<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Llm Security on MacWorks</title><link>https://macworks.dev/tags/llm-security/</link><description>Recent content in Llm Security on MacWorks</description><generator>Hugo</generator><language>en</language><atom:link href="https://macworks.dev/tags/llm-security/index.xml" rel="self" type="application/rss+xml"/><item><title>AI@X</title><link>https://macworks.dev/docs/week/ai@x/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://macworks.dev/docs/week/ai@x/</guid><description>&lt;h1 id="aix--week-of-2026-06-20-to-2026-06-26"&gt;AI@X — Week of 2026-06-20 to 2026-06-26&lt;a class="anchor" href="#aix--week-of-2026-06-20-to-2026-06-26"&gt;#&lt;/a&gt;&lt;/h1&gt;
&lt;h2 id="the-buzz"&gt;The Buzz&lt;a class="anchor" href="#the-buzz"&gt;#&lt;/a&gt;&lt;/h2&gt;
&lt;p&gt;The U.S. government is effectively attempting to nationalize and heavily regulate frontier models, clashing violently with an emerging enterprise reality where cheap, hyper-capable open-weights models are commoditizing intelligence. The Trump administration&amp;rsquo;s unprecedented mandate to stagger OpenAI&amp;rsquo;s GPT-5.6 release on a customer-by-customer basis marks a massive shift toward state-controlled AI. Simultaneously, the realization that Chinese open models like Zhipu&amp;rsquo;s GLM-5.2 can match frontier capabilities at a fraction of the cost is rapidly dismantling the trillion-dollar &amp;ldquo;compute moat&amp;rdquo; narrative that has driven recent hyperscaler valuations.&lt;/p&gt;</description></item><item><title>2026-06-22</title><link>https://macworks.dev/docs/archives/ai@x/x-2026-06-22/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://macworks.dev/docs/archives/ai@x/x-2026-06-22/</guid><description>&lt;details&gt;
&lt;summary&gt;Sources&lt;/summary&gt;
&lt;div class="markdown-inner"&gt;
&lt;ul&gt;

&lt;li&gt;&lt;a href="https://twitter.macworks.dev/levie/rss"&gt;Aaron Levie / @levie&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;&lt;a href="https://twitter.macworks.dev/karpathy/rss"&gt;Andrej Karpathy / @karpathy&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;&lt;a href="https://twitter.macworks.dev/AndrewYNg/rss"&gt;Andrew Ng / @AndrewYNg&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;&lt;a href="https://twitter.macworks.dev/AravSrinivas/rss"&gt;Aravind Srinivas / @AravSrinivas&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;&lt;a href="https://twitter.macworks.dev/awnihannun/rss"&gt;Awni Hannun / @awnihannun&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;&lt;a href="https://twitter.macworks.dev/drfeifei/rss"&gt;Fei-Fei Li / @drfeifei&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;&lt;a href="https://twitter.macworks.dev/GaryMarcus/rss"&gt;Gary Marcus / @GaryMarcus&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;&lt;a href="https://twitter.macworks.dev/sama/rss"&gt;Sam Altman / @sama&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;&lt;a href="https://twitter.macworks.dev/Steve_Yegge/rss"&gt;Steve Yegge / @Steve_Yegge&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;&lt;a href="https://twitter.macworks.dev/trq212/rss"&gt;Thariq / @trq212&lt;/a&gt;&lt;/li&gt;

&lt;li&gt;&lt;a href="https://twitter.macworks.dev/ylecun/rss"&gt;Yann LeCun / @ylecun&lt;/a&gt;&lt;/li&gt;

&lt;/ul&gt;
&lt;/div&gt;
&lt;/details&gt;


&lt;h1 id="ai-infrastructure-reality-checks-and-the-rise-of-multi-agent-orchestration--2026-06-22"&gt;AI Infrastructure Reality Checks and the Rise of Multi-Agent Orchestration — 2026-06-22&lt;a class="anchor" href="#ai-infrastructure-reality-checks-and-the-rise-of-multi-agent-orchestration--2026-06-22"&gt;#&lt;/a&gt;&lt;/h1&gt;
&lt;h2 id="highlights"&gt;Highlights&lt;a class="anchor" href="#highlights"&gt;#&lt;/a&gt;&lt;/h2&gt;
&lt;p&gt;The AI community today is sharply divided between the tangible, highly profitable enterprise gains of applied AI and the looming reality of an infrastructure bubble. While companies like Adobe and Box are proving that AI drives massive engagement and record revenues for incumbent software, skeptics are loudly warning that the trillion-dollar data center buildout simply does not align with current enterprise demand. Meanwhile, the launch of multi-agent APIs like Sakana&amp;rsquo;s Fugu signals a critical architectural shift toward routing tasks to specialized experts rather than relying on massive, monolithic models.&lt;/p&gt;</description></item></channel></rss>