AI News Today | Julian Goldie Podcast

NotebookLM + Gemini 3.5 Agentic Upgrade: Build a Fully Automated Research Machine (Agent OS + MCP)The script explains Google’s new NotebookLM update with Gemini 3.5, adding agentic research and coding capabilities, improved reasoning, and new output formats. It shows how users can start with a loose question, have NotebookLM find and add high-quality web sources, cite every answer, chat with notebooks, and generate assets like research reports, PDFs with charts, videos, podcasts, slide decks, mind maps, infographics, quizzes, and flashcards. The speaker demonstrates connecting NotebookLM to an agent operating system via MCP so multiple agents (e.g., Hermes, Claude) can collaborate, edit outputs (including AI-generated videos), and route research into other tools like SEO workflows. The update is rolling out globally starting with Google AI Ultra members, with expanded access expected later. The video also introduces a five-step “Goldie Deep Dive” framework and promotes the AI Profit Boardroom for the full setup, tutorials, resources, training, and coaching.00:00 NotebookLM Upgrade Reveal00:57 New Agentic Features Tour03:13 Outputs and Content Formats04:02 Agent OS Integration Power05:47 Research and Source Hunting06:34 Code Runner and PDF Reports07:36 Use Cases and Workflow Stack08:51 Idea Factory Automation Loop10:12 Goldie Deep Dive Framework12:05 Old Way vs New Way14:04 Key Takeaways Recap14:19 Get the Full Agent OS15:10 Closing Thanks

Hermes Jarvis: A Voice-Controlled Agent OS That Builds Apps, Opens Sites & Saves Everything to a Memory GalaxyThe script demos “Hermes Jarvis,” a voice-controlled assistant powered by the Goldie Oracle Control System that can wake on a keyword, show information, and create working apps in real time. The presenter shows it building a Japanese flashcard game, generating an SEO website, opening bing.com, teaching basic Japanese phrases, sharing a brain fact, and visualizing “latest numbers for SEO,” while supporting click-to-talk, typing input, conversation transcripts, and a fullscreen “wall mode” mission-control display. The system briefs the user with time and workspace build counts on startup, runs on Windows or Mac, can use free local models for privacy, and saves outputs into a “memory galaxy” knowledge graph (Obsidian visualized). The video contrasts this one-tab, hands-free workflow with juggling multiple tabs/tools, and promotes getting the full Agent OS and tutorials inside the AI Profit Boardroom.00:00 Hermes Jarvis Demo00:15 Voice Commands In Action01:02 Japanese Tutor Moment01:49 Goldie Oracle Explained01:56 Wake Show Create Modes03:53 Why This Beats Chat Tabs05:03 HUD Boot Up Briefing05:50 Wall Mode And Controls06:14 Agent Auto And Visuals07:20 Mission Control Future07:52 Get The Full Agent OS08:22 More Tools Memory Galaxy08:59 Boardroom Tutorials Community09:37 Wrap Up

Obsidian Memory Galaxy: Turn Your Notes Into a 3D AI Memory System (Goldie Setup)The script introduces the free “Goldie Memory Galaxy” setup that connects an Obsidian vault to an agent operating system so AI agents like Hermes can share and retain context through a visual 3D star map where each note is a star, links form constellations, and recently updated notes glow brightest. It explains how this solves scattered, forgotten notes and “blank session” AI by letting users fly through their second brain, open notes by clicking stars, and feed agents real business context such as clients and writing voice. The five layers are outlined: stars, constellations, glow, fly-through navigation, and continuous improvements via a feedback loop where agents create and update markdown notes (including exports from Omi to Obsidian). It also highlights AI Profit Boardroom access to the Agent OS updates, tutorials, zip install, community support, and weekly coaching calls.00:00 Memory Galaxy Overview01:18 Why Notes Stay Invisible02:30 From Scattered Memories to Stars04:00 Get the Free Setup05:01 Five Layers Explained05:41 Agents Plugged Into Obsidian08:03 Old Way vs New Way09:17 How the 3D Map Works10:11 What You Can Do With It11:40 Objections and Recap12:31 Agent OS Demo and Features14:36 Community Support and Coaching15:15 Final Encouragement

