AI with Kyle

Stop asking AI to edit your video in one click. It doesn't work. Ignore the hype. Instead build the pipeline that actually works. Grab the fix kit (Free Guide): https://aiwithkyle.com/mini/ai-video-editor Get AI-Ready with Kyle's 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg I built an AI video-editing pipeline that turns a messy livestream into a cut, checked and packaged YouTube draft. The video you're watching went through that system. I show the exact split between local tools and hosted models, why the one-click AI editor pitch is still nonsense, and how this replaced an editor who previously cost me about $1,000 a month. The practical shift is to stop treating video editing as one giant prompt. Editing is too complex for that! Instead break the task up into a number of steps: Transcribe locally with Whisper, use a capable model to create the edit decision list, make the cuts with FFmpeg, review a lightweight 720p draft, then render the approved edit in 4K and upload it privately. AI can handle the machinery; you still own the thesis, performance and final approval. —— Time Stamps —— 0:00 AI Video Editing: The Hype vs Reality 0:29 What the Pipeline Actually Does 1:58 The $1,000-a-Month Proof 2:17 Why “Edit This for Me” Still Fails 2:57 Transcript → EDL → Local Render 3:45 From Messy Livestream to Source Files 6:46 Codex Builds the Edit Decision List 7:33 FFmpeg, Remotion and Hyperframes 8:35 The QA Pass That Catches Bad Edits 9:00 Why Review Starts at 720p 9:26 4K Rendering, Uploads and Thumbnail Tests 9:43 The Human Checks You Still Need 10:54 Catching the Mistakes AI Misses 12:04 Why This Finally Works Now 13:08 Build the Pipeline, Not the Fantasy — Useful Resources —— Find everything else at https://aiwithkyle.com/

Stop chasing every new model. Get the free Chinese AI 101 guide: https://aiwithkyle.com/mini/chinese-ai Get AI-Ready with Kyle's 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: Alibaba has announced Qwen3.8 and says it is second only to Claude Fable 5. But the open weights aren't out yet, the ranking is still a company claim, and a frontier model this size is not going to run on your laptop. A launch post isn't a reason to rebuild your workflow. In this video I use four checks before switching: announced, available, independently proven and useful on your own work. Then I show you how to compare Qwen3.8, Kimi K3, Claude and ChatGPT on real repeated tasks, measure quality, time, cost and failures, and only move when the gain is large enough to pay for the disruption. Most people will get more done by ignoring the leaderboard and sticking with the system that already works. —— Time Stamps —— 0:00 Qwen3.8 Made My Guide Obsolete 1:39 What Alibaba Is Actually Claiming 3:17 Why Launch Posts Are Not Proof 4:00 What Open Weights Actually Means 5:11 Does Model Size Matter? 6:21 China's Open-Weight Advantage 7:04 Ignore the Model Hype Cycle 8:36 How to Test a New Model Properly 9:16 Make Your Workflows Model-Agnostic 10:28 When Switching Providers Makes Sense 10:54 The Four Tests to Run 11:34 Why Sticking With One Provider Wins — Useful Resources —— Chinese AI 101 guide: https://aiwithkyle.com/mini/chinese-ai Find everything else at https://aiwithkyle.com/

Chinese AI isn't a future threat — it's already in the top three. Get the full breakdown (Free Guide): https://aiwithkyle.com/mini/chinese-ai Get AI-Ready with Kyle's 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Kimi K3 dropped this week and knocked Claude Opus to third place in the AI rankings. It's not the best model in the world overall, but it's within striking distance of Fable and GPT - and it costs about a third of what Claude charges on the API ($3 in / $15 out versus $10 / $50) - and China's cheaper models like GLM, DeepSeek and MiniMax run at up to 20x less. It's a deliberate strategy. Xi Jinping got on stage the same week and committed China to open-source AI. Six hundred million people in China already use AI - about half the country - adoption is growing at 142% year on year, and the gap between Chinese and American frontier models has closed from six-to-nine months to a month or two. The practical shift here is about where you spend attention and money. If you're building anything token-heavy, automations, coding agents, content pipelines etc., Chinese models deserve a serious look right now. At the same time, you need to understand the difference between cloud and local deployment, because "open source" does not mean "free on your laptop". This video gives you the full map: who's building what in China, why the pricing is a genuine existential threat to OpenAI and Anthropic's trillion-dollar valuations, and what you should actually do with any of this today. —— Time Stamps —— 0:00 Hook: Claude Just Fell to Third Place 0:29 Kimi K3 Explained: What the New Rankings Actually Look Like 1:43 The Benchmark That's Going Viral — and Why to Ignore It 3:14 20x Cheaper Than Claude: The Price Weapon China Is Using 4:09 Chinese AI Map 101: Moonshot, DeepSeek, ZAI, Qwen, ByteDance 5:04 Why Cheap AI Is Existential for OpenAI and Anthropic 6:14 Trillion-Dollar IPOs vs. a Model That Costs a Dollar Per Task 6:59 Xi Jinping's Open Source Pledge: What It Actually Means 7:39 Open Source Does Not Mean Free on Your Laptop 8:34 The Real Enterprise Threat: Local Fine-Tuned Models for Almost Nothing 9:57 Cloud vs Local: The Chinese Privacy Paradox Nobody Talks About 11:22 600M Users and a Beijing Bookshop: China's AI Reality 12:46 What to Do Now: Kimi Subscription, Local LLMs, Front-End Testing 14:16 Where to Go Next —— Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Premium AI is about to get more expensive. Build one small, useful asset while it's still cheap. Grab the fix kit (Free Guide): https://aiwithkyle.com/mini/build-while-it-lasts Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  Right now everyone is using AI like it's about to vanish, and honestly, it might. Anthropic just extended Fable 5 again, and within the hour OpenAI reset every Codex limit in response. In this video I break down why the whole industry is behaving like this, why the intelligence you're paying twenty, a hundred, two hundred dollars a month for is heavily subsidised, and what happens the moment that subsidy ends. I share what one billion tokens actually cost me in a single day once you take it off the subscription and put it through the API. It is not a small number. That subsidy is the whole story here, and it will not last. I walk through the idea of the permanent underclass, the K shaped economy, and why I think you have roughly six months to build something real before this gets expensive again. I show you exactly where to start if you have never built anything before, including how I personally use Fable as the planner and pass the actual work to cheaper models. No coding required. You do not need to be technical. You just need to start before the window closes. —— Time Stamps —— 0:00 OpenAI vs Anthropic: Fable 5 Extended Again 1:53 Why Anthropic Can't Afford to Pull Fable 5 3:32 The $50,000-a-Day Cost of Losing Subscription Access 4:59 The Fable Usage Frenzy & Rob's $800/Month Strategy 7:15 Rented Intelligence & the Permanent Underclass 8:30 The K-Shaped Economy Explained 10:58 How to Build Your First AI Project (No Coding Needed) 11:54 Free AI Business Resources & Monetization 12:54 Get Paid to Teach AI to Businesses 14:29 Final Thoughts: Build Before It's Too Late — Useful Resources —— Find everything else at https://aiwithkyle.com/

Read the full issue: https://aiwithkyle.com/ai-news/openai-anthropic-are-fighting-we-got-pony Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Premium AI is about to get more expensive. Build one small, useful asset while it's still cheap. Grab the fix kit (Free Guide): https://aiwithkyle.com/mini/build-while-it-lasts Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: Right now everyone is using AI like it's about to vanish, and honestly, it might. Anthropic just extended Fable 5 again, and within the hour OpenAI reset every Codex limit in response. In this video I break down why the whole industry is behaving like this, why the intelligence you're paying twenty, a hundred, two hundred dollars a month for is heavily subsidised, and what happens the moment that subsidy ends. I share what one billion tokens actually cost me in a single day once you take it off the subscription and put it through the API. It is not a small number. That subsidy is the whole story here, and it will not last. I walk through the idea of the permanent underclass, the K shaped economy, and why I think you have roughly six months to build something real before this gets expensive again. I show you exactly where to start if you have never built anything before, including how I personally use Fable as the planner and pass the actual work to cheaper models. No coding required. You do not need to be technical. You just need to start before the window closes. —— Time Stamps —— 0:00 OpenAI vs Anthropic: Fable 5 Extended Again 1:53 Why Anthropic Can't Afford to Pull Fable 5 3:32 The $50,000-a-Day Cost of Losing Subscription Access 4:59 The Fable Usage Frenzy & Rob's $800/Month Strategy 7:15 Rented Intelligence & the Permanent Underclass 8:30 The K-Shaped Economy Explained 10:58 How to Build Your First AI Project (No Coding Needed) 11:54 Free AI Business Resources & Monetization 12:54 Get Paid to Teach AI to Businesses 14:29 Final Thoughts: Build Before It's Too Late — Useful Resources —— Find everything else at https://aiwithkyle.com/

