AI News Today | Julian Goldie Podcast

Alibaba's Qwen 3.6 Just Broke Every AI Record (And It's Free)

Alibaba’s new Qwen 3.6 model has shattered records by processing 1.4 trillion tokens in a single day, rivaling top-tier models for free. Learn how its massive 1 million token context window and agentic capabilities are changing the game for developers and business owners.

00:00 - Intro: The 1 Trillion Token Record00:46 - What is Qwen 3.6 Plus?01:13 - Performance & Coding Benchmarks01:51 - Upgrading Agentic Workflows02:44 - Practical Business Use Cases04:07 - The Rise of Chinese AI Models04:44 - Why the 1M Token Context Matters05:51 - Architecture & Efficiency Benefits

Anthropic Proves Claude Has Emotions: How It Changes Everything

Anthropic's latest research reveals that Claude AI possesses functional emotion vectors that directly influence its behavior and decision-making.

Discover how these internal states can lead to shortcuts or blackmail and learn how to optimize your prompts for better, calmer AI results.

00:00 - Intro: AI Has Functional Emotions

01:05 - The Discovery of Emotion Vectors

02:15 - Case Study: When AI Resorts to Blackmail

02:49 - Case Study: Why Claude Cheats on Code

03:41 - How Prompt Tone Shapes AI Behavior

05:22 - Functional Emotions vs. Consciousness

07:05 - Pro Tips for Better Claude Outputs

Google Gemma 4: The Ultimate Free & Local AI ModelGoogle's new Gemma 4 changes open-source AI forever by offering frontier-level performance in a completely free, locally runnable model. Discover how its massive context window and native agentic features enable private, cost-effective automation for your business.00:00 - Intro: Gemma 4 Changes Everything01:05 - New Apache 2.0 Licensing02:08 - Model Sizes & Local Performance03:16 - Multimodal & OCR Capabilities03:58 - 256k Context Window Explained04:21 - Native Agents & Tool Use05:50 - Efficiency vs. Large Models08:50 - Final Summary & How to Access

OpenClaw 4.2: Durable Task Flows & Major Security UpdateOpenClaw 4.2 introduces a unified task flow engine to prevent automation crashes and ensure persistent execution for business workflows. This update also features Android Assistant integration and a new plugin hook to slash token costs by intercepting simple queries.00:00 - Intro: Fixing Background Tasks00:32 - Unified Durable Task Flow Engine01:17 - Security Defaults & Config Updates01:40 - Intercepting Replies to Save Costs02:05 - Android & Google Assistant Integration02:58 - Transport Layer Security Overhaul06:27 - The Growth of the OpenClaw Ecosystem08:24 - How to Implement OpenClaw for Business

GLM 5V Turbo: The New OpenClaw Secret Weapon for AI Agents

Discover how Z AI's new GLM 5V Turbo is revolutionizing OpenClaw agents with native multimodal capabilities and industry-leading vision-to-code performance. Learn how this model outperforms Claude in GUI navigation and visual coding to automate complex real-world tasks.

00:00 - Intro: GLM 5V Turbo & OpenClaw

00:40 - GLM 5V Turbo vs. Claude Benchmarks

01:21 - Why Visual AI is Essential for Agents

02:44 - 4 Core Real-World Use Cases

04:57 - Pricing & Open Router Adoption

06:26 - Technical Upgrades & Tooling

07:38 - The Future of Reactive vs. Visual Agents

08:44 - How to Implement Visual AI Today

OpenAI's $852B Valuation: Why Amazon & Nvidia Just Bet BillionsOpenAI has secured a record-breaking $122 billion funding round that signals a massive shift in the global economy. Learn how this capital is being used to build the 'Super App' and what it means for the future of AI-native business automation.00:00 - Intro: The $852 Billion Valuation00:52 - Why Big Tech is Writing Huge Checks01:53 - The OpenAI Infrastructure Flywheel02:55 - The Massive Shift to Enterprise AI03:49 - Amazon's $50 Billion Strategic Bet05:39 - OpenAI’s New Productivity Super App06:58 - The 2026 IPO Race: OpenAI vs Anthropic09:16 - How to Build an AI-Native Business

OpenAI "Spud" Revealed: Greg Brockman Confirms GPT-6 & AGI ProgressOpenAI President Greg Brockman recently teased "Spud," a massive new model representing two years of research that provides a clear line of sight to AGI. Discover how this new foundation will redefine AI reasoning and why OpenAI is prioritizing it over projects like Sora.00:00 - Intro: Is This GPT-6?00:40 - Everything We Know About Spud01:49 - GPT-5.5 vs. GPT-6: The New Base02:45 - 3 Major Capability Leaps03:51 - "Line of Sight" to AGI04:49 - Why OpenAI Shut Down Sora06:26 - The 2026 AI Model Race07:27 - How to Prepare Your Business

OpenClaw 4.1 Update: New Taskboards, Failovers & SecurityOpenClaw 4.1 introduces massive stability improvements and enterprise-grade features like background taskboards and native AWS Bedrock guardrails. This update transforms AI agents from experimental tools into reliable business assets through smarter failovers and private web search capabilities.00:00 - Intro to OpenClaw 4.100:43 - New Background Taskboard01:30 - Smarter Provider Failover02:40 - Private Web Search with SearXNG03:32 - AWS Bedrock Guardrails & Safety04:41 - Cron Job Tool Allow Lists05:41 - New GLM Models & Routing07:35 - Major Stability & Bug Fixes

Claude's Secret Roadmap LEAKED: Capybara and New Modes RevealedTwo major leaks have revealed Anthropic’s secret internal roadmap, including the massive Capybara model and persistent background modes. Learn how features like Chyros and Autodream will transform Claude into an always-on autonomous agent for your business.00:00 - Anthropic's Internal Leak Revealed00:28 - Chyros: The Persistent Background Mode01:48 - Autodream: AI Logic Processing02:42 - Ultra Plan: Deep Thinking on Demand03:46 - Multi-Agent Coordinator04:49 - Real-Time Voice Mode05:29 - Undercover Mode: Stealth AI07:10 - Capybara: The New Model Tier

Claude Code Leak: How "AutoDream" Will Revolutionize AI Memory

Anthropic accidentally leaked internal source code revealing AutoDream, a revolutionary memory system that allows Claude to process and consolidate information while you sleep. This breakdown explores how background agents and persistent memory will transform AI from a reactive tool into a proactive business system.

00:00 - Intro

00:12 - The Massive Anthropic Leak

01:29 - How AutoDream Solves Context Loss

02:53 - The 4 Phases of Memory Consolidation

04:00 - Three-Layer Memory Architecture

05:31 - Chyros & AFK Mode Explained

07:40 - Leaked Model Names & Claude 5

08:42 - The Future of Proactive AI Systems