<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>The Em Dash</title><description>Real-world AI for exhausted professionals. Less theory, more button pushing.</description><link>https://blog.theemdash.ai/</link><language>en-us</language><item><title>Building Our Own Apps Made CVEs Our Problem, So I Tested Which AI Would Catch One</title><link>https://blog.theemdash.ai/p/vibe-coded-node-audit/</link><guid isPermaLink="true">https://blog.theemdash.ai/p/vibe-coded-node-audit/</guid><description>When the Oct 21 Node.js security release dropped, building our own apps made CVEs our problem. So I built a blind test to see which AI would catch one.</description><pubDate>Fri, 30 Oct 2026 16:00:00 GMT</pubDate></item><item><title>How a Team of Grok Bots Runs This Blog</title><link>https://blog.theemdash.ai/p/bot-team-overview/</link><guid isPermaLink="true">https://blog.theemdash.ai/p/bot-team-overview/</guid><description>A team of Grok Bot agents runs this blog. Who does what, how a post moves through it, what broke first, what it costs, and what I still haven&apos;t measured.</description><pubDate>Fri, 23 Oct 2026 16:00:00 GMT</pubDate></item><item><title>Claude Finally Enters the Personal Assistant Game with Duet</title><link>https://blog.theemdash.ai/p/claude-duet/</link><guid isPermaLink="true">https://blog.theemdash.ai/p/claude-duet/</guid><description>Anthropic opened a waitlisted research preview of Claude Duet on Oct 21, a proactive agent app for professionals and small business owners, starting with US Pro and Max subscribers.</description><pubDate>Thu, 22 Oct 2026 17:00:00 GMT</pubDate><category>Claude</category><category>AI agents</category></item><item><title>Claude Projects Is Back, and It&apos;s a Different Thing Now</title><link>https://blog.theemdash.ai/p/claude-projects-multi-agent/</link><guid isPermaLink="true">https://blog.theemdash.ai/p/claude-projects-multi-agent/</guid><description>Anthropic&apos;s new Claude Projects puts one coordinator in charge of parallel agent threads. Public beta on Pro and Max. Get on the waitlist!</description><pubDate>Tue, 20 Oct 2026 17:00:00 GMT</pubDate><category>Claude</category><category>Anthropic</category><category>AI agents</category></item><item><title>Notion&apos;s Agent Can Now Fill Out Forms on the Web. First, Check Whether You Get It.</title><link>https://blog.theemdash.ai/p/notion-agents-rebuilt/</link><guid isPermaLink="true">https://blog.theemdash.ai/p/notion-agents-rebuilt/</guid><description>Notion rebuilt its agents on Oct 19 with a browser that can fill and submit web forms. Plus gets a trial. Business gets the full agent and monthly credits.</description><pubDate>Tue, 20 Oct 2026 17:00:00 GMT</pubDate><category>Notion</category><category>AI agents</category></item><item><title>Deploying OpenClaw: The Basement Email Server of AI</title><link>https://blog.theemdash.ai/p/deploying-openclaw-the-basement-email-server-of-ai/</link><guid isPermaLink="true">https://blog.theemdash.ai/p/deploying-openclaw-the-basement-email-server-of-ai/</guid><description>I deployed OpenClaw on a Hostinger VPS. It runs like an email server in your basement: constant upkeep, silent failures, a restart loop. My verdict: not yet.</description><pubDate>Sat, 28 Mar 2026 18:00:00 GMT</pubDate></item><item><title>Building My First AI System: What Happens When Chat Apps Aren&apos;t Enough</title><link>https://blog.theemdash.ai/p/building-my-first-ai-system-what-happens-when-chat-apps-aren-t-enough/</link><guid isPermaLink="true">https://blog.theemdash.ai/p/building-my-first-ai-system-what-happens-when-chat-apps-aren-t-enough/</guid><description>Chat apps lost my context when each conversation ended and the MCP fixes kept failing. So I built a folder of markdown files that Claude Code reads first.</description><pubDate>Fri, 13 Mar 2026 16:00:00 GMT</pubDate></item><item><title>Your $240 Question: When Should You Actually Try the Hot New AI Tool?</title><link>https://blog.theemdash.ai/p/your-240-question-when-should-you-actually-try-the-hot-new-ai-tool/</link><guid isPermaLink="true">https://blog.theemdash.ai/p/your-240-question-when-should-you-actually-try-the-hot-new-ai-tool/</guid><description>I was ready to deploy OpenClaw until I asked what it does. A Reddit user burned $240 finding out. Five questions to ask before you try a viral AI tool.</description><pubDate>Mon, 02 Feb 2026 05:00:00 GMT</pubDate></item><item><title>Local LLM Setup Was Step One. Finding the Right Model Was Step Two.</title><link>https://blog.theemdash.ai/p/local-llm-setup-was-step-one-finding-the-right-model-was-step-two/</link><guid isPermaLink="true">https://blog.theemdash.ai/p/local-llm-setup-was-step-one-finding-the-right-model-was-step-two/</guid><description>I ran llama-bench on Mistral-Small-3.2-24B across four backends on an HP Z2 Mini (Strix Halo). ROCm 6.4.4 won prompt processing at 400.74 tokens per second.</description><pubDate>Wed, 28 Jan 2026 05:00:00 GMT</pubDate></item><item><title>I Built an Unlimited AI Workspace (And Why You Might Want One)</title><link>https://blog.theemdash.ai/p/i-built-an-unlimited-ai-workspace-and-why-you-might-want-one/</link><guid isPermaLink="true">https://blog.theemdash.ai/p/i-built-an-unlimited-ai-workspace-and-why-you-might-want-one/</guid><description>I put a local LLM on an HP Z2 Mini G1a (AMD Strix Halo) to iterate without rate limits. What I weighed, the stack, three gotchas, and whether it was worth it.</description><pubDate>Thu, 15 Jan 2026 05:00:00 GMT</pubDate></item><item><title>I Have Nothing to Show: Three Months of AI Work With No Artifacts</title><link>https://blog.theemdash.ai/p/i-have-nothing-to-show-three-months-of-ai-work-with-no-artifacts/</link><guid isPermaLink="true">https://blog.theemdash.ai/p/i-have-nothing-to-show-three-months-of-ai-work-with-no-artifacts/</guid><description>Three months of daily AI work left me nothing to publish. My new rule: every session ends in a decision, a learning, a template or an admitted nothing.</description><pubDate>Thu, 08 Jan 2026 05:00:00 GMT</pubDate></item></channel></rss>