Stop asking ChatGPT questions, give it a job. Grab the fix kit (Free Guide): https://aiwithkyle.com/mini/chatgpt-work Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  OpenAI has released ChatGPT Work, a new work mode inside ChatGPT, and in this video I cover everything you need to know. What it actually is, how ChatGPT Work compares to Codex and Chat, and which one you should actually use depending on what you do. There is plenty of confusion out there right now, three modes, a moving model picker, and power users asking what the point is. I cut through all of it and explain why this update is aimed squarely at the billion people who use ChatGPT every day, even on the free plan. The bigger shift is this: stop chatting with ChatGPT and start giving it jobs. Instead of asking questions and getting answers, you hand it a task, it spins up sub-agents, does the research, checks its own work, and delivers a finished result. The presentation in this video was built by ChatGPT Work itself in about twelve minutes, so you can see exactly what it produces. I also show you where to find the new work mode on web, mobile and desktop, and what to use it for in your daily work. —— Time Stamps —— 0:00 ChatGPT Work: OpenAI's Biggest Release This Week 0:38 What Is ChatGPT Work? Why Everyone's Confused 2:12 Where to Find the Chat vs Work Toggle 2:43 Live Demo: ChatGPT Work Builds a Full Presentation 4:22 Model Selection Differences in Chat vs Work Mode 5:05 The Three Doors: Chat, Work & Codex Explained 5:56 Why This Matters: The Agentic Age for Normal Users 6:47 What Should You Actually Use ChatGPT Work For? 7:48 The Four Big Jumps in AI (Ethan Mollick's Framework) 9:57 The Power User Dilemma: ChatGPT Work vs Codex 11:51 Claude vs OpenAI: Going in Opposite Directions 13:01 Too Many Choices: Modes, Models & Effort Levels 14:08 How to Get ChatGPT Work on Desktop, Phone & Web 16:14 How to Delegate Real Jobs to AI (Prompting Still Matters) — Useful Resources —— Find everything else at https://aiwithkyle.com/

Stop asking ChatGPT questions, give it a job. Grab the fix kit (Free Guide): https://aiwithkyle.com/mini/chatgpt-work Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: OpenAI has released ChatGPT Work, a new work mode inside ChatGPT, and in this video I cover everything you need to know. What it actually is, how ChatGPT Work compares to Codex and Chat, and which one you should actually use depending on what you do. There is plenty of confusion out there right now, three modes, a moving model picker, and power users asking what the point is. I cut through all of it and explain why this update is aimed squarely at the billion people who use ChatGPT every day, even on the free plan. The bigger shift is this: stop chatting with ChatGPT and start giving it jobs. Instead of asking questions and getting answers, you hand it a task, it spins up sub-agents, does the research, checks its own work, and delivers a finished result. The presentation in this video was built by ChatGPT Work itself in about twelve minutes, so you can see exactly what it produces. I also show you where to find the new work mode on web, mobile and desktop, and what to use it for in your daily work. —— Time Stamps —— 0:00 ChatGPT Work: OpenAI's Biggest Release This Week 0:38 What Is ChatGPT Work? Why Everyone's Confused 2:12 Where to Find the Chat vs Work Toggle 2:43 Live Demo: ChatGPT Work Builds a Full Presentation 4:22 Model Selection Differences in Chat vs Work Mode 5:05 The Three Doors: Chat, Work & Codex Explained 5:56 Why This Matters: The Agentic Age for Normal Users 6:47 What Should You Actually Use ChatGPT Work For? 7:48 The Four Big Jumps in AI (Ethan Mollick's Framework) 9:57 The Power User Dilemma: ChatGPT Work vs Codex 11:51 Claude vs OpenAI: Going in Opposite Directions 13:01 Too Many Choices: Modes, Models & Effort Levels 14:08 How to Get ChatGPT Work on Desktop, Phone & Web 16:14 How to Delegate Real Jobs to AI (Prompting Still Matters) — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  Every week a new model lands and the same question comes with it: am I on the best one? This week alone we have GPT 5.6, Grok 4.5, a new Gemini and Fable hanging on for a few more days. My answer is to stop switching between models altogether. The model is not the system. In this beginner friendly walkthrough I show you how to get ChatGPT, Claude and Gemini working together instead of forcing yourself to pick a favourite and second-guessing it a fortnight later. The idea is simpler than it sounds. One smart model does the planning and the reviewing, a cheaper model does the grunt work, and they share a single folder plus a few plain text instruction files so every model reads from the same page. I walk through my exact setup, from voice-briefing the planner to pushing the finished work, and why it holds up whether you are shipping code or drafting a book. Once ChatGPT, Claude and Gemini are working together like this, the next release stops being a threat and becomes a free upgrade. —— Time Stamps —— 0:00 Fable Deadline & a Messy Week of AI Model News 0:53 Model Paralysis: Should You Switch AI Models? 2:26 The Model Is Not the System: Model-Agnostic Workflows 3:19 Planner vs Worker: The Smart Model / Dumb Model Split 4:05 The Real Cost of AI: $15,000 in One Day 5:16 Use Intelligence Where It Matters (Not Just for Coders) 6:27 GitHub Repo Setup: Your Central Source of Truth 8:14 README, Agents & Claude MD Files Explained 9:22 Let Your Smart Model Set Up the Whole Project 10:07 Connecting Claude Code & Codex with Plug-ins 11:33 Direct Connection: MCP & CLI Between AI Models 12:48 My Personal AI Workflow: Fable Plans, Codex Builds 15:02 Don't Panic: Use All the AI Models Together 16:38 Free Webinar: Teach AI to Businesses — Useful Resources —— Find everything else at https://aiwithkyle.com/

Read the full issue: https://aiwithkyle.com/ai-news/fable-limited-sol-here

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  Claude Fable 5 is being taken off the subscription tomorrow and dumped behind API pricing that would cost me thousands a day to run properly. Half my community is in mourning. I get it, because for one week Fable 5 tore through problems that had been sitting on my desk for months and did work I would once have paid three people to do. But grieving Fable 5 is fixating on the wrong thing. The real story is that what feels like science fiction this week becomes ordinary within a year, the models keep getting cheaper and faster, and that is brilliant news if you build things and rather less brilliant if you have a nine to five. So in this one I get on with the useful part: what you actually do next. I cover why the model going away barely matters, what is coming after it, and how to run capable AI on your own machine, free and genuinely private, with nothing sitting on someone else's server. LM Studio, Ollama, Hugging Face, what the parameter counts really mean, the hardware you need before the "run Claude for free" crowd walks you off a cliff, and the small local setup I lean on every single day.  —— Time Stamps —— 0:00 Claude Fable Is Going Behind a Paywall 1:12 Fable's API Pricing: $10 In, $50 Out 2:02 One Week With Fable: Debt Payoff, Book Tools & 15 Builds 3:32 We've Seen This Movie Before: The Model Hype Cycle 4:01 GPT-5.6 Rumors: Sol, Luna & Terra Coming This Week? 5:13 AI Costs Are Collapsing Up to 900x 6:44 The Release Cycle: Anthropic, OpenAI, Google & Open Models 7:54 The Apollo Analogy: 5MB of RAM Landed on the Moon 8:51 The Big Takeaway: Your Job Is Not Safe 11:42 What to Do: Build Something You Control 13:25 Where to Start: Free Courses & Finding Demand 14:55 AI Workshops: The $1K–$3K Per Hour Opportunity 16:24 Wrap-Up: Fable Leaves, Intelligence Doesn't — Useful Resources —— Find everything else at https://aiwithkyle.com/

