AI News Today | Julian Goldie Podcast

Ultimate Guide to Hermes AI Agent Operating System

Learn how to set up and optimize the Hermes Agent Operating System to manage all your AI agents from one central mission control. This guide covers everything from SEO content pipelines and cloud hosting to using Obsidian as a permanent memory layer for smarter AI.

00:00 - Intro: AI Mission Control

00:53 - Local Setup & Security

02:38 - Running Agent OS in the Cloud

03:19 - AI SEO & Backlink Automation

05:12 - Choosing the Best AI Model

07:02 - Hermes vs. Anti-Gravity

08:02 - Obsidian: The Agent Memory Layer

13:27 - Syncing ChatGPT Data

OpenAI Codex Goal Mode: How to Automate Everything

OpenAI just dropped massive updates for Codex, including the game-changing Goal Mode for autonomous tasks. Learn how to integrate these updates into an Agent Operating System to manage all your AI agents and workflows in one place.

00:00 - Intro

01:28 - 4 Major Codex Updates

02:06 - Why You Need Goal Mode

03:11 - Codex Chat vs Goal Mode

04:16 - Managing Goal History

05:26 - Past Sessions & Templates

06:10 - Previewing Your Workspace

07:52 - Is Codex Only for Developers?

09:11 - Summary & Community Offer

Google Anti-gravity 2.0: Build a Powerful AI Agent OS

Learn how to transform Google Anti-gravity 2.0 from a simple chat box into a high-performance Agent Operating System with permanent memory. This guide breaks down the Goldie Gravity Grid framework to create a self-improving AI flywheel for your business.

00:00 - 00:00 - Intro: Anti-gravity + Agent OS

00:44 - How Anti-gravity 2.0 Works

01:58 - Customizing the Agent Dashboard

03:01 - Gemini 3.5 Flash & Agent Features

05:10 - The Goldie Gravity Grid Framework

07:08 - The Anti-gravity CLI Explained

10:11 - The AI Flywheel & Memory Loop

11:56 - Accessing the Agent OS Files

Get Claude Code for FREE: Full Setup Guide (Agent OS)

Discover how to access the power of Claude Code completely for free using open-source tools and free APIs on Open Router. Learn to build a professional-grade Agent Operating System featuring live code previews, persistent Obsidian memory, and managed agent fleets.

00:00 - Intro

00:37 - How Free Claude Code Works

01:31 - Live Workspace & Preview Features

02:36 - Using Free APIs via Open Router

04:30 - Building a Persistent Memory System

05:31 - Managing an AI Agent Fleet

06:47 - Debunking Claude Code Myths

09:38 - Bonus: Ruffalo & Agent Swarms

Turn NotebookLM Into a Self-Running Content Machine (Goldie Infinite Knowledge Engine Setup)The script explains a new setup for using Google’s free NotebookLM as a “full content machine” that can turn a single source (websites, PDFs, Google Docs, audio, etc.) into 12 different media formats such as podcasts, videos, slide decks, infographics, FAQs, reports, and flashcards. It argues most users waste NotebookLM’s potential by using it manually and describes the “Goldie Infinite Knowledge Engine,” a three-layer system: a knowledge vault for storing sources, a local agent operating system that connects to NotebookLM via an MCP bridge so agents like Claude, Hermes, and OpenClaw can operate it, and an infinite loop that regenerates content as new knowledge is added. The script contrasts an 8-hour manual workflow with a 20-minute automated process and promotes the AI Profit Boardroom for the full setup, prompts, roadmap, coaching, tutorials, and community.00:00 NotebookLM Power Unlock00:51 Cleaner Agent OS Setup03:01 NotebookLM Basics and Outputs03:46 Why Standalone Fails04:56 Goldie Knowledge Engine05:14 Three Layers Explained07:23 Manual vs Automated Workflow09:08 Dashboard and Obsidian Memory10:42 Real Use Cases Ideas11:24 Boardroom Offer and Proof12:35 Wrap Up Community Wins

OpenClaw 5.2: Huge AI Agent Update & Discord Voice Support

Discover the massive improvements in OpenClaw 5.2, including Discord voice handoff and enhanced security via the "Doctor" tool. Learn how to streamline your AI agent workflow, fix Windows install issues, and scale your operations with the new Agent Operating System.

