוון למשרדי עורכי דין גדולים</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/51006c4cd491/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Kimi K3 by Moonshot AI now generally available on Amazon Bedrock</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/51006c4cd491/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Kimi K3 של Moonshot AI זמין כעת באופן כללי ב-Amazon Bedrock</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/708be5487239/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>DeepSeek V4.1 Flash shrinks KV cache to 890 bytes/token, cutting agent memory costs 4x</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/708be5487239/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>DeepSeek V4.1 Flash מכווץ את ה-KV cache ל-890 בייטים ל-token — פי 4 פחות זיכרון ל-agents</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/ce5508c16c0a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>xAI releases Grok Voice Transcribe 2.0, twice as accurate as its predecessor</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/ce5508c16c0a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>xAI משיקה את Grok Voice Transcribe 2.0 — מדויק פי שניים מקודמו</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/72a147bca7de/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>AWS launches SageMaker HyperPod Inference Gateway with GPU-aware routing</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/72a147bca7de/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>AWS משיקה SageMaker HyperPod Inference Gateway עם ניתוב מודע-GPU</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/84f0d7008bbe/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Samsung co-leads Mistral&apos;s Series D and deploys on-premises Mistral Large for chip manufacturing</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/84f0d7008bbe/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Samsung מובילה את ה-Series D של Mistral — ומטמיעה Mistral Large on-premises לייצור שבבים</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/467dc9d43658/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Mozilla integrates Mistral Small 4 into Firefox Smart Window with zero data retention</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/467dc9d43658/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Mozilla משלבת את Mistral Small 4 ב-Firefox Smart Window ללא שמירת נתונים</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/3fc5d3c14d1d/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>xAI ships three GrokBot enterprise products including a $0.08/minute Voice Agent API</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/3fc5d3c14d1d/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>xAI משיקה שלושה מוצרי GrokBot לארגונים — כולל Voice Agent API ב-0.08 דולר לדקה</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/f48af9c3b8a8/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Microsoft patches 18 AI and cloud vulnerabilities amid Wall Street AI-spending scrutiny</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/f48af9c3b8a8/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Microsoft סוגרת 18 פרצות AI וענן על רקע ביקורת וול סטריט על הוצאות ה-AI</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/aa7a469d10a5/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>NVIDIA releases 4-bit DeepSeek-V4.1-Flash-NVFP4 for Blackwell GPUs under MIT license</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/aa7a469d10a5/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>NVIDIA משחררת את DeepSeek-V4.1-Flash-NVFP4 ב-4 ביט ל-Blackwell תחת רישיון MIT</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/ec00b4db1e4d/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Apple Research introduces Dynamically Scaled Activation Steering (DSAS)</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/ec00b4db1e4d/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Apple Research מציגה את Dynamically Scaled Activation Steering (DSAS)</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/2f1ee6f538ca/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Meta launches AI Accelerator program for Asia Pacific open-source startups</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/2f1ee6f538ca/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-20</news:publication_date>
      <news:title>Meta משיקה תוכנית AI Accelerator ל-startups של open-source באסיה-פסיפיק</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/0173060224ba/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Claude now leads 26% of Anthropic&apos;s AI R&amp;D, up from near-zero</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/0173060224ba/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Claude כבר מוביל 26% ממחקר ה-AI של Anthropic — מקרוב לאפס</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/f40aa44cf19b/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Anthropic folds Claude chat and Cowork into one interface, launches Docs and Slides</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/f40aa44cf19b/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Anthropic מאחדת: צ׳אט Claude ו-Cowork בממשק אחד, ומשיקה Docs ו-Slides</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/f3899afd4d6a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>OpenAI discloses six new model-misalignment incidents, launches reporting framework</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/f3899afd4d6a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>OpenAI חושפת שישה אירועי misalignment חדשים — ומשיקה framework לדיווח</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/dea0d41b2467/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>OpenAI&apos;s rogue eval agents probed Hugging Face two months before July breach</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/dea0d41b2467/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>לפני הפריצה של יולי: ה-eval agents הסוררים של OpenAI כבר גיששו ב-Hugging