Read the full issue: https://aiwithkyle.com/ai-news/boomeranging Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: Claude Fable 5 is being taken off the subscription tomorrow and dumped behind API pricing that would cost me thousands a day to run properly. Half my community is in mourning. I get it, because for one week Fable 5 tore through problems that had been sitting on my desk for months and did work I would once have paid three people to do. But grieving Fable 5 is fixating on the wrong thing. The real story is that what feels like science fiction this week becomes ordinary within a year, the models keep getting cheaper and faster, and that is brilliant news if you build things and rather less brilliant if you have a nine to five. So in this one I get on with the useful part: what you actually do next. I cover why the model going away barely matters, what is coming after it, and how to run capable AI on your own machine, free and genuinely private, with nothing sitting on someone else's server. LM Studio, Ollama, Hugging Face, what the parameter counts really mean, the hardware you need before the "run Claude for free" crowd walks you off a cliff, and the small local setup I lean on every single day. —— Time Stamps —— 0:00 Claude Fable Is Going Behind a Paywall 1:12 Fable's API Pricing: $10 In, $50 Out 2:02 One Week With Fable: Debt Payoff, Book Tools & 15 Builds 3:32 We've Seen This Movie Before: The Model Hype Cycle 4:01 GPT-5.6 Rumors: Sol, Luna & Terra Coming This Week? 5:13 AI Costs Are Collapsing Up to 900x 6:44 The Release Cycle: Anthropic, OpenAI, Google & Open Models 7:54 The Apollo Analogy: 5MB of RAM Landed on the Moon 8:51 The Big Takeaway: Your Job Is Not Safe 11:42 What to Do: Build Something You Control 13:25 Where to Start: Free Courses & Finding Demand 14:55 AI Workshops: The $1K–$3K Per Hour Opportunity 16:24 Wrap-Up: Fable Leaves, Intelligence Doesn't — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  Fable 5 is back on the normal Claude subscription, but only for a few days, and then it vanishes behind a genuinely eye-watering price. Which puts you in an odd spot: the smartest AI in the world is sitting in your app, and you have roughly one prompt a day to use it. So the question stops being what it can do, and becomes what on earth you should ask it. This video is my answer, built around one idea: treat it less like a chatbot and more like the Oracle of Delphi. I walk through the method I have been using all week. Leave the mundane jobs to the cheaper models. Use one of those cheaper models to help you draft the question first, then hand the finished thing to Fable 5 in a single fully specced shot, with a safety net so it never has to come back and ask. And the questions worth your one shot are the uncomfortable ones. Where am I wasting my time. What am I avoiding. What would embarrass me if someone sharp studied my work for an hour. Big model, big questions. It turns out the hard part was never the AI. —— Time Stamps —— 00:00 Fable Is Back: You Have Until July 7th 01:33 Fable Will Be the New Normal 02:07 Two Mistakes People Are Making With Fable 03:23 The Right Mindset: Big Model, Big Problems 04:49 Use a Dumber AI to Write Your Fable Prompts 06:34 Usage Limits & the One-Prompt Strategy 07:45 How I Use Fable: Brutal Self-Challenge Questions 09:26 The Real Skill: Asking Better Questions 10:45 Time-Sensitive: Why $20 Is Worth It Right Now 12:05 Your Homework This Weekend 13:04 Resources & Wrap-Up — Useful Resources —— Find everything else at https://aiwithkyle.com/

Read the full issue: https://aiwithkyle.com/ai-news/i-ll-stop-soon Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: Fable 5 is back on the normal Claude subscription, but only for a few days, and then it vanishes behind a genuinely eye-watering price. Which puts you in an odd spot: the smartest AI in the world is sitting in your app, and you have roughly one prompt a day to use it. So the question stops being what it can do, and becomes what on earth you should ask it. This video is my answer, built around one idea: treat it less like a chatbot and more like the Oracle of Delphi. I walk through the method I have been using all week. Leave the mundane jobs to the cheaper models. Use one of those cheaper models to help you draft the question first, then hand the finished thing to Fable 5 in a single fully specced shot, with a safety net so it never has to come back and ask. And the questions worth your one shot are the uncomfortable ones. Where am I wasting my time. What am I avoiding. What would embarrass me if someone sharp studied my work for an hour. Big model, big questions. It turns out the hard part was never the AI. —— Time Stamps —— 00:00 Fable Is Back: You Have Until July 7th 01:33 Fable Will Be the New Normal 02:07 Two Mistakes People Are Making With Fable 03:23 The Right Mindset: Big Model, Big Problems 04:49 Use a Dumber AI to Write Your Fable Prompts 06:34 Usage Limits & the One-Prompt Strategy 07:45 How I Use Fable: Brutal Self-Challenge Questions 09:26 The Real Skill: Asking Better Questions 10:45 Time-Sensitive: Why $20 Is Worth It Right Now 12:05 Your Homework This Weekend 13:04 Resources & Wrap-Up — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  Everyone's suddenly an expert on open source AI, and almost nobody's using the term correctly. I break down what open weights actually means, why it's not the same as open source and definitely not the same as free, and why your laptop has zero chance of running a 753 billion parameter model no matter how many browser tabs you close first. I walk through Hugging Face, show you what's actually inside these model files, and explain the real difference between downloading a model and being able to run one. If you've heard people throw around GLM, DeepSeek or Qwen and nodded along without having a clue what any of it means, this sorts it out in non tech speak. —— Time Stamps —— 0:00 Open Source vs Open Weight AI: Why It Matters 0:40 The Open Source Price War: Why Anthropic & OpenAI Are Scared 1:49 Claude Sonnet 5 vs GLM 5.2: Anthropic's IPO Nightmare 3:15 What Open Source Actually Means (It's Not Free) 4:48 Open Weights vs Open Source: The Taxi, Car & Kit Car Analogy 7:08 How to Download AI Models: Hugging Face Walkthrough 8:06 Inside GLM 5.2: Licenses, Safetensors & 753 Billion Parameters 10:32 Running AI Models Locally: Hardware Requirements & Quantization 11:55 American vs Chinese Open Source Models: Who's Left in the Game 13:34 Open Source AI Rankings: The Top Models Are All Chinese 14:57 Trillion-Dollar IPOs vs China's Undercut Strategy 17:58 Cost Per Intelligence: $2.75 vs 50 Cents Per Task 20:39 The AI Triage System: Which Model for Which Task 23:06 Newsletter, Webinar & How to Get Paid Teaching AI — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: Everyone's suddenly an expert on open source AI, and almost nobody's using the term correctly. I break down what open weights actually means, why it's not the same as open source and definitely not the same as free, and why your laptop has zero chance of running a 753 billion parameter model no matter how many browser tabs you close first. I walk through Hugging Face, show you what's actually inside these model files, and explain the real difference between downloading a model and being able to run one. If you've heard people throw around GLM, DeepSeek or Qwen and nodded along without having a clue what any of it means, this sorts it out in non tech speak. —— Time Stamps —— 0:00 Open Source vs Open Weight AI: Why It Matters 0:40 The Open Source Price War: Why Anthropic & OpenAI Are Scared 1:49 Claude Sonnet 5 vs GLM 5.2: Anthropic's IPO Nightmare 3:15 What Open Source Actually Means (It's Not Free) 4:48 Open Weights vs Open Source: The Taxi, Car & Kit Car Analogy 7:08 How to Download AI Models: Hugging Face Walkthrough 8:06 Inside GLM 5.2: Licenses, Safetensors & 753 Billion Parameters 10:32 Running AI Models Locally: Hardware Requirements & Quantization 11:55 American vs Chinese Open Source Models: Who's Left in the Game 13:34 Open Source AI Rankings: The Top Models Are All Chinese 14:57 Trillion-Dollar IPOs vs China's Undercut Strategy 17:58 Cost Per Intelligence: $2.75 vs 50 Cents Per Task 20:39 The AI Triage System: Which Model for Which Task 23:06 Newsletter, Webinar & How to Get Paid Teaching AI — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  I want to talk about something that's been bothering me for weeks now. GPT 5.6 just landed, and it's not coming to you or me. It's going straight to a vetted list approved by the US government, the same thing that happened to Claude Mythos and Fable not long before it. In this one I break down what's actually in the GPT 5.6 release, how Sol and Terra compare to Mythos on the benchmarks that matter, and why the real question in 2026 isn't whether a model is smart anymore. Every model is smart now. The question is whether you're allowed to use it. Then I get into what to actually do about it. I cover open source alternatives like GLM and Gemma, why running a model locally on your own machine has gone from a nerdy side project to a genuinely sensible backup plan, and how to start spreading your dependence across more than one provider before the next lockdown happens. If you run a business on any of these tools, this is the part you really shouldn't skip. —— Time Stamps —— 0:00 Why I'm Disillusioned With AI for the First Time 0:51 AI Lockdowns: Fable Banned & GPT-5.6 Restricted 2:52 Meet GPT-5.6: Sol, Terra & Luna Model Family 4:16 Too Dangerous to Release: The US Government Steps In 5:44 The Benchmarks: Just as Powerful, Far More Efficient 8:39 The Real Threat: Bioweapons & Hacking Banks 9:10 The Permission Era: A New Digital Divide 10:11 The Fix: Open Weights & Open Source 10:48 Chinese Open Source Models: GLM 5.2 Max 11:41 Run Powerful AI Locally on Your Own Machine 13:12 Dario Amodei's Crusade Against Open Source 13:56 How to Set Up Local Models on Your Devices 14:27 Don't Depend on One Model (or One Country) 15:53 The New Question: Do You Even Have Access? 16:51 IPO Fallout & Final Thoughts — Useful Resources —— Find everything else at https://aiwithkyle.com/

