AI News Today | Julian Goldie Podcast

GPT-5.6 Sol, Terra & Luna: Limited Preview, Benchmarks, Access (and Why Systems Beat Models)The episode breaks down OpenAI’s GPT-5.6 release, explaining that it’s currently a limited preview available only to about 20 partner organizations via Codex and the API at the request of the US government, with general access hoped for in the coming weeks and possibly restricted by country. It introduces the three new models—Sol (frontier, with Sol Ultra and Sol), Terra (balanced, competitive with GPT-5.5 at roughly half the cost), and Luna (fast, affordable for high-volume work)—and highlights benchmark claims such as Sol Ultra’s 91.9 on Terminal Bench 2.1 while noting other tests where competitors still lead. The speaker argues that chasing new model launches is a mistake and recommends building a model-proof “agent operating system” with a swap layer, multiple models on tap, task routing, and owned memory so models can be swapped in and out with minimal effort, then promotes access to their Agent OS and community.00:00 GPT 5.6 Overview00:23 Preview Access Limits00:42 Meet Sol Terra Luna01:45 Release Timing Uncertainty03:12 Benchmarks Breakdown05:27 Stop Chasing Models07:54 Model Proof System08:16 Swap Layer Routing Memory10:43 Common Objections Answered12:36 Agent OS Offer13:22 Community Tour14:19 Final Call To Action

Agent OS Q&A: Using Codex, Claude Code, Local LLMs, GLM 5.2, Team Setups, Updates & AI Video EditingJulian answers recent community questions about his Agent Operating System, a system for linking and orchestrating AI agents with features like voice control via Hermes Jarvis, a Memory Galaxy connected to an Obsidian vault, Kanban orchestration, and a local agent engine. He explains multiple ways to use a Codex subscription inside Agent OS (Codex tab, Hermes with GPT‑4.5 and image generation, OpenCloude integration, and a new Codex plugin inside Claude Code), and emphasizes building repeatable systems rather than relying on prompts or specific models. He discusses Qwen 3.6 vs Qwen Agent World, options for connecting Codex/Claude Code to local models via config changes or Ollama, and why GLM 5.2 performs strongly versus other models while being openly available. He also covers simplifying Agent OS for teams with hidden UI sections and SOPs, faster AI video assembly via Agent OS/Remotion workflows, how to update Agent OS via the latest zip and update file, and creating a shared company “brain” with a shared Obsidian vault.00:00 Agent OS Overview00:48 Using Codex Inside OS02:24 Systems Over Prompts04:00 Qwen Agent World Talk04:37 Local LLM Connections05:16 Why GLM 5.2 Wins07:08 Sharing OS With Team09:14 Fast SaaS Video Creation10:37 Best Ollama Model Setup11:48 Updating Agent OS Builds12:27 Memory Galaxy Workflow13:06 Team Profiles Shared Brain14:24 Wrap Up Join Boardroom

GPT 5.6 (Sol/Terra/Luna) & Claude Mythos/Fable 5 Are Gated—Here’s How to Beat Frontier Models Without AccessThe script covers two announcements: OpenAI’s GPT 5.6 models Sol, Terra, and Luna are released only in a limited US government–requested preview to a small group of trusted partners, with no confirmed public release date, and Anthropic’s Claude Mythos 5 and Fable 5 are also gated to about 100 companies via a deal referenced from CNBC. The speaker argues that instead of waiting for inaccessible frontier models, users can reach similar or better results using multi-model systems: Hermes Agents’ Mixture of Agents, Sakana Fugu, and Fusion, which combine multiple models and use judging/fusion to improve outputs and benchmark performance (including results shown on Goldy Bench). The episode promotes the AI Profit Boardroom community for access to the Agent OS, tutorials, updates, and coaching.00:00 Two Big AI Updates00:11 GPT 5.6 Preview Release01:11 Sol Terra Luna Breakdown01:33 Why Models Are Gated02:10 Claude Mythos Fable News03:03 Best Workaround Systems03:12 Hermes Mixture of Agents05:34 Sakana Fugu Benchmarks06:28 Fusion Beats Single Models09:23 Cost Local Model Options10:13 Stop Chasing Models11:32 Recap and Access Limits12:08 AI Profit Boardroom Tour13:42 Final Thanks

