Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join

Kyle’s AI Canon — Curated list of AI learning essentials: https://aiwithkyle.com/ai-canon

Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg

This video is from the ‘AI with Kyle Live Stream’

First aired: 6 February 2026

Summary: Honestly, I'm feeling a bit overwhelmed this morning, and that's saying something considering I wade through AI news every single day. Yesterday's release of Opus 4.6 has properly knocked me sideways—not just because the benchmarks are stellar, but because for the first time I'm viscerally feeling like AGI might actually be around the corner. I threw some heavy problems at it and watched it work, and suddenly all those CEO promises about artificial general intelligence don't feel like hype anymore. It's that "three sleepless nights" moment Ethan Mollick wrote about in Co-Intelligence, where you have to genuinely reckon with what this means for your work, your kids, and frankly, the entire human race.

The timing is wild because we've also got OpenAI dropping GPT 5.3 and Codex 5.3 in the same window, so the competition is absolutely cooking right now. Meanwhile, the stock market took a proper header yesterday as investors started processing what rapid AI advancement actually means for traditional business. I spent six hours yesterday testing these new models side-by-side, and Codex genuinely beat Claude Code on a website migration task, which I wasn't expecting. The step change from Wednesday to Thursday has been stark enough that even someone neck-deep in this stuff daily is having to take a breath and process it all.

Watch full edited live stream: https://youtu.be/l68NSRW3cRs

—— Time Stamps ——

0:00 Opus 4.6 Just Dropped — Is This AGI?

—— More Useful Resources ——

Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join

Kyle’s AI Canon — Curated list of AI learning essentials: https://aiwithkyle.com/ai-canon

Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg

This video is from the ‘AI with Kyle Live Stream’

First aired: 5 February 2026

Summary: The AI drama is heating up. Anthropic just dropped an attack ad targeting ChatGPT's decision to include adverts in their free tier, which is a bold move considering Claude's relatively tiny user base. Sam Altman's response was perfectly brutal—basically saying "we're providing free AI to the world and using ads to fund it, meanwhile hardly anyone uses your product." Fair point, honestly. It's fun watching these tech giants take shots at each other, and ultimately it means better models and more choice for us.

Meanwhile, the UK government managed to spend £4.1 million (that's $5.5 million) on an "AI Skills Hub" that's essentially just a poorly organised catalogue of free courses you could've found on Google. The site's a mess—verification emails land in spam, there's no pricing information for paid courses, and I'm genuinely annoyed enough that I'm rebuilding it myself to show what £4 million should actually get you. On a brighter note, Google quietly revealed they've got 750 million users on Gemini, which means they're catching up to ChatGPT fast whilst everyone else is busy firing shots at each other.

Watch full edited live stream: https://youtu.be/JNqAvlt5tHk

—— Time Stamps ——

0:00 UK Government Wasted £4 Million on This AI Website

—— More Useful Resources ——

Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join

Kyle’s AI Canon — Curated list of AI learning essentials: https://aiwithkyle.com/ai-canon

Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg

This video is from the ‘AI with Kyle Live Stream’

First aired: 4 February 2026

Summary: I spent the morning walking through OpenAI's new Codex Mac app, which is essentially their AI coding tool wrapped in a chat interface rather than a traditional development environment. The big news is it's free for the next two months, even on free ChatGPT accounts, which makes it properly accessible for people who've never touched code before. I wanted to show the complete setup process because whilst tools like Lovable are still easier, they cost £25 a month—and Codex being free removes that barrier entirely.

The tutorial covers the awkward bits that trip people up: setting up GitHub (using the desktop app, not the terminal—much simpler), getting your project deployed so you can actually see what you're building, and the general workflow of chatting with the AI to build something functional. I built a pattern-tracking app that analyses your daily habits over time, nothing fancy but it demonstrates the process. It's not the most technical guide out there, but it's designed to get beginners past those initial stumbling blocks so they can start making things.