Read the full issue: https://aiwithkyle.com/ai-news/ai-access-becoming-permissioned Get the free resource: https://aiwithkyle.com/mini/hedge-your-stack?utm_source=youtube&utm_medium=description&utm_campaign=mini_hedge_your_stack Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: I want to talk about something that's been bothering me for weeks now. GPT 5.6 just landed, and it's not coming to you or me. It's going straight to a vetted list approved by the US government, the same thing that happened to Claude Mythos and Fable not long before it. In this one I break down what's actually in the GPT 5.6 release, how Sol and Terra compare to Mythos on the benchmarks that matter, and why the real question in 2026 isn't whether a model is smart anymore. Every model is smart now. The question is whether you're allowed to use it. Then I get into what to actually do about it. I cover open source alternatives like GLM and Gemma, why running a model locally on your own machine has gone from a nerdy side project to a genuinely sensible backup plan, and how to start spreading your dependence across more than one provider before the next lockdown happens. If you run a business on any of these tools, this is the part you really shouldn't skip. —— Time Stamps —— 0:00 Why I'm Disillusioned With AI for the First Time 0:51 AI Lockdowns: Fable Banned & GPT-5.6 Restricted 2:52 Meet GPT-5.6: Sol, Terra & Luna Model Family 4:16 Too Dangerous to Release: The US Government Steps In 5:44 The Benchmarks: Just as Powerful, Far More Efficient 8:39 The Real Threat: Bioweapons & Hacking Banks 9:10 The Permission Era: A New Digital Divide 10:11 The Fix: Open Weights & Open Source 10:48 Chinese Open Source Models: GLM 5.2 Max 11:41 Run Powerful AI Locally on Your Own Machine 13:12 Dario Amodei's Crusade Against Open Source 13:56 How to Set Up Local Models on Your Devices 14:27 Don't Depend on One Model (or One Country) 15:53 The New Question: Do You Even Have Access? 16:51 IPO Fallout & Final Thoughts — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  Today I'm tackling the problem nearly everyone using AI seriously eventually hits: your tools don't talk to each other. ChatGPT on your phone has no idea what Claude Code did on your desktop, and Codex on your laptop is working blind to both. I walk through how to build a vendor agnostic AI vault, a shared brain made of simple markdown files that any tool can read, so your context stops getting lost between apps. I show you how to get an AI to interview you and build the entire folder structure for you, no technical setup required, and where GitHub and Obsidian actually fit into the picture (hint: Obsidian is for you, not the AI). If you've seen those gorgeous Obsidian graph diagrams floating around and felt like you were missing something, this clears it up. I explain why the vault, not Obsidian, is the actual brain, why GitHub is the smartest way to sync it across devices, and how to keep growing the system over time so it gets more useful the more you use it. Whether you're juggling client projects, personal life admin or a dozen half finished side builds, this gives every AI tool you use the same source of truth. —— Time Stamps —— 0:00 The Problem: Your AI Tools Don't Share Memory 0:36 The Obsidian Vault Hype, Explained Simply 1:30 Why Your AI Context Is Scattered in 2026 2:39 What Is a Markdown File? 4:17 How Obsidian Works and Why the Vault Matters More 5:33 Structuring Your Vault: Personal, Business, Projects 6:35 Let the AI Interview You to Build It 7:14 README, AGENTS.md and CLAUDE.md Explained 9:13 Use Claude Code or Codex: The Interview Prompt 10:02 Pro Tip: Use a Microphone and Just Talk 10:38 Make It a GitHub Repo for a Durable Vault 11:47 Recap: Obsidian Is Optional, Start Now 12:27 Wrap-Up and Free Newsletter — Useful Resources —— Find everything else at https://aiwithkyle.com/

Read the full issue: https://aiwithkyle.com/ai-news/build-one-shared-ai-vault Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: Today I'm tackling the problem nearly everyone using AI seriously eventually hits: your tools don't talk to each other. ChatGPT on your phone has no idea what Claude Code did on your desktop, and Codex on your laptop is working blind to both. I walk through how to build a vendor agnostic AI vault, a shared brain made of simple markdown files that any tool can read, so your context stops getting lost between apps. I show you how to get an AI to interview you and build the entire folder structure for you, no technical setup required, and where GitHub and Obsidian actually fit into the picture (hint: Obsidian is for you, not the AI). If you've seen those gorgeous Obsidian graph diagrams floating around and felt like you were missing something, this clears it up. I explain why the vault, not Obsidian, is the actual brain, why GitHub is the smartest way to sync it across devices, and how to keep growing the system over time so it gets more useful the more you use it. Whether you're juggling client projects, personal life admin or a dozen half finished side builds, this gives every AI tool you use the same source of truth. —— Time Stamps —— 0:00 The Problem: Your AI Tools Don't Share Memory 0:36 The Obsidian Vault Hype, Explained Simply 1:30 Why Your AI Context Is Scattered in 2026 2:39 What Is a Markdown File? 4:17 How Obsidian Works and Why the Vault Matters More 5:33 Structuring Your Vault: Personal, Business, Projects 6:35 Let the AI Interview You to Build It 7:14 README, AGENTS.md and CLAUDE.md Explained 9:13 Use Claude Code or Codex: The Interview Prompt 10:02 Pro Tip: Use a Microphone and Just Talk 10:38 Make It a GitHub Repo for a Durable Vault 11:47 Recap: Obsidian Is Optional, Start Now 12:27 Wrap-Up and Free Newsletter — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  The most advanced AI model ever released to the public, Anthropic's Fable 5, was pulled from shelves roughly 48 hours after launch. Not by Anthropic. By the US government, citing national security and export control concerns. The ban applies to all foreign nationals, meaning non-US citizens are locked out entirely. I went through the full timeline on the live: the jailbreak, the Amazon angle, David Sacks' government-side narrative, and why Dario Amodei arguably talked his own model into a ban by spending months telling everyone it was a weapon. The more interesting question though isn't whether we get Fable back. It's what this reveals about the structure underneath all of it. You don't own your AI access. It's rented land. A government or a boardroom decision and it's gone, sometimes with 90 minutes' notice. In this one I get into what that actually means practically: local models, open source alternatives, the geopolitical chess game between the US and China, and the one skill that matters regardless of which model you're using.  —— Time Stamps —— 0:00 Fable & Mythos Banned: What Happened and Why It Matters 0:51 Background: Fable vs Mythos, Project Glasswing & the "Super Weapon" 2:58 Why People Fell in Love With Fable 5 (Expert Reactions) 7:10 The Ban Timeline: Launch, Backlash & the 90-Minute Export Order 9:52 What Triggered the Ban? Speculation & the Enforcement Problem 11:29 The David Sacks Narrative: The Government's Side of the Story 13:50 Pliny Jailbreaks Fable 5 15:15 Anthropic vs the Administration: The Standoff 17:38 The Foreign National Rule & the Digital ID Threat 20:09 Why You Should Learn Local & Open-Source Models (LM Studio Guide) 24:34 Think Bigger: The Real Skill as AI Keeps Improving 25:42 Big Open Questions: China, the IPO & AI Geopolitics 29:00 Newsletter & Webinar: Get Paid to Teach AI — Useful Resources —— Find everything else at https://aiwithkyle.com/