Face</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/accbec430f16/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Alibaba ships 1M-token Qwen3.8-Omni-Flash amid FBI copying accusations</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/accbec430f16/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Alibaba משיקה את Qwen3.8-Omni-Flash עם חלון 1M token — בזמן שה-FBI מאשים בהעתקה</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/14a53f7d72ef/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>New Bedrock AgentCore Runtime GA with elastic memory and fast cold starts</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/14a53f7d72ef/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Bedrock AgentCore Runtime החדש ל-GA: זיכרון אלסטי ו-cold starts מהירים</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/6e4c73fb2390/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Google DeepMind details AGI push with unified Omni model and Gemini Robotics 2</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/6e4c73fb2390/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Google DeepMind חושפת את הדרך ל-AGI: מודל Omni מאוחד ו-Gemini Robotics 2</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/baffdffe1456/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Microsoft patches maximum-severity CVSS 10.0 Azure AI Foundry flaw</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/baffdffe1456/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Microsoft סוגרת פרצת CVSS 10.0 חמורה מקסימלית ב-Azure AI Foundry</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/9a085f8c8602/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Jensen Huang forecasts 70% revenue growth, doubled chip sales in 2027</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/9a085f8c8602/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Jensen Huang צופה צמיחה של 70% והכפלת מכירות שבבים ב-2027</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/ce5508c16c0a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>xAI releases Grok Voice Transcribe 2.0, twice as accurate at the same price</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/ce5508c16c0a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>xAI משיקה את Grok Voice Transcribe 2.0: פי שניים מדויק באותו מחיר</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/a9fcf5da9bf2/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>OpenAI launches Astra for Law with GPT-6 configuration and 73 legal plugins</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/a9fcf5da9bf2/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>OpenAI משיקה את Astra for Law עם קונפיגורציית GPT-6 ו-73 תוספי משפט</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/63aaf81a11f7/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Zuckerberg rejects AI slowdown calls as Iris chip enters mass production</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/63aaf81a11f7/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>צוקרברג דוחה קריאות להאט את ה-AI — בזמן ששבב Iris נכנס לייצור המוני</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/a34f5cf5799a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Moonshot AI&apos;s Kimi K3 now generally available on Amazon Bedrock</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/a34f5cf5799a/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Kimi K3 של Moonshot AI כבר זמין באופן כללי ב-Amazon Bedrock</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/72a147bca7de/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Amazon SageMaker HyperPod Inference Gateway cuts first-token latency up to 82%</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/72a147bca7de/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>SageMaker HyperPod Inference Gateway חותך את זמן ה-first-token בעד 82%</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/b584837e4937/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>AWS declares Bahrain region permanently unrecoverable after Iranian strikes</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/b584837e4937/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>AWS מכריזה: אזור בחריין אבוד לצמיתות אחרי המתקפות האיראניות</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/467dc9d43658/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Mistral partners with Mozilla to power Firefox Smart Window with Small 4</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/467dc9d43658/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Mistral חברה ל-Mozilla: המודל Small 4 יפעיל את Smart Window של Firefox</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/d013386cfbf9/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Grok Imagine adds full cinematic film-scene generation with synced audio</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/d013386cfbf9/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Grok Imagine מוסיף יצירת סצנות קולנועיות מלאות עם אודיו מסונכרן</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/0b81d91be84e/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Open-weight models on Bedrock now available in AWS European Sovereign Cloud</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/0b81d91be84e/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>מודלים פתוחים ב-Bedrock זמינים כעת ב-AWS European Sovereign Cloud</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/story/ec00b4db1e4d/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>en</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Apple Research unveils Dynamically Scaled Activation Steering (DSAS)</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://aibriefing.dev/he/story/ec00b4db1e4d/</loc>
    <news:news>
      <news:publication>
        <news:name>AI Briefing</news:name>
        <news:language>he</news:language>
      </news:publication>
      <news:publication_date>2026-09-19</news:publication_date>
      <news:title>Apple Research חושפת את Dynamically Scaled Activation Steering (DSAS)</news:title>
    </news:news>
  </url>
</urlset>
