<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
        <title>Tokens in the Stream - Why LLMs are inherently insecure and prompt injection will persist</title>
        <link>https://tube.yeet.fm/videos/watch/8a108b17-4f49-4d70-ae82-5a2af0cfbe08</link>
        <description>Turns out, every chatbot conversation runs on a messy hack at the heart of language models, making prompt injection an unsolved—and possibly unsolvable—security threat. Steve and Leo unravel the research that explains why “roles” in AI aren’t what you think they are. • Understanding the controversy surrounding “AI Model Distillation” • Anthropic moves to make their most powerful Mythos 5 model more widely available. • Bitwarden’s “Secrets Manager” offering prevents agentic and prompt injection abuse. • The astounding and disturbing truth about the way conversation AI actually works Show Notes - https://www.grc.com/sn/SN-1093-Notes.pdf Hosts: Steve Gibson and Leo Laporte Security Now episode 1093 More Info: https://twit.tv/shows/security-now/episodes/1093 Sponsors: • http://threatlocker.com/twit • http://box.com/AI • http://hoxhunt.com/securitynow • http://guardsquare.com • http://material.security Download or subscribe to Security Now at https://twit.tv/shows/security-now You can submit a question to Security Now at the GRC Feedback Page. For 16kbps versions, transcripts, and notes (including fixes), visit Steve's site: https://grc.com/ also the home of the best disk maintenance and recovery utility ever written Spinrite 6. Join Club TWiT for Ad-Free Podcasts! Support what you love and get ad-free audio and video feeds, a members-only Discord, and exclusive content. Join today: https://twit.tv/clubtwit Products we recommend: https://www.amazon.com/shop/twitnetcastnetwork TWiT may earn commissions on certain products. Follow us: Website: https://twit.tv/ Bluesky: https://bsky.app/profile/twit.tv Mastodon: https://mastodon.social/@TWiT Facebook: https://www.facebook.com/TWiTNetwork Instagram: https://www.instagram.com/twit.tv/ X (Twitter): https://x.com/TWiT TikTok: https://www.tiktok.com/@twittok Reddit: https://www.reddit.com/r/thisweekintech/ LinkedIn: https://www.linkedin.com/company/twit-llc About us: TWiT.tv is a technology podcasting network located in the San Francisco Bay Area with the #1 ranked technology podcast This Week in Tech hosted by Leo Laporte. Every week we produce dozens of hours of content on a variety of programs including Tech News Weekly, MacBreak Weekly, Windows Weekly, Security Now, Intelligent Machines, and more.</description>
        <lastBuildDate>Sat, 10 Oct 2026 03:24:42 GMT</lastBuildDate>
        <docs>https://validator.w3.org/feed/docs/rss2.html</docs>
        <generator>PeerTube - https://tube.yeet.fm</generator>
        <image>
            <title>Tokens in the Stream - Why LLMs are inherently insecure and prompt injection will persist</title>
            <url>https://tube.yeet.fm/client/assets/images/icons/icon-96x96.png</url>
            <link>https://tube.yeet.fm/videos/watch/8a108b17-4f49-4d70-ae82-5a2af0cfbe08</link>
        </image>
        <copyright>All rights reserved, unless otherwise specified in the terms specified at https://tube.yeet.fm/about and potential licenses granted by each content's rightholder.</copyright>
        <atom:link href="https://tube.yeet.fm/feeds/video-comments.xml?videoId=8a108b17-4f49-4d70-ae82-5a2af0cfbe08" rel="self" type="application/rss+xml"/>
    </channel>
</rss>