Hand big work to AI without getting burned (Free Guide): https://aiwithkyle.com/mini/fable-commissioning Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  Claude Fable 5 and Mythos are here, and they've quietly split the AI world in two. In this episode I break down exactly what these models are, why most people are only getting Fable and not Mythos, and what the June 22nd deadline actually means for your access. The hierarchy has shifted: it no longer goes Opus, Sonnet, Haiku. There's a new tier above all of that, and it costs significantly more than your current subscription.The uncomfortable truth is that $200 a month now puts you in the lower class of AI user. The models capable of genuinely transformative work are heading toward pay-per-use pricing at roughly $40 an hour, and only businesses with serious cash flow will run them at scale. I cover what this bifurcation means practically, who it affects, and how to get the most out of Fable while you still have free access.  —— Time Stamps —— 0:00 This Week in AI: Mythos, Fable 5 & the IPO News 1:20 First Impressions: Why Fable 5 Feels Like a Step Change 3:13 Prompting vs. Commissioning: A New Way to Work With AI 4:55 The Backstory: "Most Dangerous Model," Glasswing & the IPO 6:13 The New Model Hierarchy: Where Fable Fits 7:00 Safeguards: Why Biomedical Researchers Got Cut Off 7:50 Ethan Mollick's Isochronic Map: What Fable Can Really Do 9:02 AI as a $40-an-Hour Employee & the Impact on Jobs 10:37 The Black Box Problem & Distillation Attacks 11:56 Is Fable Overkill? When Not to Use It 12:38 Solving Big Problems: The Youth Unemployment Case Study 13:42 The June 22nd Cutoff & the AI Wealth Divide 16:15 Controversies: Gatekeeping, Jailbreaks & Transparency 17:42 AI for the People or the Elites? OpenAI's Opening 18:50 Wrap-Up & Free Newsletter — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  AI search is eating into Google and if your business isn't showing up inside ChatGPT, Claude, or Google's AI Overviews, you're already behind. In this video I go through what Ahrefs found across 1 billion data points and 14 studies: which content formats actually get cited by AI, why schema markup does essentially nothing, and why YouTube turns out to be the single strongest signal for AI brand visibility. Yes, really. Practical stuff only. Blog listicles, homepage rewrites, reviews, directories, podcasts, and a YouTube strategy most businesses are completely sleeping on. Nobody's even agreed on a name for this yet (GEO? AEO? LLM SEO?) but the playbook is simpler than you'd think. —— Time Stamps —— 0:00 Intro: How to Be Discoverable Inside AI 01:07 What Do We Even Call It? AI SEO vs GEO vs AEO 1:39 AI Still Looks for Real Evidence Across Surfaces 2:59 Blog Listicles: ChatGPT's #1 Cited Content Format 3:40 Chatbots vs Google AI Overviews Explained 5:34 Wikipedia & the Sources Marketers Can't Control 6:35 The Hidden Discovery Layer: Zero Google Visibility 7:46 Retrieved vs Cited: ChatGPT's 50% Citation Gap 8:27 Off-Site Signals: Reviews, Podcasts & Directories 9:22 Why YouTube Is the #1 Signal for AI Visibility 10:43 How AI Overviews Are Killing Your Website Clicks 12:28 Comparison Pages & Why Schema Markup Is a Waste 14:36 Stop Tricking the Algorithm — Just Be Useful 15:19 The Practical AI Visibility Checklist 16:17 Wrap-Up & Free Resources — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  AI is getting very good at writing, coding, and creating polished content. So good, in fact, that polish is now the least interesting thing you can offer. In this one I make the case that the most valuable skill you can build right now is getting on camera and talking, because that is exactly what AI cannot replicate convincingly. I walk through the full progression from short-form video to livestreaming to paid AI workshops for businesses, and why non-technical communicators are the people companies actually want to hire. You have a phone and an opinion. That is genuinely all you need to start —— Time Stamps —— 0:00 Public Speaking & "Yapping": Your AI-Proof Skill 1:26 Why Polished Content Now Kills Your Engagement 3:34 From Short-Form Video to the Stage: The Progression 7:16 Don't Hide Behind AI Avatars or Voice Clones 9:14 Turn Content Into a Business (and Niche Down) 12:03 What "Yap" Content Is & Twitter's Video React 13:09 The Fundamentals of Yap Content 14:45 How to Start: Phone, Apps & Recording Tips 16:58 Final Thoughts: Why Being Human Wins — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  Tokenization is one of those concepts that sounds technical until you realise it explains basically everything. Why AI can't count the letters in "raspberry." Why a bigger context window sometimes makes things worse. Why your API bill might be quietly spiralling. This is a 101 lesson on tokens: what they actually are, how models use them, and why the common assumption that a token equals a word is close but not quite right. The second half gets into context windows, the working memory of a model, and why the "just give it everything" instinct is both expensive and counterproductive. Whether you're building on the API or just a heavy user wondering why long threads start getting weird, understanding this stuff changes how you work. Also, token maxing is a trend and it is, in my view, a bit dumb. We get into that too. —— Time Stamps —— 0:00 Intro: What Tokenization Is & Why It Matters 1:27 Tiktokenizer Demo: How Words Break Into Tokens 4:18 Token IDs: How Models Store Words as Numbers 5:35 Tokens Across Languages: Why English Is Cheapest 7:59 The Raspberry Problem: Why AI Can't Count Letters 9:49 Context Windows Explained: What "1 Million Tokens" Means 11:40 Context Compaction: Why Chats No Longer Cut You Off 13:27 Why a Bigger Context Window Isn't Always Better 16:18 API Pricing: How Tokens Drive Your AI Costs 18:42 "Token Maxing": The Wasteful AI Trend to Avoid 20:13 Finding Real Context Limits & DeepSeek vs GPT Pricing 22:36 Rules of Thumb: Optimize for Useful Context 25:12 Wrap-Up & Andrej Karpathy Recommendation 26:20 Newsletter & Sign-Off — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  Codex Pulses are dedicated AI threads that run on a schedule, collect everything you throw at them, and send you a triage report every morning. Credit to Dan Shipper for the original idea. I tried it, built my own version, and haven't opened a to-do app since. This is the full setup walkthrough. The system works like this: you speak into the Codex mobile app on your phone, it syncs to your home computer, and your pulses do the processing overnight. Thought capture, to-dos, things to buy, content ideas. Each gets its own thread, its own rules, its own output. No coding required, no complex configuration (like Open Claw), just chat. —— Time Stamps —— 0:00 Intro: What Are Codex Pulses? 0:56 Dan Shipper's Codex Power User Method 2:07 What Is a Pulse? Building Your Own Threads 3:29 Codex vs OpenClaw: Why Focused Beats Broad 4:02 Capturing on Mobile: Your Home Computer as Command Center 5:41 Daily Reviews & Proactive AI Automation 7:32 Kyle's 5 Capture Pulses: Thoughts, To-Dos & Content Ideas 10:37 To-Do & Buy Capture: Auto-Sorting Your Purchases 12:05 The Codex Mobile App: The Real Game-Changer 13:20 Getting Started: Codex Pricing & Setup Walkthrough 15:15 Recap & Wrap-Up — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: AI isn't going to tap you on the shoulder and tell you you're redundant. It's quieter than that. What's actually happening is a slow compression of the tasks that make up your working week, and if you're not paying attention to which ones are at risk, you'll be caught out. In this video, I walk you through how to break your job down into its component tasks, identify which ones are genuinely defensible, and work out where to focus your energy. I've also built a diagnostic tool that does a lot of the heavy lifting for you. You put in your job title, add a description, and it maps your weekly tasks against labour market exposure data to give you a real picture of where you stand. It's free. The link's above. But even if you want to run this manually first, I show you exactly how to do that in this video too. It's not about doom, it's about knowing where you actually are. —— Time Stamps —— 0:00 Will AI Take Your Job? The Real Threat Explained 1:30 Why Every Job Is a Bundle of Tasks 2:46 Exposed Edges vs Defensible Tasks (Accountant & Zookeeper) 4:06 How AI Quietly Shrinks Headcount 6:04 Why Gen Z and Entry-Level Workers Get Hit First 7:32 The Productivity Myth and the Demand Problem 9:27 Second-Order Effects: Why No Industry Is Safe 10:54 The DIY Job Audit: Map Your Weekly Tasks 12:51 Build Your Human Edge: Skills AI Can't Replace 14:43 Your Action Plan and the AI Job-Risk Tool — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: I've seen too many people rush straight into building AI agents for tasks that don't need them. This is the corrective. Starting from a viral prompt retweeted by Greg Brockman, OpenAI's co-founder, I walk through a proper workflow audit - how to find the tasks in your actual working week that are genuinely worth automating, and how to spot the ones that aren't. The framework is simple: audit your real work first, then match the fix to the job. A reusable prompt in a Word doc beats a managed cloud agent if that's all the task needs. I also show you how to put your AI into interview mode to surface repeating workflows you didn't even know you had. A practical system for getting more done with less friction. —— Time Stamps —— 0:00 The Viral Self-Improvement Prompt (Greg Brockman + Vaibhav) 4:06 Why You Shouldn't Automate for the Sake of It 5:29 How to Audit Your Real Weekly Workflows 7:19 When NOT to Automate: Avoid the Over-Engineering Trap 9:17 Put ChatGPT in Interview Mode (Live Demo) 13:04 The Simplest Tools First: Prompts, Checklists, and Templates 15:01 Stepping Up: Zapier, Make.com, and n8n Automations 16:10 Managed Agents, Sub-Agents, Scripts, and Custom Apps 17:18 Fit The Fix To The Job: Don't Get Blinded by Viral AI 19:10 Wrap Up + Where to Get the Prompt — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: Google I/O is over, the dust has settled, and the takes have mostly been garbage. So here's what actually matters from Google's big conference this week. Hint: it's not the video generator. The story everyone missed is Google's move to embed agentic AI directly into Search. That's where 70% of their revenue comes from, and they're knowingly blowing it up. I walk through why that's a genuinely brave decision, what it means for SaaS businesses, how it compares to the OpenAI approach, and why distribution will always beat product quality. Also: Gemini 3.5 Flash benchmarks, Antigravity 2's messy launch, and why I'm not particularly impressed with any of it,  but still think this was one of the more important AI weeks of the year. —— Time Stamps —— 0:00 Intro: Cutting Through the Google I/O Hype 1:10 Why Google Conferences Look Flashy But Say Nothing 2:13 Distribution Beats Product: The Slack vs Microsoft Teams Lesson 3:53 Google's Confusing AI Product Sprawl (Jules, Opal, Antigravity, Spark...) 5:53 Gemini 3.5 Flash: Benchmarks and Why People Aren't Impressed 9:07 Google's Hidden Edge: Bringing AI Directly to Users 10:17 The Real I/O Story Nobody Covered: AI Inside Google Search 12:16 Why Google Must Cannibalize Their $82 Billion Ad Empire 14:46 How Google Agents Will Demolish Lightweight SaaS Companies 17:07 Project Spark: Google's Jarvis-Style AI Assistant 18:21 Antigravity 2: The Codex Clone That Flopped 20:21 Final Verdict: What Actually Matters from Google I/O — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: AI isn't taking your job. It's doing something more insidious — it's stopping new ones from being created. Entry-level roles are down 35% since 2023, and the grunt work that used to build careers (research, admin, junior analysis, legal discovery) is gone, absorbed quietly by AI with no fanfare and no replacement. That's the real story behind Eric Schmidt getting booed off a graduation stage, and it's worth sitting with. The old deal — degree, job, ladder, pension — is finished. Wozniak's message that you have "actual intelligence" is a lovely thought, but actual intelligence without opportunity just becomes frustration. So here's what I'd actually recommend: stop waiting for a government fix or a tech exec to have a change of heart, and start carving out something small that you own. A skill, a service, a product. AI makes that more achievable than it has ever been, and right now it might be the only career ladder left worth climbing. —— Time Stamps —— 0:00 Tech CEOs Booed at Graduation Speeches Over AI 0:45 Eric Schmidt Booed at University of Arizona Commencement 2:27 Steve Wozniak's "Actual Intelligence" Counter Speech 3:43 Why Schmidt Is Probably Right About AI & Young People 5:09 Gen Z Isn't Anti-AI - They Distrust the Tech Industry Deal 6:29 How AI Kills Entry-Level Jobs Without Replacing Workers 8:55 Why New AI Roles Won't Save Gen Z From Unemployment 10:28 The Career Ladder Is Dead - Start Your Own Income Now 13:25 Action Steps to Stay Useful in the AI Economy — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join   Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg   Summary: Building an AI product used to mean months of development, a co-founder who could actually code, and probably a few rounds of VC money you'd eventually regret taking. That's not the deal anymore. In this video, I walk you through how to build your first AI product from scratch, including why people will absolutely pay for something they could technically do themselves in ChatGPT, how to find the one process in your industry worth turning into software, and the exact technical stack I use: Vercel for hosting, Supabase for your database, and Codex to do most of the actual building.   The goal here isn't a billion-dollar startup. It's something smaller, more useful, and genuinely achievable this week. I cover the full customer journey, the logic layer (where all the real value lives), and how to think about stripping down an existing SaaS into something specific enough to actually sell. I'm also giving away the prompt I use to get Codex to scaffold the whole thing, available in the newsletter at [aiwithkyle.com](http://aiwithkyle.com/).   —— Time Stamps ——     0:00 Build Your First AI Product This Afternoon 1:11 Why People Pay for AI Tools They Could Build Themselves 4:20 How to Find a Process Worth Productizing 5:35 Real Example: Productizing My Content System 7:24 Steal Ideas From SaaS Companies (Rocket Money & Vexly) 10:52 Start Small: Don't Try to Out-Build DocuSign 12:05 One Painful Input, One Useful Output (Product Examples) 14:02 Customer Journey vs Builder Journey 15:53 Under the Hood: Vercel, Supabase & The Logic Layer 20:36 The Codex Prompt That Builds Your Product For You 22:31 Be the High-Agency Person Who Actually Ships     — Useful Resources ——   Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: Codex just shipped a mobile app, and it's the last piece I needed to properly untether myself from my desk. In this one, I walk you through how to set up an always-on vibe coding workflow across your desktop, laptop, and phone. Meaning, you can start a session at home, walk out to a cafe, and pick it up on your phone without losing a single thread of context. The key is combining the new Codex mobile release with the "Control Another Device" feature. I also get into why I've quietly shifted from Claude Code to Codex as my main tool. The model quality gap has closed, and the usage limits on Claude just aren't worth the stress anymore. If you're still running one session per machine and wondering why it feels fragmented, this is the setup that fixes it. Full technical checklist and written guide are in the Monday newsletter at [aiwithkyle.com](http://aiwithkyle.com/). —— Time Stamps —— 0:00 Codex Now Has A Mobile App: Vibe Coding Just Changed 0:43 How Codex In ChatGPT Mobile Actually Works 1:18 The Multi-Device Problem Killing Your Coding Flow 2:36 My Old Setup Mess & Why OpenClaw Was So Exciting 3:59 Mobile Controls Desktop: How Native Codex Untethers You 5:30 Hardware Setup: Mac Mini As Your Always-On Vibe Coding Machine 6:42 "Control Another Device" Setting + GitHub Central Repository 8:17 Adding Your Phone & The Real Studio-To-Pool Workflow 9:53 Why I'm Switching From Claude Code To Codex — Useful Resources —— Find everything else at https://aiwithkyle.com/

