AI News Today | Julian Goldie Podcast

GLM 5.1: The New 8-Hour Autonomous AI Agent Changing Coding

China's Z AI just released GLM 5.1, a massive open-source model that ranks top three globally for coding and can work autonomously for 8 hours straight. Discover how its 'long horizon' capabilities allow it to self-correct and outperform giants like GPT-4o and Gemini.

00:00 - Intro: GLM 5.1 Changes AI Agents

01:18 - Long Horizon Task Capability

02:44 - AI That Closes the Loop

04:06 - Benchmarks vs GPT & Gemini

05:23 - Mixture of Experts Architecture

06:19 - The 8-Hour Autonomous Linux Build

07:50 - The Open Source Revolution

11:06 - Building Agentic Workflows

Perplexity AI's Huge Finance Update: Manage Money with AI

Perplexity AI has integrated with Plaid, allowing you to link your financial accounts and manage your money using simple AI prompts. This update transforms the search engine into a personal financial hub that builds custom dashboards, tracks budgets, and forecasts cash flow automatically.

00:00 - Intro: Perplexity x Plaid

00:38 - The End of Manual Budgeting

01:22 - Custom AI Financial Dashboards

02:03 - 5 Powerful Use Cases

04:20 - Security: Is Your Data Safe?

06:05 - Pricing & Availability

06:45 - The Shift to a Personal AI Hub

09:10 - How to Get Started Today

Claude Code + Obsidian: Build a Stateful AI Knowledge Base

Stop using AI like a search engine and start building a stateful knowledge base that grows smarter over time. This workflow uses Claude Code and Obsidian to turn raw information into a permanent business asset inspired by Andrej Karpathy.

00:00 - Intro: The Karpathy Workflow

01:04 - The 3-Folder Architecture

02:02 - Claude Code as a Compiler

03:43 - From Wiki to Custom Reports

04:47 - Integrating Obsidian via MCP

06:36 - Stateless vs. Stateful AI

08:26 - 2-Day Implementation Plan

Hermes AI Agent: Why Everyone is Switching from OpenClaw

Hermes has become the fastest-growing AI agent of all time by introducing a revolutionary closed-loop learning system that builds its own skills. Discover how this autonomous tool outperforms the competition with native memory, cross-platform gateways, and the ability to manage complex business workflows.

00:00 - Intro

00:35 - Who is Nous Research?

01:16 - Hermes vs. OpenClaw: The Memory Gap

02:20 - The Closed Gap Learning Loop

03:27 - 40+ Built-in Tools and Gateways

06:15 - The One-Command Migration Tool

08:56 - Advanced Features: Profiles & Sub-Agents

12:06 - The Future of Set-and-Forget AI

Anthropic Just Changed AI Agents: The Claude Advisor Tool

Anthropic's new Advisor Tool gives your AI agents Opus-level intelligence at a fraction of the cost by allowing cheaper models to consult a senior strategist only when needed. Discover how this update makes complex business automations more reliable and cost-effective than ever before.

00:00 - Intro: A New Era for AI Agents

00:34 - How the Advisor Tool Works

01:22 - Performance Benchmarks & Cost Savings

01:53 - Business Use Cases & Reliability

03:47 - Under the Hood: Technical Details

06:30 - The Future of Managed Agents

07:53 - When to Use the Advisor Tool

09:52 - How to Get Started Today

Claude Code Monitor: The End of Polling and Birth of Background AI

Anthropic just released the Monitor tool for Claude Code, shifting AI from costly polling to efficient event-driven background execution. Discover how this update saves tokens and enables fully automated, always-on agents that respond to real-world events in real time.

00:00 - Intro: Claude Code's New Monitor Tool

00:12 - The Problem with Polling and Token Costs

01:35 - Automating Fixes with Log Monitoring

02:42 - Business and Non-Coder Use Cases

03:48 - Technical Deep Dive: Breaking the Agent Loop

05:15 - How to Start Using the Monitor Tool Today

06:47 - The Future of Persistent AI Infrastructure

08:14 - Your Practical Experiment for This Week

Google Gemini's New Notebooks: Your AI Second Brain Is HereGoogle just integrated NotebookLM features directly into Gemini, transforming it from a simple chatbot into a powerful project management tool. Discover how to upload 100+ sources and organize your work into dedicated notebooks to supercharge your business productivity.00:00 - Intro: Gemini's Game-Changing Update00:41 - What are Gemini Notebooks?01:10 - Uploading Sources & Conversational AI01:52 - NotebookLM Syncing & Integration02:38 - Practical Use Cases for Agencies03:42 - Rollout Schedule & Availability05:24 - How to Use Notebooks for Content06:23 - The Future of AI Collaboration

GLM 5.1: The First Open-Source AI That Works for 8 Hours StraightGLM 5.1 is a massive breakthrough in open-source AI, capable of working autonomously for up to eight hours to solve complex tasks without human intervention. Discover how this model outperforms GPT-4 and Claude in benchmarks and learn how to leverage its long-horizon capabilities for your business.⁠00:00⁠ - Intro: A New AI Shift⁠00:17⁠ - 8 Hours of Autonomous Work⁠00:58⁠ - Benchmarking GLM 5.1 vs GPT & Claude⁠02:23⁠ - Experiment 1: 6x Performance Boost⁠03:32⁠ - Experiment 2: Hardware Optimization⁠04:09⁠ - Experiment 3: Building a Desktop from Scratch⁠05:32⁠ - How to Access & Run GLM 5.1⁠06:29⁠ - The Future of Long-Horizon AI

Hermes Agent 0.8: The Massive Update You Need to SeeHermes 0.8 brings over 200 updates including background notifications, live model switching, and native Google Gemini support. This release significantly improves AI agent reliability and introduces advanced remote code execution capabilities.00:00 - Intro: Hermes 0.8 Overview00:15 - Background Notifications & Live Switching01:32 - GPT Reliability & Native Gemini Support02:51 - Smart Timeouts & Slack/Telegram Upgrades03:42 - MCP Security & Remote Code Execution05:08 - Smart Automation & System Logging06:45 - How to Update Hermes

GLM 5.1: The Sanctioned AI Beating GPT-5 & Claude

GLM 5.1 is China's latest AI model that thrives despite US sanctions by running efficiently on Huawei hardware. Discover how this open-source powerhouse uses iterative processing to shatter performance ceilings and why it represents a massive shift for local AI workflows.

00:00 - Intro: China’s Shocking New AI

00:50 - The Ceiling-Breaking Model

01:13 - How it Gets Better Over Time

02:50 - Free, Open, and Local

03:22 - Benchmarks vs GPT & Claude

04:48 - Building Without Nvidia

06:26 - Flipping the AI Cost Model