Hermes Idea Foundry: Turn 1-Line Ideas into Shipped Apps with AI Agents (Obsidian + Human Approval Gate)The script demos a new Hermes project manager framework called the Goldie Idea Foundry inside an agent operating system that captures ideas, has agents classify and route them, generates a build plan, and requires only one human approve/reject step before a project manager agent and sub-agents implement and ship real deliverables viewable in a dashboard. Built outputs (e.g., an SEO calculator, a to-do list app, a meditation timer, webpages) can be previewed, rebuilt, and are saved as plain markdown in an Obsidian vault connected to a “memory galaxy” for future context. The presenter also shows a voice-activated “Hermes Jarvis” that can control the computer, discusses using local models or free APIs to manage token costs, contrasts the old idea-to-nothing workflow with the new pipeline, answers a few audience questions, and promotes the AI Profit Boardroom for the full setup, updates, coaching calls, and resources.00:00 Hermes Idea Pipeline00:40 Quick Build Demos01:18 Obsidian Memory Sync01:57 Human Gate Approval03:14 Goldie Idea Foundry04:08 Why Ideas Die04:58 Live Q&A Updates05:45 Hermes Jarvis Voice06:22 Five Step Workflow07:54 Old Way vs New09:31 Proof It Ships09:49 Board and Ownership10:41 Costs and Local Models10:56 Offer and Beliefs11:49 Member Results12:22 Models and Recap14:27 Full System Tour15:05 Community and Closing

Build Apps by Voice with Free Claude Code + Goldie Agent Factory (No Paid APIs)The video demonstrates using free Claude Code inside the Goldie Agent Factory to build projects by speaking prompts and watching code generate in real time, with live previews for items like a to-do list app, a snake game, and an SEO agency landing page. It highlights that the setup can run on free APIs and lets users swap among many free models (including options like Nemotron 3 Ultra and Gemma 4 variants), with outputs varying by API/model choice. The creator shows how the workspace helps manage and revisit previous builds in one place and explains the Agent Factory’s simple workflow connected to an agent operating system with multiple agents available. The video closes by directing viewers to the AI Profit Ballroom for a full Free Claude Code guide, an Agent OS setup zip and tutorials, daily lessons, community Q&A, and weekly coaching calls.00:00 Voice Coding Overview00:29 Live Build Demo01:34 To Do App Preview02:27 Free Models and APIs03:08 Landing Page Build04:09 Agent OS Workspace04:30 Easy Setup Proof04:58 Get the Full Guide05:36 Community and Coaching06:25 Final Thanks and Link

Hermes Agent Operating System: Mission Control, Jarvis Voice Mode, Studio & Long-Horizon Goals (Full Setup Tour)The script tours an “agent operating system” that makes Hermes Agent easier to control and far more powerful than using it in the terminal, featuring a mission control dashboard, a chat UI with saved conversation history, voice input, and email sending via a connected dedicated inbox. It demonstrates real-time talk mode enabled by Minimax M3, plus “Hermes Jarvis,” a voice butler-style assistant that can execute computer commands like opening Google. The system includes Hermes Studio for generating videos, images, and audio using Minimax and Grok OAuth, with Remotion as an alternative video skill, and an artifacts area to preview created outputs. It also covers MCP connections (e.g., Gmail), a Kanban-based multi-agent orchestration workflow, long-horizon “goal mode” with an internal judge, and new v0.16 management features like sessions, logs, scheduled tasks, channels, profiles, and provider keys, then promotes the AI Profit Boardroom for the full setup, tutorials, coaching, and community support.00:00 Hermes Agent OS Overview00:16 Mission Control Dashboard00:28 Chat Interface and History00:59 Email and Voice Input01:33 Talk Mode Real Time Voice01:53 Jarvis Computer Control03:03 Hermes Studio Media Creation04:15 Why an Agent OS Matters04:46 UI Models and Skills Tour05:30 Kanban Agent Orchestration06:23 Goal Mode Long Horizon Tasks07:57 Artifacts Preview Workspace08:30 MCP Connections Explained09:01 Manage Sessions and Scheduling10:32 More Pipelines and Wrap Up11:03 AI Profit Boardroom Community11:25 Training Setup and Coaching12:29 Testimonials and Final Thanks