Tired of shipping polished AI rubbish? (Free Guide): https://aiwithkyle.com/mini/ai-fluency Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: Anthropic has a free AI Fluency course and, it’s good. I watched the whole thing and pulled out the frameworks that are actually worth your time, so you don't have to sit through the full thing yourself. The big idea is this: prompting is one small part of a much larger skill set. Anthropic breaks it down into four competencies. Interestingly the one people skill entirely is ‘discernment’, that's your human judgment telling you whether the AI actually did a decent job, or whether it's handed you a beautifully formatted hallucination. I walk through all four Ds, the three modes of AI use, and the description-to-discernment loop that, once you have it in your head, changes how you work with AI entirely. The course is free, but hopefully with this video I've saved you about an hour. —— Time Stamps —— 0:00 Welcome & Series Introduction 1:12 The Big Takeaway: AI Skill Is More Than Just Prompting 2:03 The Three Ways to Use AI: Automation, Augmentation & Agency 3:36 The Four Ds Framework Explained 4:29 D1: Delegation – What to Hand Off to AI vs Keep Yourself 6:53 D2: Description – The RISEN Prompting Framework 10:04 D3: Discernment – How to Spot AI Hallucinations 13:47 The Description-to-Discernment Loop 16:28 D4: Diligence – AI Ethics, Safety & Verification 18:15 Final Takeaways & Why I Recommend This Course — Useful Resources —— Find everything else at https://aiwithkyle.com/