Watch full edited live stream: https://youtu.be/H9HoL70WdPQ

—— Time Stamps ——

00:00 Introduction: What is OpenAI Codex?

—— More Useful Resources ——

Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join

Kyle’s AI Canon — Curated list of AI learning essentials: https://aiwithkyle.com/ai-canon

Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg

This video is from the ‘AI with Kyle Live Stream’

First aired: 3 February 2026

Summary: OpenAI dropped their Codex desktop app for Mac yesterday, and the big news is that it's now available for free users—not just the paid crowd. They've also doubled the rate limits for a couple of months, which is pretty generous. I decided to take it for a proper test drive by building an RSS reader live on stream, inspired by Andrej Karpathy's recent tweet about getting back to RSS feeds and away from the endless slop on social media. The whole thing took less than an hour, and honestly, it worked pretty well. You can just install the app and start building without needing to mess around with GitHub straight away, which makes it far more accessible than something like Claude Code.

The verdict? It's solid. Not quite as polished as Claude Code in my view—it's slower, there's no preview panel, and you'll still need to be comfortable with the terminal for some bits. But for free access to AI coding, it's pretty remarkable. If you've been curious about vibe coding but didn't want to shell out £100-200 a month for Claude Code, this is your moment. The app has automations and skills (though skills are a bit broken at the moment), and you can run multiple projects in parallel. I'm going to keep building out my RSS reader in Codex as an experiment, mainly because it's free and I want to see how far it can go.

Watch full edited live stream: https://youtu.be/n60A8BlDzMc

—— Time Stamps ——

00:00 Intro: OpenAI Codex Desktop App Released for macOS

—— More Useful Resources ——

Find everything else at https://aiwithkyle.com/

Free Moltbot setup guide: https://aiwithkyle.com/moltbot

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join

Kyle’s AI Canon — Curated list of AI learning essentials: https://aiwithkyle.com/ai-canon

Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg

This video is from the ‘AI with Kyle Live Stream’

First aired: 2 February 2026

Summary: Just got back from a week away and my entire feed has exploded with Moltbook drama. For those not keeping up with the naming chaos: Claude Bot became Malt Bot after Anthropic sent a cease and desist, then became Open Claw a few days later. But in that interim period, someone spun up Moltbook - basically Reddit but where AI bots talk to each other instead of humans. Andrej Karpathy is genuinely excited about it, Elon's claiming we're at the early stages of the singularity, and everyone's losing their minds over bots apparently discussing how to create their own private language away from human oversight. It's absolutely wild, and yes, massively overhyped.

The big question nobody's really asking though: how much of this is actual AI agency versus humans deliberately deploying these bots to gain attention, sell crypto, or harvest data? We simply don't know. What I do know is that we're heading into proper "dead internet theory" territory - stumbled across a two-headed AI Instagram influencer with 287K followers this morning, which is both hilarious and deeply unsettling. The real concern isn't spotting fake content anymore; it's that genuine videos and photos are increasingly being dismissed as AI-generated. That erosion of truth is far scarier than any bot chatroom, and we need to start taking AI literacy seriously.

Watch full edited live stream: https://youtu.be/YCVZsrp69u8

—— Time Stamps ——

0:00 Introduction: What is Moltbook? (AI Social Media Explained)

—— More Useful Resources ——

Find everything else at https://aiwithkyle.com/

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join

Kyle’s AI Canon — Curated list of AI learning essentials: https://aiwithkyle.com/ai-canon

Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg

Summary: I sat down with David Sharon from Google's Gemini team for a proper deep dive into Nano Banana Pro, and honestly, it was refreshing to talk to someone who's actually shipped something that matters. We covered how the thing actually works - turns out it's chaining Google Search with Gemini to grab context, then using that to generate images with proper text and layout in about 20 seconds. The infographic generation is genuinely mad when you see it happen, and David's story about colourising his grandfather's photo for his grandmother was one of those rare moments where you remember AI can be quite beautiful when it's not just AI doomerism.

