AI with Kyle - Daily AI News and Updates

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 Matters1:27 Tiktokenizer Demo: How Words Break Into Tokens4:18 Token IDs: How Models Store Words as Numbers5:35 Tokens Across Languages: Why English Is Cheapest7:59 The Raspberry Problem: Why AI Can't Count Letters9:49 Context Windows Explained: What "1 Million Tokens" Means11:40 Context Compaction: Why Chats No Longer Cut You Off13:27 Why a Bigger Context Window Isn't Always Better16:18 API Pricing: How Tokens Drive Your AI Costs18:42 "Token Maxing": The Wasteful AI Trend to Avoid20:13 Finding Real Context Limits & DeepSeek vs GPT Pricing22:36 Rules of Thumb: Optimize for Useful Context25:12 Wrap-Up & Andrej Karpathy Recommendation26:20 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 Method2:07 What Is a Pulse? Building Your Own Threads3:29 Codex vs OpenClaw: Why Focused Beats Broad4:02 Capturing on Mobile: Your Home Computer as Command Center5:41 Daily Reviews & Proactive AI Automation7:32 Kyle's 5 Capture Pulses: Thoughts, To-Dos & Content Ideas10:37 To-Do & Buy Capture: Auto-Sorting Your Purchases12:05 The Codex Mobile App: The Real Game-Changer13:20 Getting Started: Codex Pricing & Setup Walkthrough15:15 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 Explained1:30 Why Every Job Is a Bundle of Tasks2:46 Exposed Edges vs Defensible Tasks (Accountant & Zookeeper)4:06 How AI Quietly Shrinks Headcount6:04 Why Gen Z and Entry-Level Workers Get Hit First7:32 The Productivity Myth and the Demand Problem9:27 Second-Order Effects: Why No Industry Is Safe10:54 The DIY Job Audit: Map Your Weekly Tasks12:51 Build Your Human Edge: Skills AI Can't Replace14:43 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 It5:29 How to Audit Your Real Weekly Workflows7:19 When NOT to Automate: Avoid the Over-Engineering Trap9:17 Put ChatGPT in Interview Mode (Live Demo)13:04 The Simplest Tools First: Prompts, Checklists, and Templates15:01 Stepping Up: Zapier, Make.com, and n8n Automations16:10 Managed Agents, Sub-Agents, Scripts, and Custom Apps17:18 Fit The Fix To The Job: Don't Get Blinded by Viral AI19:10 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 Hype1:10 Why Google Conferences Look Flashy But Say Nothing2:13 Distribution Beats Product: The Slack vs Microsoft Teams Lesson3:53 Google's Confusing AI Product Sprawl (Jules, Opal, Antigravity, Spark...)5:53 Gemini 3.5 Flash: Benchmarks and Why People Aren't Impressed9:07 Google's Hidden Edge: Bringing AI Directly to Users10:17 The Real I/O Story Nobody Covered: AI Inside Google Search12:16 Why Google Must Cannibalize Their $82 Billion Ad Empire14:46 How Google Agents Will Demolish Lightweight SaaS Companies17:07 Project Spark: Google's Jarvis-Style AI Assistant18:21 Antigravity 2: The Codex Clone That Flopped20:21 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 AI0:45 Eric Schmidt Booed at University of Arizona Commencement2:27 Steve Wozniak's "Actual Intelligence" Counter Speech3:43 Why Schmidt Is Probably Right About AI & Young People5:09 Gen Z Isn't Anti-AI - They Distrust the Tech Industry Deal6:29 How AI Kills Entry-Level Jobs Without Replacing Workers8:55 Why New AI Roles Won't Save Gen Z From Unemployment10:28 The Career Ladder Is Dead - Start Your Own Income Now13:25 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 Changed0:43 How Codex In ChatGPT Mobile Actually Works1:18 The Multi-Device Problem Killing Your Coding Flow2:36 My Old Setup Mess & Why OpenClaw Was So Exciting3:59 Mobile Controls Desktop: How Native Codex Untethers You5:30 Hardware Setup: Mac Mini As Your Always-On Vibe Coding Machine6:42 "Control Another Device" Setting + GitHub Central Repository8:17 Adding Your Phone & The Real Studio-To-Pool Workflow9:53 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 Introduction1:12 The Big Takeaway: AI Skill Is More Than Just Prompting2:03 The Three Ways to Use AI: Automation, Augmentation & Agency3:36 The Four Ds Framework Explained4:29 D1: Delegation – What to Hand Off to AI vs Keep Yourself6:53 D2: Description – The RISEN Prompting Framework10:04 D3: Discernment – How to Spot AI Hallucinations13:47 The Description-to-Discernment Loop16:28 D4: Diligence – AI Ethics, Safety & Verification18:15 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 Machine0:50 Why an Audience Is Non-Negotiable for Business2:06 Why Education Content Beats Entertainment4:12 What I Publish Every Single Day6:08 The AI Slop Problem: Don't Automate Everything7:40 The Full Workflow Overview12:19 Step 1: AI-Powered Topic Selection13:42 Step 2: Voice Notes to Slide Decks15:18 Live Streaming (and the AI Interview Alternative)17:44 Step 3: Transcripts to Newsletters & Articles19:14 Carousels, PDFs & 100+ Emails a Day20:56 Short Form Scripts & Where to Use AI23:27 The Self-Improving Feedback Loop24:46 Tools Don't Matter, Workflows Do (My Stack)26:56 How to Start Small Without Burning Out28:40 Why Your Humanity Becomes More Valuable & Outro
— Useful Resources ——
Find everything else at https://aiwithkyle.com/