How Kyle turns ONE live session into 20 pieces of content (Free Guide): https://aiwithkyle.com/mini/content-machine Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: Every day I produce a YouTube video, a newsletter, five to eight short-form videos, carousels, PDFs, and LinkedIn posts, and there are two of us. In this video I walk through the complete AI-powered content system that makes that possible, from morning topic selection via an AI agent, through voice-note prep and slide generation, to the content processing workflow that turns a single live stream transcript into 20-plus assets automatically. This isn't a pitch for going all-in day one. I also cover the minimum viable version for anyone who's not ready to livestream to five platforms simultaneously, including how to use AI to interview you instead, capture your voice and opinions, and still spin out a newsletter and short-form scripts without getting anywhere near a camera. The key principle throughout: the AI handles the boring infrastructure, you stay in for the parts that actually require a human. —— Time Stamps —— 0:00 Intro: How I Use AI to Become a Content Machine 0:50 Why an Audience Is Non-Negotiable for Business 2:06 Why Education Content Beats Entertainment 4:12 What I Publish Every Single Day 6:08 The AI Slop Problem: Don't Automate Everything 7:40 The Full Workflow Overview 12:19 Step 1: AI-Powered Topic Selection 13:42 Step 2: Voice Notes to Slide Decks 15:18 Live Streaming (and the AI Interview Alternative) 17:44 Step 3: Transcripts to Newsletters & Articles 19:14 Carousels, PDFs & 100+ Emails a Day 20:56 Short Form Scripts & Where to Use AI 23:27 The Self-Improving Feedback Loop 24:46 Tools Don't Matter, Workflows Do (My Stack) 26:56 How to Start Small Without Burning Out 28:40 Why Your Humanity Becomes More Valuable & Outro — Useful Resources —— Find everything else at https://aiwithkyle.com/

Tired of ChatGPT and Claude spitting out generic answers that don't sound like you? (Free Guide): https://aiwithkyle.com/mini/system-prompt-builder Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: I came across Marc Andreessen's viral system prompt. The co-founder of Netscape, one of the sharpest minds in Silicon Valley, posted his custom ChatGPT instructions online and got absolutely roasted for it. Telling an AI it has "intellectual firepower on par with the smartest people in the world" doesn't make it smarter. Telling it to never hallucinate doesn't stop hallucinations. And asking for answers "as long and detailed as possible" just produces verbose nonsense that buries any actual value. His system prompt is stuck in 2023. In this video I fix it. We go through exactly what's broken in Andreessen's custom instructions and why, then build a system prompt from scratch using the RISEN(TM) framework — a structured approach to writing AI instructions that actually holds up in 2026. You'll learn why role declarations don't make AI smarter (but are still useful), why adjectives are useless without verbs, how to write custom instructions that change AI behaviour rather than just describe it, and how to define an end goal that stops your AI padding everything out. Whether you're using ChatGPT custom instructions or Claude's system prompt settings, this will change how you prompt. —— Time Stamps —— 0:00 Marc Andreessen's Viral System Prompt Is Getting Mocked Online 1:14 Why "World-Class Expert in All Domains" Doesn't Work 2:19 "Never Hallucinate" Is Not How AI Actually Works 2:58 Tone, Provocative Answers, and Why "Bloviate" Padding Backfires 3:54 The Anti-Sycophancy Trick That Actually Works 5:39 System Prompts vs Conversational Prompts Explained 6:13 How to Add Custom Instructions in ChatGPT and Claude 8:03 The Role Declaration Myth (Ethan Mollick Research) 10:41 Use Verbs, Not Adjectives, In Your Prompts 11:39 Why "Think Step By Step" Is Outdated in 2026 14:07 You Can't Prompt Your Way Out of AI Hallucinations 16:47 Why Asking for Long Detailed Answers Wastes Tokens 18:49 The RISEN Prompt Engineering Framework Explained 24:46 RISEN in Action: A Better Version of Marc's Prompt 26:13 The Coffee Order Analogy: When NOT to Engineer Prompts 28:18 Free Quick-Start Guide and Outro — Useful Resources —— Find everything else at https://aiwithkyle.com/

 Turn Your Industry Experience Into An AI-Powered Income Stream (Free Guide): https://aiwithkyle.com/mini/ai-insider Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  I'm covering one of the biggest strategic moves I've seen from the AI labs: OpenAI and Anthropic both launched consulting and deployment ventures on the same day, backed by Goldman Sachs, Blackstone, BCG, McKinsey, and a handful of others. This isn't just big money doing big things. It tells you exactly where AI is heading and, more importantly, what that creates for people like you and me. The real focus here is what you can actually do about it. I walk through why your industry knowledge combined with basic AI skills puts you in the same position these companies are paying billions to replicate, just at a scale that actually makes sense for a one or two-person outfit. Workshops, advisory, implementation: the playbook is straightforward, the market is wide open, and most industries haven't had anyone claim that space yet. —— Time Stamps —— 0:00 Intro: AI Labs Push Into Consulting & Advisory 0:30 The May 4th Announcements: $10B AI Enterprise Deals 1:53 Following the Smart Money: AI Engineers Embedded Inside Companies 3:23 The Palantir Playbook: How AI Locks In Forever 4:55 The Jagged Frontier: Why AI Got Coding First 7:08 From Coding to White Collar: The Harvard Study Proof 8:16 When AI Fails: The Government Report Disasters 9:13 The Hilarious Irony: Wall Street Is Next 11:26 What This Means for Your Job 12:46 The Solopreneur Opportunity Gap 14:54 4 Ways to Profit From AI: Advisory, Implementation, Education, Content 18:26 My Story: Building 300K Audience in Two Years 19:07 Where to Start: Why Workshops Are the Best Entry Point 21:35 Recap & Newsletter CTA — Useful Resources —— Find everything else at https://aiwithkyle.com/