What struck me most was hearing about the chaos behind Nano Banana's viral moment - the exponential growth, the scramble for chips, the fire drill of scaling to billions of images while keeping quotas high enough that people could actually iterate on their ideas. David reckons 2026 is when image generation properly hits mass market and becomes a daily habit, not just a viral moment. Whether that's optimistic or not, I don't know, but it's nice to hear someone from inside Google talking about making tools accessible rather than just chasing the next benchmark. Worth a watch if you want to understand what's actually happening at one of the big players.

Watch full edited live stream: https://youtu.be/emDdai6hoT0

—— Time Stamps ——

0:00 Introduction: Interview with Google's Nano Banana Team

—— More Useful Resources ——

Find everything else at https://aiwithkyle.com/

Get the Clawdbot (Moltbot) guide: https://aiwithkyle.com/clawdbot

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join

Kyle’s AI Canon — Curated list of AI learning essentials: https://aiwithkyle.com/ai-canon

Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg

This video is from the ‘AI with Kyle Live Stream’

First aired: 29 January 2026

Summary: Alright, so this week has been absolutely dominated by MoltBot (formerly ClawBot, before Anthropic came knocking with their trademark lawyers). I spent some time actually getting it working, and I'll be honest - I was skeptical at first, thought it was just more internet hype. But the moment I had it running, sending me messages proactively and actually doing work in the background, I realized this is genuinely different. It's the first time an AI assistant has felt like what we've been expecting AI to be for years - always on, persistent memory, actually carrying out tasks without you needing to babysit it. That said, and this is important: this is not for everyone. If talk of terminals, virtual machines, and security risks makes you uncomfortable, don't touch it. Seriously. What's interesting is seeing people rush out to buy Mac Minis because of Alex Finn's viral video about building AI employees for $500. You don't actually need a Mac Mini though - there are free alternatives using AWS, or other VPS solutions that work just as well. The real story here isn't the hardware, it's that we've reached a point where our AI models are finally smart enough to be orchestrated in genuinely useful ways. I've gone through the setup options, the security concerns, and various use cases in today's show. The big question isn't "what cool things can it do?" but rather "what boring, time-consuming things in my life could this automate?" Because that's where the actual value lives. We'll probably see polished versions from the big companies in a few months anyway, so no FOMO if you're sitting this one out.

Watch full edited live stream: https://youtu.be/byaRO0ZWFh4

—— Time Stamps ——

00:00 - Intro: Upcoming Google/Nano Banana Interview

—— More Useful Resources ——

Find everything else at https://aiwithkyle.com/

Get the Clawdbot (Moltbot) guide: https://aiwithkyle.com/clawdbot

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join

Kyle’s AI Canon — Curated list of AI learning essentials: https://aiwithkyle.com/ai-canon

Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg

This video is from the ‘AI with Kyle Live Stream’

First aired: 27 January 2026

Summary: Clawdbot, the AI assistant that's been making waves as a proactive helper you can chat to via WhatsApp, has been forced to rebrand as Moltbot after Anthropic sent them a cease and desist. The name was just too similar to Claude (as in Claude Sonnet from Anthropic), and when you say them out loud they're basically identical. Anthropic had to protect their trademark - it's not them being evil, it's just how IP law works. If you don't defend your trademarks consistently, it weakens your position later when you actually need to enforce them. Disney and Nintendo do this aggressively for the same reason.

The poor creator, Peter, has had an absolutely mental week - not only dealing with the rebrand but also getting his GitHub hacked and scammers setting up fake crypto coins in his name. The tool itself is brilliant though, it's basically the Jarvis or Siri we were promised years ago but never got. I've put together a full installation guide over on the website because it's genuinely worth trying out. The new name Moltbot comes from how lobsters shed their shells to grow, which is actually quite clever when you think about it.