Hermes Agent Skills Hub Update: Browse, Scan, Trust Tiers & One-Click Skill InstallsThe script walks through Hermes Agent’s new Skills Hub update, showing how to update Hermes and then browse, search, preview, scan, and one-click install skills from multiple connected hubs (including skills.sh, ClawHub, Clawed Marketplace, and GitHub). It highlights faster search versus the old method, built-in security scanning with verdicts that can block installs, and the importance of reviewing each skill’s skill.md instructions and trust tiers (trusted, built-in, community) with color coding and feedback. The presenter warns that installing too many skills can slow the agent, increase token usage, and cause confusion, but skills can be searched and toggled off. Finally, it promotes an “agent operating system” inside the AI Profit Boardroom for managing agents, skills, memory, Kanban orchestration, and media generation, with training, resources, coaching calls, and a community.00:00 New Skills Hub Update00:33 Update Hermes Agent01:20 Browse and Install Skills01:49 Search and Safety Verdicts02:39 Old vs New Workflow03:01 Scan and Review Skills03:46 Benefits and Cautions04:40 Goldie Skill Vault Framework05:16 Four Layers of Trust06:34 Quick Update Reminder06:49 Agent OS Overview07:45 Video Skill Example08:30 AI Profit Boardroom Offer08:49 How to Get Access09:23 Wrap Up

Anthropic’s New Warning: AI Agents Are Closing the Loop (And It’s Accelerating Fast)This episode discusses a new Anthropic report warning about rapidly accelerating AI capabilities, arguing that autonomous agents are “closing the loop” by running code, delegating work, and potentially enabling recursive self-improvement. The creator shares firsthand experience building an Agent OS and a voice-activated “Hermes Jarvis” system largely from simple prompts, describing feeling increasingly unnecessary while also noting the risk of not understanding or fixing systems when they break. Quotes from Anthropic employees highlight how AI is driving large increases in code output and changing collaboration dynamics by removing human reciprocity. The script questions future hiring needs, suggests humans’ current advantage is big-picture thinking and taste, and explains Anthropic’s view that slowing development would be beneficial but unrealistic in a competitive race without government intervention. The episode ends by promoting access to the creator’s AI Profit Ballroom community, trainings, and the Agent OS system.00:00 Anthropic Sounds Alarm00:29 AI Progression Timeline01:38 Building Agent OS Fast02:47 Hermes Jarvis Demo03:35 Inside Anthropic Quotes05:08 Feeling Useless and Lost06:42 Humans vs AI Advantage08:25 What Humans Still Do10:03 Slow Down or Accelerate10:55 Recursive Self Improvement Risk11:23 Why Pausing Is Hard12:32 Offer and Community Wrap

Free Claude Code Setup: Run Claude Code on Free APIs with an Agent OS (OwlAlpha + OpenRouter)The script explains how “Claude Code is now free” by using a free, open-source project that sits between the Claude Code CLI and the API, allowing it to plug into free model APIs (e.g., OwlAlpha, Gemma 4, and other OpenRouter free options). The presenter demonstrates building projects in real time inside a custom Agent Operating System dashboard with chat/voice input, workspace previews, and persistent memory via an Obsidian-based system, claiming this enables unlimited experimentation without token worries. Examples include a playable snake game, visual animations, real apps, and a 30-second showcase video generated using Remotion. The stack is described as Claude Code CLI + “Free Claude Code” + a free model “brain” (OwlAlpha with a million-token context window and native tool use) integrated into an Agent OS, with setup available via a zip file through the AI Profit Boardroom community.00:00 Free Claude Code Intro00:34 Demos and Results00:59 Live Build Snake Game01:32 How Free Setup Works02:55 Video Agents With Remotion03:31 Why Free Beats Tokens04:30 Best Free Models List05:35 Agent OS Access06:12 OwlAlpha Brain Explained07:21 Four Pillars Architecture10:27 Myths and Objections12:26 Easy Install and Community13:44 Wrap Up and Next Steps

Run Google Gemma 4 Locally for Free in 1 Click (Ollama + Hermes Agent Setup)The episode shows how to run Google’s newly released Gemma 4 locally for free using Ollama and connect it to Hermes Agent in a single click/prompt, highlighting new quantization-aware training weights that reduce memory requirements while maintaining quality. The presenter demonstrates Gemma 4 running inside an agent operating system with Hermes, including tool access (terminal, browser use on Mac), shared memory via an Obsidian database, and multi-agent coordination, plus quick model switching in the Hermes dashboard (including filtering for free models and using local models for sub-agents). It contrasts local/offline private use versus paid subscriptions and explains a five-step “Goldie Iron Brain” framework (vault, wiring, shared memory, dashboard, updates). The video also mentions Hermes v0.16 MCP management features and promotes the AI Profit Boardroom community and training.00:00 One Click Gemma 400:27 Setup in Hermes01:08 Why Local Models Win02:09 Free Model Switching03:35 What Is Gemma 404:28 Goldie Iron Brain05:28 Hermes Dashboard Tour07:23 Future Proof Updates08:01 Recap and Next Steps08:39 Join The Community09:33 Final Thanks