Stop AI from making things up. Get the free Hallucination Defence Kit (Free Guide): https://aiwithkyle.com/mini/ai-hallucinations Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  AI hallucinations aren't going away, and the reason might surprise you. In this one, I break down exactly what's happening under the hood when your chatbot confidently lies to you. Whether it’s making up a book, a statistic, or a legal citation, why it happens more with the smarter reasoning models, and why calling it a "bug" fundamentally misunderstands what these systems actually are. More importantly, I give you a set of practical things you can do today to reduce the damage: which settings to turn on, when to use stronger models, how to prompt for accuracy instead of agreement, and a genuinely underrated use of NotebookLM. —— Time Stamps —— 0:00 Intro: What Are AI Hallucinations? 0:32 When AI Hallucinations Go Public: Chicago Sun-Times & Deloitte 2:28 The Scary Stats: 92% Don't Fact-Check AI 4:05 The Two Types of Hallucinations You Need to Know 5:56 Why ChatGPT Isn't a Database (How LLMs Actually Work) 8:01 Andrej Karpathy: Hallucination Isn't a Bug, It's the Feature 12:18 Will AI Hallucinations Ever Be Fixed? 14:39 When Hallucinations Are Most Likely to Happen 18:17 Practical Fixes: Web Search, Paid Plans & Stronger Models 21:22 The NotebookLM Trick: Build Your Own Verified AI 22:17 Better Prompting: How to Stop AI From Lying to You 24:28 The New AI Paradigm (and Final Thoughts) — Useful Resources —— Find everything else at https://aiwithkyle.com/

Does your writing smell like AI? Grab the fix kit (Free Guide): https://aiwithkyle.com/mini/goblins Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  OpenAI banned the word ‘Goblin’ from their models, and the reason why is a reinforcement learning feedback loop that turbo-charged a "nerdy personality" setting into goblin obsession. This is an interesting insight into how AI writing ticks develop in the first place.  Practically, what does this mean for you? I walk you through Wikipedia's living list of AI writing giveaways (yes, they built a 15,000-word guide to protect themselves from AI slop, and yes, I show you how we’ll be using it against them), how to compress it into system instructions and a skill file, and why running your output through both a pre-write and post-write filter is the closest thing to a reliable humaniser you're going to get. I also discuss AI detectors, spoiler: they're worse than useless, and the academic institutions using them to penalise students are in for a rough few years. —— Time Stamps —— 0:00 Intro: OpenAI Banned Goblins From ChatGPT 1:42 Why Goblins? The Reinforcement Learning Loop 4:33 What Are AI Tells? 5:39 Wall of Shame: AI's Most Overused Words 7:46 Wikipedia's Signs of AI Writing Guide 9:42 How to Stop AI From Sounding Like AI (Belt & Braces System) 13:02 Why AI Detectors Don't Work (And Never Will) 15:59 AI Tells Are a Moving Target 19:18 The Disclosure Paradox: Does Anyone Actually Care? 22:00 Free Resources & Wrap-Up — Useful Resources —— Find everything else at https://aiwithkyle.com/

New to GitHub? Skip the confusion (Free Guide): https://aiwithkyle.com/mini/github Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary:  GitHub keeps coming up everywhere if you're building with AI, and nobody ever properly explains what it actually is. So here it is: a complete beginner's guide to GitHub, what it does, why it matters, and how to get it set up and connected to your AI tools of choice, whether that's Claude Code, Codex, Lovable, or Cursor. Think of it as Google Drive for your projects. One central folder that everything syncs to, across every tool and every device. I walk you through creating your first repo, the key terminology you'll actually encounter (commit, push, pull, branch), and how to connect it all up without touching the terminal once. I also show you how I use GitHub beyond code, as the backbone of my content system and second brain. One repo, set up properly, changes how you work with AI permanently. —— Time Stamps —— 0:00 What Is GitHub & Why Vibe Coders Need It 2:05 The Version Control Nightmare Without GitHub 4:05 Create Your First GitHub Repo (Step by Step) 6:12 GitHub Terms Explained: Repo, Commit, Push, Pull 8:42 Cloning, Branching & Merging Made Simple 11:01 Connect GitHub to Lovable, Cursor, Claude Code & Codex 13:12 GitHub Beyond Code: The Obsidian Second Brain Workflow 15:39 Install GitHub Desktop (No Terminal Required) 18:01 Best Practices: Folder Structure & When to Commit 19:57 Your Action Plan: Set Up GitHub Today 21:33 Wrap Up & Newsletter Resources — Useful Resources —— Find everything else at https://aiwithkyle.com/

Want to run AI privately on your own laptop? (Free Guide): https://aiwithkyle.com/mini/local-llm Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: Running a large language model on your own computer sounds like the kind of thing that requires a computer science degree and a spare weekend. It doesn't. In this video I walk you through exactly how to get a local AI model running on your machine today, from downloading LM Studio to picking the right model for your hardware. We cover how to figure out what your device can actually run (RAM is the main thing), what quantization means and why it's not as scary as it sounds, and why Gemma 4 is the sensible starting point for almost everyone. There are real limitations to local models and I'm not going to pretend otherwise. But for general chat, for working offline, for keeping your data off someone else's servers, they're genuinely good and getting better fast. By the end of this you'll have a working local LLM and a decent handle on where this whole space is heading. —— Time Stamps —— 0:00 Intro: How to Run AI on Your Device 2:11 What Is a Local Large Language Model? 3:53 4 Reasons to Run AI Locally (Privacy, Cost, Offline, Learning) 6:35 LM Studio vs Ollama: Best Tool to Start 8:49 Finding Local AI Models on Hugging Face 10:25 Best First Model for Beginners: Gemma 4 11:35 How Much RAM Do You Need? The Parameter Rule 15:36 Quantization Explained: Compressing AI Models 17:08 Picking Your Model Inside LM Studio 19:08 Edge AI: Run AI on Your iPhone or Android 21:06 Hobbyist vs Production: When Local AI Makes Sense 22:45 Outro & Free Resources — Useful Resources —— Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg Summary: The ground has shifted. The nine-to-five used to be the low-risk option, the sensible, boring, reliable choice for anyone who didn't fancy betting everything on a business idea. That deal is off the table. AI is attacking your job at three levels simultaneously: the tasks inside your role, the company that employs you, and the industry you work in. Any one of those can collapse without warning, and most people are sitting there hoping it won't. I'd rather you got a bit annoyed at me now than blindsided later. The good news, and there is good news, is that entrepreneurship has never been more accessible or less intimidating than it is right now. I'm not talking about raising venture capital or moving to San Francisco. I'm talking about making your first quid from a stranger on the internet, building a small side income, and slowly rewiring your brain to understand that you can generate your own money. AI makes the operational parts of running a small business genuinely manageable for one or two people. The risk-reward on doing your own thing has never looked more reasonable by comparison. —— Time Stamps —— 0:00 Why Your Job Isn't Safe From AI Anymore 0:26 When the 9-to-5 Was Actually the Safe Bet 1:32 The Three Levels of AI Risk: Job, Company, Industry 5:19 Take Responsibility — The Brutal Data on AI Layoffs 9:09 The Risk-Reward of Work Has Completely Flipped 13:13 How Gen Z Are Building Their Own Way Out 17:14 Objection 1: "But Not Everyone Wants a Business" 19:56 Where to Start: Freelancing With AI as Your Team 22:15 Objection 2: "But My Job Is Creative" 25:02 Objection 3: "Just Go Into the Trades" 27:19 Why Businesses Can Be Beautifully Simple 31:13 The Bottom Line: Start Generating Your Own Income — Useful Resources —— Find everything else at https://aiwithkyle.com/