00:00 - 00:00 - Intro

00:44 - Discord Voice Integration

02:04 - OpenClaw Doctor: Security First

03:09 - Smarter Model Status & Tracking

04:01 - XAI Remote Server Login

05:02 - Windows Installation Fixes

06:47 - Scaling with Agent OS

09:06 - Safety: Backups & Updating

Six Days That Changed AI Security: Claude Mythos, Cloudflare’s Harness, the Pentagon Task Force & Japan’s BanksBetween May 15–21, 2026, six rapid events signaled a major shift in AI-driven cybersecurity: Claude Mythos Preview became the first model to solve both UK AI Security Institute cyber attack simulations, including a 32-step offensive chain, completing it 6/10 times. Cloudflare’s Project Glasswing postmortem then admitted the model alone is “the wrong tool for achieving high coverage,” describing a multi-agent vulnerability discovery harness. Anthropic and Cloudflare launched Claude Managed Agents on Cloudflare, “decoupling the brain from the hands” by running the core loop on Anthropic while execution can run on Cloudflare. Cloudflare’s CSO said Mythos reasoning resembled a senior researcher and could chain low-severity bugs into high-risk exploits, while noting guardrails can fail under rephrasing. The Pentagon announced an offensive AI task force, and Japan’s three mega banks began preparing for expected autumn AI-driven attacks via a 36-entity working group. The script argues the common pattern is building systems around frontier models rather than deploying them raw, and promotes an “agent operating system” via AI Profit Boardroom.00:00 Six Days That Changed AI01:01 Mythos Cracks 32 Steps03:27 Cloudflare Harness Reality Check05:17 Managed Agents Brain Hands Split06:55 Senior Researcher Level Findings10:01 Pentagon Goes Offensive11:18 Japan Banks Brace For Autumn12:33 Systems Not Tools Takeaway12:58 Agent OS Pitch And Demo14:23 What Is Claude Mythos16:32 Build Your Own Agent OS18:04 Wrap Up And Thanks

Hermes Agent OS: The Free Setup to Automate Lead Outreach, Content & Support (Community Q&A)Julian explains how Hermes Agent Operating System can automate key business tasks like lead outreach, content calendars, and customer support, sharing community wins from Amanda and Jerry who set up scheduled tasks in under 30 minutes. He answers questions about using open-source agents with a mix of closed and free models via APIs (e.g., OpenRouter/OwlAlpha) and discusses Gemini Omni for video creation. He shows how to enable web research in Hermes using Firecrawl, BrowserBase, and Groq for Twitter search, and outlines a LinkedIn/X content approach using AI avatar videos and caption posts with prompts. He recommends prioritizing AI services around lead generation, demonstrates connector options in Claude, suggests using daily scheduled task check-ins to track KPI progress, advises simplifying agent setups when they break, and points viewers to AI Profit Boardroom training, zip files, prompts, coaching calls, and the community wins section.00:00 Hermes Agent OS Intro00:40 Community Wins Spotlight01:25 Open Source vs Models02:09 Gemini Omni Quick Look02:32 Web Research With Hermes03:17 LinkedIn Content System04:30 Agent Harness Explained04:53 Setup Guides And Leads05:52 Build In Public Strategy06:52 Best AI Service Focus07:37 Boardroom Access And QnA08:28 Jerry Win Scheduled Tasks09:30 Claude MCP Connectors10:24 Three Videos A Day11:34 KPI Tracking Scheduled Reports13:29 Free Manus Alternatives14:15 Agent Vibes TTS Project14:37 Omi Second Brain Overview16:36 Biggest Time Saving Automation18:05 Skool Autoresponder Options19:32 Hermes Setup Troubleshooting21:01 Best Free Setup Tips21:45 Wrap Up And Next Steps

Hermes Agent’s Free Browser Upgrade: 250 Skills + Agent OS to Run Web Tasks 24/7 (No Code)The script explains a major free upgrade to Hermes Agent that enables AI agents to browse the web and complete tasks like bookings, lead generation, data tracking, and form filling using a free library of 250 pre-built browser skills from Browse.sh/BrowserBase. It walks through basic setup in Hermes Agent, including installing Browse via an npm command, listing and installing skills, and demonstrates simple web navigation. The presenter introduces an “Agent Operating System” as a control tower to manage many browser agents from one screen and outlines the “conductor stack” framework: the user sets vision, Hermes translates plain English into actions, browser skills perform site tasks, Agent OS provides SOPs/memory/routing, and outputs can run 24/7. It addresses common myths about complexity and fragility and promotes the AI Profit Boardroom for full guides, files, prompts, training, and support.00:00 Hermes Browser Upgrade01:35 What Browser Agents Do02:04 Install Browse Skills02:35 Explore Skill Catalog03:10 Agent OS Control Tower05:22 Conductor Stack Framework06:14 Live Demo In Hermes09:30 Myths And Proof11:40 How The Stack Fits12:51 Recap And Boardroom14:17 Q&A Small Business Uses15:58 Closing Thoughts

How to Run Claude Code for Free Forever (1M Context Tokens + Agent OS Memory Stack)The script explains how to use an open-source project called Free Claude Code to run the Claude Code CLI for free by routing it to alternative model providers, specifically a free one-million-token context model called OwlAlpha on OpenRouter. It demonstrates the setup inside a custom “Agent Operating System” dashboard that unifies multiple AI agents (e.g., Claude, OpenClaude, Hermes, Gemini) with persistent memory and tools via an Obsidian vault, plus workflows like SEO tools, a studio for video/image/TTS generation, Kanban task management, journals, and status monitoring. The presenter outlines a four-pillar stack (model, Free Claude Code, memory/connections, dashboard), addresses common myths about Claude Code and free models, and directs viewers to a paid AI Profit Boardroom for guides, prompts, files, coaching calls, and community support.00:00 Free Claude Code Hook00:42 Live Demo In Agent OS02:03 What Free Claude Code Is02:55 Routing Models With OpenRouter03:46 Agent OS Explained05:41 Four Pillar Stack06:29 Obsidian Memory Engine08:34 Six Myths Debunked11:23 Why Agent OS Wins13:22 Recap And Next Steps14:09 Boardroom Offer And Support15:09 Community Proof And Wrap