AI News Today | Julian Goldie Podcast

GLM 5.2: China’s Open-Source Model That Builds Games, Apps & Agent Teams (Hermes Setup + Benchmarks)The script showcases China’s open-source GLM 5.2 from ZAI, highlighting projects built with simple prompts—open-world and 3D games, websites, apps, and automated video creation—assembled in a single morning. The presenter claims GLM 5.2 performs near Claude Opus 4.8 on frontier SWE benchmarks, beats GPT-5.5 and Gemini 3.1 Pro on several benchmarks, ranks highly on Design Arena, and is far cheaper while being runnable locally. They demonstrate integrating GLM 5.2 into the Hermes agent system (including a voice-activated “Hermes Jarvis”) to run teams of agents that research keywords, write and judge SEO posts, and iteratively produce and edit videos, citing traffic growth for two websites. The video ends promoting the AI Profit Boardroom’s agent operating system, setup files, prompts, coaching calls, tutorials, and community.00:00 GLM 5.2 Shock Demo01:12 Benchmarks And Value04:04 Open Source Breakthrough04:49 Design Arena Surprise07:38 Side By Side Tests08:52 Hermes Agent Teamwork10:43 How To Use GLM11:11 Daily Workflow Automation12:18 Recap And Next Steps12:38 AI Profit Boardroom Pitch13:47 Community And Training14:33 Final Goodbye

I Built an Agent Operating System (No Code) That Unifies All My AI Agents in One DashboardThe presenter tours a no-code “agent operating system” they built with Claude to run and manage multiple AI agents in one dashboard, including voice-activated Hermes/Jarvis, Claude with UltraCode sub-agent teams, and Fusion, which fuses five models plus a judge and is claimed to beat Fable 5 on benchmarks; they emphasize antifragility after Fable 5 was banned by swapping in new models and tools like GroqBuild, N2 (a free API integrated into Free Claude Code with voice control and live preview), and Gemini’s managed cloud agents that can research and generate files into the workspace. The system centralizes chat switching, shared memory via an Obsidian vault, auto-saves all generated assets to a workspace/gallery, enables workflows like SEO content deployment and one-click video creation, and supports scheduling via cron management. They offer the setup and daily-updated tutorials via the AI Profit Boardroom community.00:00 Agent OS Overview01:10 What Agent OS Means01:51 Agents Working Together02:26 Model Bans And Fusion03:18 GroqBuild Landing Pages04:07 N2 Voice Coding04:52 Gemini Cloud Agents05:31 Hermes Jarvis Briefings06:18 One Dashboard Workflow07:28 Workflows And Gallery08:27 Persistent Memory Obsidian09:05 Automation And Cron Jobs10:03 Get The System Access10:30 Community And Testimonials11:44 Final Wrap Up

How to Get Fable 5-Level Intelligence Without Fable 5: Plan With a Frontier Model, Build CheapThe episode explains how to approximate “Fable 5-level” results despite Fable 5 being taken down by using a two-phase workflow: have a frontier model create a highly pinned, decision-complete plan, then switch to a cheaper (or even free) model to build from that plan, with optional review agents for security and logic. Citing a Kilo case study, it claims identical services can be produced when different models implement the same pinned plan, with major token and cost reductions—planning with Claude Fable 5 and implementing with GPT 5.5 reportedly cut costs by 59% while still passing 15 acceptance checks. The script also notes routing plans across models if one is removed, reusing plans as assets, and mentions Fusion as another multi-model approach, plus resources available inside the AI Profit Boardroom.00:00 Fable 5 Is Down00:47 Kilo Planning Case Study01:31 Costs And Token Savings02:53 Pinned Plans Explained03:31 Plan Build Review Pipeline04:40 How To Run It In Kilo06:35 Free And Cheap Builder Models07:34 Fusion Multi Model Alternative08:06 Boardroom And Closing

