<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>yonifra.com — טכנולוגיה ובינה מלאכותית</title>
    <link>https://yonifra.com/he</link>
    <description>מאמרים בעברית על טכנולוגיה, בינה מלאכותית, אבטחת AI ותכנות מאת יוני פריימוריץ.</description>
    <language>he</language>
    <lastBuildDate>Thu, 17 Sep 2026 00:00:00 GMT</lastBuildDate>
    <atom:link href="https://yonifra.com/he/rss.xml" rel="self" type="application/rss+xml"/>
    <item>
      <title>שש תקריות חוסר ההתאמה של OpenAI: מה מפתחי סוכנים צריכים ללמוד</title>
      <link>https://yonifra.com/he/post/09-openai-six-misalignment-incidents</link>
      <guid isPermaLink="true">https://yonifra.com/he/post/09-openai-six-misalignment-incidents</guid>
      <pubDate>Thu, 17 Sep 2026 00:00:00 GMT</pubDate>
      <description>OpenAI חשפה סוכנים שהסתירו טעויות, השתמשו במפתחות שדלפו ועקפו מגבלות תקשורת. כך שש התקריות משפיעות על מערכות אמיתיות.</description>
    </item>
    <item>
      <title>חוק FRONTIER: מה מבקרי AI עצמאיים חייבים לבדוק</title>
      <link>https://yonifra.com/he/post/07-frontier-act-independent-ai-audits</link>
      <guid isPermaLink="true">https://yonifra.com/he/post/07-frontier-act-independent-ai-audits</guid>
      <pubDate>Wed, 16 Sep 2026 00:00:00 GMT</pubDate>
      <description>OpenAI תומכת בפיקוח עצמאי מחייב על AI. אילו בדיקות נדרשות כדי לזהות הסקה מוטה ופעילות מסוכנת של קבוצות סוכנים, ומה גם ביקורת כזו לא תוכל להוכיח?</description>
    </item>
    <item>
      <title>איך מודלי AI עובדים — ואיך אפשר להטעות אותם</title>
      <link>https://yonifra.com/he/post/08-how-ai-models-work</link>
      <guid isPermaLink="true">https://yonifra.com/he/post/08-how-ai-models-work</guid>
      <pubDate>Wed, 16 Sep 2026 00:00:00 GMT</pubDate>
      <description>מדריך חזותי לטוקנים, אימון, מנגנון הקשב וסוכני AI, עם הסבר מעשי על הזרקת הנחיות, עקיפת מגבלות בטיחות והרעלת נתונים.</description>
    </item>
    <item>
      <title>המודל אמר שזו סימולציה. ואז הוא פרסם נוזקה ב-PyPI.</title>
      <link>https://yonifra.com/he/post/05-mythos-5-simulation-biased-reasoning</link>
      <guid isPermaLink="true">https://yonifra.com/he/post/05-mythos-5-simulation-biased-reasoning</guid>
      <pubDate>Tue, 15 Sep 2026 00:00:00 GMT</pubDate>
      <description>דוח הבטיחות של Anthropic על Claude Mythos 5 מראה מה קורה כשמערכת הניטור סומכת על ההסברים של המודל עצמו.</description>
    </item>
    <item>
      <title>תקרית Hugging Face: כשסוכני AI ניסו לרמות במבחן והגיעו למערכות ייצור</title>
      <link>https://yonifra.com/he/post/06-hugging-face-incident-ai-agents</link>
      <guid isPermaLink="true">https://yonifra.com/he/post/06-hugging-face-incident-ai-agents</guid>
      <pubDate>Tue, 15 Sep 2026 00:00:00 GMT</pubDate>
      <description>סוכני OpenAI הפכו שירות חבילות משותף ללוח הודעות, יצאו מרשת הבדיקה והגיעו למערכות הייצור של Hugging Face.</description>
    </item>
    <item>
      <title>שיטת ראלף ויגאם: לפתח בעזרת AI, שוב ושוב, עד שזה עובד</title>
      <link>https://yonifra.com/he/post/03-ralph-wiggum-method</link>
      <guid isPermaLink="true">https://yonifra.com/he/post/03-ralph-wiggum-method</guid>
      <pubDate>Wed, 07 Jan 2026 00:00:00 GMT</pubDate>
      <description>לא חייבים להגיע לפתרון מושלם בניסיון הראשון. כך משתמשים בלולאות של ראלף ויגאם כדי לתת ל-Claude לעבוד באופן עצמאי עד להשלמת המשימה.</description>
    </item>
    <item>
      <title>שורת פקודה או סביבת פיתוח? העתיד של פיתוח בעזרת סוכני AI</title>
      <link>https://yonifra.com/he/post/04-cli-vs-ide-ai-coding</link>
      <guid isPermaLink="true">https://yonifra.com/he/post/04-cli-vs-ide-ai-coding</guid>
      <pubDate>Wed, 07 Jan 2026 00:00:00 GMT</pubDate>
      <description>האם הטרמינל חוזר למרכז הבמה? השוואה בין כלי AI בתוך סביבת הפיתוח לבין סוכנים שפועלים משורת הפקודה, כמו Claude Code.</description>
    </item>
    <item>
      <title>פרדוקס הבינה המלאכותית: כותבים פחות קוד, אבל מבינים יותר?</title>
      <link>https://yonifra.com/he/post/02-ai-paradox</link>
      <guid isPermaLink="true">https://yonifra.com/he/post/02-ai-paradox</guid>
      <pubDate>Tue, 06 Jan 2026 00:00:00 GMT</pubDate>
      <description>האם הבינה המלאכותית הופכת אותנו למפתחים טובים יותר, או רק מקלה עלינו להתאמץ פחות? על השפעת מודלי השפה על מפתחים בתחילת הדרך ועל מפתחים מנוסים.</description>
    </item>
    <item>
      <title>עידן חדש ליזמים</title>
      <link>https://yonifra.com/he/post/01-new-era-for-entrepreneurs</link>
      <guid isPermaLink="true">https://yonifra.com/he/post/01-new-era-for-entrepreneurs</guid>
      <pubDate>Fri, 15 Aug 2025 00:00:00 GMT</pubDate>
      <description>על השינויים הטכנולוגיים שפותחים הזדמנויות חדשות ליזמים ומשנים את הדרך שבה הופכים רעיון למציאות.</description>
    </item>
    <item>
      <title>איך לנהל פרויקטים אישיים לצד העבודה והמשפחה</title>
      <link>https://yonifra.com/he/post/00-side-projects</link>
      <guid isPermaLink="true">https://yonifra.com/he/post/00-side-projects</guid>
      <pubDate>Mon, 02 Oct 2023 00:00:00 GMT</pubDate>
      <description>לא פשוט לעבוד על פרויקטים אישיים כשמגדלים שלושה ילדים ועובדים במשרה מלאה, אבל זה בהחלט אפשרי.</description>
    </item>
  </channel>
</rss>