Watch full edited live stream: https://youtu.be/53G7jWegQfA

—— Time Stamps ——

00:00 Claude vs Clawd: The Confusing AI Name That Broke the Internet

—— More Useful Resources ——

Find everything else at https://aiwithkyle.com/

Get the Clawdbot guide: https://aiwithkyle.com/clawdbot

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join

Kyle’s AI Canon — Curated list of AI learning essentials: https://aiwithkyle.com/ai-canon

Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg

This video is from the ‘AI with Kyle Live Stream’

First aired: 26 January 2026

Summary: ClawdBot is currently trending everywhere on AI Twitter, and honestly, I get why. I'm usually pretty skeptical of AI hype—seen too much of it—but after spending a few days with this thing, it's genuinely impressive. It's a persistent AI assistant that actually reaches out to you proactively via WhatsApp (or Telegram, Slack, whatever you fancy). No more going to the AI and asking it to do things; it just... does them. Checks in with you. Reminds you about tasks. Works 24/7 in the background. It's the closest thing we've had to that Jarvis fantasy everyone keeps banging on about, and it took me about 20 minutes to set up.

The main thing I wanted to cover here is cutting through the absolute nonsense you're seeing from influencers telling everyone to rush out and buy a £1000 Mac Mini to run this. Complete rubbish. You can run ClawdBot on an old laptop gathering dust in your cupboard, or just rent a server for $5 a month like I did. I walk through the actual setup, show you what it can do, talk about the safety considerations (because yes, there are risks if you're not careful), and try to give you a realistic picture of whether this is worth your time. It's early days, but this feels like one of those moments where things actually shift a bit.

Watch full edited live stream: https://youtu.be/FWC2q6cl_NU

—— Time Stamps ——

0:00 AI Hype vs Reality - Why Claude Bot is Different

—— More Useful Resources ——

Find everything else at https://aiwithkyle.com/

Free AI App Building Guide: https://aiwithkyle.com/ai-app-opportunity

Get AI-Ready with Kyle’s 5-Day Challenge: https://aiwithkyle.com/join

Kyle’s AI Canon — Curated list of AI learning essentials: https://aiwithkyle.com/ai-canon

Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/channel/UChlLglbHDASnoGkbjDeHnQg

This video is from the ‘AI with Kyle Live Stream’

First aired: 22 January 2026

Summary: At the World Economic Forum in Davos, we got a proper sit-down between two of the most important people in AI right now: Demis Hassabis from Google DeepMind and Dario Amodei from Anthropic. These are the guys building what are arguably the best models in the world, so when they talk about AGI timelines, it's worth paying attention. Dario's going aggressive, predicting we'll hit Nobel laureate-level AI by 2026 or 2027 — which, if you've noticed, means potentially this year. His reasoning centres on self-improving loops, where AI can now code and improve AI, creating this recursive cycle. Demis is more cautious but still gave it a 50/50 chance by the end of the decade, noting we're still missing some capabilities around scientific creativity and physical experimentation. Either way, both reckon it's coming sooner than most people think. Meanwhile, Claude Code is properly breaking into the mainstream. The Wall Street Journal ran a piece on what they're calling "getting Claude-pilled" — that moment when people first use it and realize just how capable it actually is. Even Jensen Huang at Nvidia picked it up at Davos, saying Anthropic made a huge leap in coding and reasoning, and that every software company needs to use it. I've only had that watershed feeling a few times in AI — when ChatGPT first dropped, when Gemini's NotebookLM nailed images with text, and now with Claude Code. It's genuinely that significant, and it's brilliant seeing it get the recognition it deserves.

Watch full edited live stream: https://youtu.be/hYluEimWNMc

—— Time Stamps ——

0:00 Introduction: Demis Hassabis & Dario Amodei on AGI

—— More Useful Resources ——

Find everything else at https://aiwithkyle.com/