The Infinite Context Engine: Make Your AI Agents Remember Everything (Obsidian + Claude + Omi)The video introduces a system that combines Obsidian, Claude, and shared AI agents to solve the “blank slate” problem where agents forget context between chats. Using a shared Obsidian vault visualized as a “memory galaxy,” the setup automatically captures daily activity, stores it as linked markdown notes, and lets any agent (Claude, OpenClaude, Hermes, and CLIs) read the same file-path-based memory so each conversation starts with full context. The workflow is described as a four-layer loop—capture with the free Omi app, organize and tag notes (e.g., PARA), store locally in Obsidian, and deploy across agents—where every new chat is written back into the vault so the vault trains the agents and the agents enrich the vault over time with zero manual upkeep after setup.00:00 Never Forget Anything00:50 Memory Galaxy Demo01:52 Infinite Context Engine02:44 Blank Slate Problem03:37 Three Tool Pipeline03:49 Omi Captures Daily04:38 Obsidian Stores Vault06:04 One Brain Agents06:32 Four Layers Loop08:45 Galaxy Navigation10:04 Old Way Vs New11:18 Free Setup Promise11:35 Roadmap And Offer13:02 Final Thanks

Ornif 1.0 (9B) Local Model Demo: Build Apps Privately, Offline & Free in Agent OSThe video introduces the local AI model Ornif 1.0 (9B) and demonstrates how it’s integrated into an agent operating system to run AI agents locally for free on Mac or PC. The presenter shows Ornif planning and coding a habit tracker in HTML, previewing and saving creations in a workspace, and highlights benchmark comparisons where Ornif outperforms Qwen 3.5 (including some larger models) and Gemma 4 on tests like Terminal Bench. Key benefits emphasized are privacy, offline use, no API costs or limits, and a 9.5GB model building real apps on-device. The script also covers using Ornif with workflows like Hermes, Agent Kanban, local chat, and a loop engine for bug-finding and self-fixing, plus where to get setup help via the AI Profit Boardroom.00:00 Meet Ornif 9B00:38 Agent OS Local Setup01:16 Building a Habit Tracker02:09 Benchmark Breakdown03:30 Local First Benefits04:16 Previewing the App05:00 Using Hermes and Agents06:46 Self Fixing Local Engine08:15 More Agent OS Workflows09:20 Common Objections Answered09:41 Join the Community10:37 Final Thanks

GPT-5.6 Is Gated: Staggered Rollout, Customer-by-Customer Approval, and How to Stay ProductiveThe episode discusses reports that GPT-5.6 Plus is being rolled out in a staggered, limited preview rather than to everyone at once, allegedly after the US government asked OpenAI to limit the release following the Fable 5 incident where a powerful model was taken down after three days. Citing articles and reports from Axios, Silicon Angle, The Information, and a tweet by Stephanie Palazzolo, it describes a customer-by-customer approval process during the preview period, with a hoped-for wider rollout a couple of weeks later (possibly mid-July), and notes Sam Altman reportedly said this isn’t OpenAI’s preferred long-term approach. The script contrasts this gated US trend with open releases of Chinese models like GLM 5.2, and argues the solution is building an “Agent OS” system that can swap models quickly so work continues even when models are gated or pulled, promoting the AI Profit Boardroom and its tools, training, tutorials, community, and coaching.00:00 GPT-5.6 Limited Access00:42 Government Staggered Rollout01:40 Approval By Customer02:15 China Open Source Contrast03:06 New Normal For Launches04:43 How Gated Launch Works05:18 Why This Sets Precedent07:21 Stop Relying On Drops09:07 Agent OS Approach09:50 AI Profit Boardroom Pitch10:36 Wrong Assumptions Now11:51 Live System Examples12:34 Key Takeaways And Fix13:29 Boardroom Features Tour14:41 Final Call To Action