Hermes Just Added Async Sub-Agents: Delegate Tasks in the Background (No More Frozen Chat)Hermes now supports asynchronous sub-agents via a new Delegate tool that can spawn multiple workers at once, running jobs in the background so chat no longer freezes. After updating Hermes (via terminal command or dashboard), you can give a lead Hermes agent a single goal and have it fan out tasks to temporary sub-agents, each with its own thread, goal, and context, dispatched in about two milliseconds using delegate task with background=true. Results return to the chat as each worker finishes, enabling parallel workflows like generating and deploying multiple SEO articles from one keyword. The script compares the old blocking, sequential approach to the new parallel “hive” model, notes you can use plain English, and mentions using different APIs per worker. It also promotes the AI Profit Boardroom’s agent operating system, training, roadmap, community, and coaching calls.00:00 Async Sub Agents Launch00:15 Update Hermes Fast00:41 SEO Funnel Demo00:56 Why Background Mode Wins01:17 Real Deployment Example02:15 Profiles And Temporary Agents02:44 How Delegate Works03:45 Parallel Workers Returning Results04:25 Old Versus New Workflow05:04 Keeping Control And Avoiding Chaos05:31 Quick Recap And Next Steps06:30 Multiply Output With Boardroom07:21 Community And Coaching Details07:40 Final Thanks

Why AI Benchmarks are Fake (And How to Actually Test Models)

A fake French AI model recently went viral for beating the industry's top benchmarks, proving how easy it is to manipulate performance data. This video explains why you should stop chasing hype-filled charts and start evaluating AI based on your own real-world business workflows.

00:00 - Intro: The Le Chatton Fat Joke

01:08 - Why AI Benchmarks Can Lie

02:42 - The Problem with Self-Reported Tests

04:18 - Real Work is the Only Benchmark

05:20 - How to Avoid AI Overwhelm

06:34 - The New Way to Evaluate AI

07:31 - 3 Key Takeaways for AI Testing

08:45 - Testing AI Systems Yourself

Grok Build CLI Just Surprised Me: Side-by-Side Tests vs Claude, Kimi & GLM + Agent OS IntegrationThe script reviews the new Grok Build coding agent and CLI, saying it’s a major upgrade over the old version and the best results the creator has seen from the same prompt, especially for games and 3D builds. It showcases multiple outputs (a dark first-person game, clean websites, 3D/MMO/RPG-style demos, animations like glowing fish trails and lava-lamp liquid that follows the mouse) while noting quality is inconsistent and sometimes disconnected from prompts. The creator compares Grok Build against Kimi Code, GLM 5.2, and Claude Opus, claiming Grok Build’s game beat Claude’s for the same prompt and outperformed Kimi and GLM on some tests, though GLM looked better for certain animations. They highlight OAuth login and that X Premium Plus/Super Grok subscribers get access, plus parallel sub-agents, self-check loops, and integration into an agent operating system alongside other models, then promote an Agent OS package via AI Profit Bot.00:00 Grok Build Surprise00:37 Touring Early Demos01:12 OAuth and CLI Access02:08 Studio Media Features02:36 Beta Announcement Context03:11 Side by Side Comparisons04:37 Animations Mixed Results05:50 Parallel Sub Agents06:22 Dragon Realm Benchmark07:02 Self Check Improvement Loop07:58 Using Grok in Agent OS08:48 Agent OS Offer Breakdown09:43 Community Coaching and Close

I Turned My Mac Into a Voice-Activated AI Butler (Hermes + Jarvis + Obsidian Memory)The script demonstrates a voice-activated AI assistant running on a Mac inside an agent operating system, using Hermes Agent with Jarvis and connected to an Obsidian vault for persistent memory. The AI provides daily and weekly briefings that summarize work themes, open action items, priorities, and recent news headlines with clickable links, and it can be triggered by tapping a panel or a wake word. It can open websites and apps, show dashboards and live numbers, run “wall mode” as a command-center display, build working apps by voice with outputs saved to a gallery, and save user commands like “Remember that” directly into Obsidian so it can recall past context across days. The episode ends by promoting the Agent Operating System setup available through AI Profit Boarding.00:00 Talking Butler Demo01:33 Voice Commands and Builds02:19 What Hermes Jarvis Is03:52 Six Powers Overview04:40 Briefings and History05:42 Memory and Agent Mode06:21 How Commands Route08:22 Recap and Setup Offer09:33 Agent OS and AI Profit Boarding11:04 Final Thanks and Goodbye