Stop AI “Slop” Web Design: Install Claude Design Skills for Unique, On-Brand PagesThe video explains how to use Claude for better web design by installing or creating “design skills” that prevent generic AI-built websites with purple gradients, emojis, and identical card layouts. It demonstrates before-and-after examples showing how teaching Claude a set of open-source UI/UX principles can produce differentiated landing pages, guides, and full sites with topic-matched palettes, fonts, whitespace, a single bold signature element, and light animation. The script references skill sources like SkillMD, ShadCN UI, and UI/UX Pro Max, notes the need to review or rewrite external skills for safety, and shows how these rules can be wired into an agent operating system (Claude/OpenClaude/Hermes) with shared memory so pages align with brand and goals. It ends by pitching the AI Profit Boardroom for the full system, prompts, tutorials, and coaching.00:00 AI Design Slop Problem00:33 Before and After Results02:03 Open Source Design Skills02:44 Install Safely or Rewrite03:21 One Click Roles and Palettes04:28 Real World Examples06:14 Design Moves That Work08:43 Seven Before After Demos10:00 Old Way vs New Engine11:47 Where Skills Live in Agent OS12:22 Full System Offer12:53 Beliefs and Takeaways13:48 Final Recap and CTA

GPT 5.5 Instant: Faster Replies, Better Intent Understanding + How to Use It in Hermes AgentThe script introduces OpenAI’s newly released GPT 5.5 Instant, describing it as faster and better at understanding intent, following complex constraints, and improving shopping and local recommendations. It notes the rollout begins with subscribers today and expands to free users tomorrow, and explains it can be used inside Hermes Agent via OAuth with an existing ChatGPT subscription, including image generation. The presenter shows a side-by-side speed comparison against Groq Build using a “tell me a joke” prompt, with GPT 5.5 Instant responding in about three seconds versus about nine seconds. The video discusses potential use cases for lightweight chat and delegation, mentions speculation that it may be a stopgap due to an expected GPT 5.6 delay, and offers to run benchmarks on Goldie Bench. It ends by promoting the AI Profit Boardroom community and Hermes agent operating system resources.00:00 GPT 5.5 Instant Launch00:41 Hermes Setup and Access01:00 Speed Test Side by Side02:09 Outputs and Rollout Details03:29 Why It Matters Use Cases04:27 Release Timing and Setup Commands05:05 Benchmarks and Recap06:08 Agent OS and Community Pitch07:10 Wrap Up

GPT-5.6 May Be Government-Gated: What It Means (and Why Systems Matter More Than Models)The script discusses reports from Axios and The Information claiming the Trump administration asked OpenAI to stagger GPT-5.6’s release, limiting access to government-approved partners during a preview period due to security and cybersecurity concerns, with two federal offices vetting and approving customers one by one and only a hoped-for wider rollout later. It links this to the “Fable 5” situation where a powerful model was taken down after a few days, suggesting a new reality where frontier US models may be delayed or never broadly released. The speaker weighs pros (safety and vetting) and cons (US competitiveness as Chinese open-weight models like GLM rapidly improve and remain widely available), then advises building workflows as model-agnostic systems so models can be swapped without disrupting work, and promotes an agent operating system offered via AI Profit Ballroom.00:00 GPT 5.6 Restricted00:52 Fable 5 Fallout01:19 Government Approval Queue03:31 Who Gets Access04:17 Global Competition Risk06:18 Build Model Agnostic Systems07:39 Automation Examples09:09 Agent OS Pitch10:01 Local Private Options10:42 Wrap Up and Outlook

My AI Agent Operating System (Mission Control to Automate SEO, Content, Outreach & Builds)The speaker demos an “agent operating system” dashboard that centralizes and orchestrates multiple AI agents to automate building projects, generating and publishing content, and running workflows. The system includes a group chat for CLIs, an idea-to-implementation pipeline, and tools like Hermes Oracle to pull trending X news, draft posts, and publish SEO blog content directly to WordPress. Hermes Jarvis enables voice-controlled building with transcripts and wake-word mode, while an email outreach SaaS module finds, enriches, validates leads, drafts campaigns, and manages inbox/sent and sending limits. The OS also features model panels like Sakana Fugu and Fusion for multi-model “team” outputs, a workspace to store all builds, a “memory galaxy” knowledge graph that auto-logs activity, and automations for Google Search Console keyword research, avatar videos, and media generation, with an option to get the setup via the AI Profit Boardroom community.00:00 Agent OS Overview00:24 Mission Control Dashboard01:30 Hermes Oracle Trends02:45 Personalized SEO Publishing03:39 Hermes Jarvis Voice04:17 Email Outreach SaaS05:33 Fugu Boardroom Models06:48 Local Engine Fusion08:07 Build Process Benchmarks09:02 Memory Galaxy Graph10:42 Automating Daily Workflows13:34 Community Setup Offer14:54 Final Wrap Up