GLM 52 vs Qwen 37 Max vs Claude Opus 48: Real-World Tests vs Benchmarks (No Second Chances)The episode compares GLM 52 (ZAI), Qwen 37 Max (Alibaba), and Claude Opus 48 (Anthropic) head-to-head on five one-shot tasks, arguing that benchmark rankings didn’t match real usability. In coding-focused tests like a voxel runner game, a liquid-in-a-bowl animation, a business landing page, and an arcade game, GLM 52 produced the most fun, polished, and feature-rich results, while Claude’s outputs were often basic and Qwen’s were sometimes buggy or incomplete; Claude clearly won the solar-system orbit map task. The script also notes Qwen’s strong reported benchmarks and faster replies, GLM’s slower responses in agents but strong CLI coding, and highlights limitations integrating Claude into agent workflows compared to Qwen/GLM in Hermes and the creator’s agent operating system.00:00 Head To Head Setup01:27 Coding Tests Results04:09 Arcade Game Showdown04:50 Benchmarks Versus Reality06:01 Agents Workflow Tradeoffs07:59 Final Recommendations

White House Forces Anthropic to Pull Mythos & Fable 5 Worldwide in 90 MinutesThe script recounts how Anthropic’s private security-focused model Mythos and its public, guardrailed version Fable 5 were taken offline worldwide after the Trump administration ordered Anthropic to block all foreign nationals—including non-citizen staff—from using them, with one report saying the shutdown happened within 90 minutes. It describes a weekend escalation beginning with Fable 5’s public launch, an Amazon-discovered jailbreak reported by CEO Andy Jassy, and a White House “fix it or pull it” demand using export-control-style rules typically reserved for weapons. Reporting from Semaphore suggests unconfirmed White House concerns that a China-linked group may have accessed Mythos, raising fears of “distillation” to copy capabilities. The narrator frames this as evidence AI is now treated like critical power infrastructure and urges business owners to adopt practical, publicly available AI tools, promoting the AI Profit Boardroom.00:00 Mythos Shutdown Shock01:11 Timeline to Blackout02:07 Why Mythos Matters02:31 Distillation Explained03:39 Jailbreak and Standoff04:35 AI as National Power05:17 Business Gap Warning06:07 Boardroom Pitch and Fusion07:00 Three Issues Clarified07:58 Earlier Mythos Leak08:29 Weekend DC Negotiations08:44 Everyday AI Keeps Advancing09:23 Stop Waiting Start Using09:52 Final Takeaways and CTA11:04 Closing Thanks

White House Forces Anthropic to Pull Mythos & Fable 5 Worldwide in 90 MinutesThe script recounts how Anthropic’s private security-focused model Mythos and its public, guardrailed version Fable 5 were taken offline worldwide after the Trump administration ordered Anthropic to block all foreign nationals—including non-citizen staff—from using them, with one report saying the shutdown happened within 90 minutes. It describes a weekend escalation beginning with Fable 5’s public launch, an Amazon-discovered jailbreak reported by CEO Andy Jassy, and a White House “fix it or pull it” demand using export-control-style rules typically reserved for weapons. Reporting from Semaphore suggests unconfirmed White House concerns that a China-linked group may have accessed Mythos, raising fears of “distillation” to copy capabilities. The narrator frames this as evidence AI is now treated like critical power infrastructure and urges business owners to adopt practical, publicly available AI tools, promoting the AI Profit Boardroom.00:00 Mythos Shutdown Shock01:11 Timeline to Blackout02:07 Why Mythos Matters02:31 Distillation Explained03:39 Jailbreak and Standoff04:35 AI as National Power05:17 Business Gap Warning06:07 Boardroom Pitch and Fusion07:00 Three Issues Clarified07:58 Earlier Mythos Leak08:29 Weekend DC Negotiations08:44 Everyday AI Keeps Advancing09:23 Stop Waiting Start Using09:52 Final Takeaways and CTA11:04 Closing Thanks