The skinny on what's happening in AI - straight from the previous live session:
Highlights
🚀 Grok 4 Launches: Best Model Yet, But at What Cost?
Discussed at 01:10
xAI launched Grok 4 yesterday and the model itself is genuinely impressive. Especially because xAI are new to the party.
Problem is, it comes with baggage: earlier this week Grok 3 called itself "MechaHitler" and spewed anti-Semitic content before they pulled it offline. Now they've released Grok 4 with zero explanation, no model card, no post-mortem on what went wrong.
Leather jackets are now required by AI execs apparently
Kyle's take: The tech is solid, but the lack of transparency is mental. No serious business will touch this with a bargepole. Why risk using a model that could turn into a Nazi bot when you could use good old reliable ChatGPT or Claude?
It's like that saying "nobody ever got fired for buying IBM" - enterprises need safe, predictable options.
xAI have also launched a £300/month "Super Grok Heavy" subscription but even most heavy AI users (like me) don't need the $200/month tiers for Claude and ChatGPT, so who's this for?
We’ll see I suppose! But Grok 4 feels like a footnote that’ll be quickly forgotten when the big boys release GPT-5.
Source: xAI launch and artificial analysis benchmarks
💰 Nvidia Hits $4 Trillion - The Accidental AI Kings?
Discussed at 22:30
Nvidia just became the first company ever to hit $4 trillion market cap. Mental when you think they used to just make graphics cards for gamers. Turns out the parallel processing that renders Crysis is exactly what's needed to train AI models. It's a weird historical coincidence that's just made Jensen Huang's company the most valuable on Earth.
Kyle's take: This shows how much money is being thrown at AI infrastructure right now. Every time someone like Elon builds a massive data centre, it's stuffed with Nvidia chips. They're basically the only game in town for AI training, which explains the astronomical valuation.
If AI is a goldrush it’s NVIDIA selling the shovels. For anyone who's been holding Nvidia stock - well done, you've probably made a fortune (not financial advice, obviously!!).
📱 Meta Poaches Apple's AI Chief - Kicking Them While They're Down
Discussed at 31:35
Meta just poached Ruoming Pang, a senior at Apple Intelligence with a 100-strong team of engineers. He's now joining Meta's "super intelligence lab" (everyone's got dramatic names these days). This feels particularly brutal because Apple is already miles behind in AI - Siri still can't set an egg timer properly after 13+ years.
Kyle's take: This seems mean. Apple are already so far behind. Apple had such an easy win with upgrading Siri to proper AI, but they've fumbled it. As someone surrounded by Apple kit, it's painful watching them fall so far behind. Meta are apparently throwing $100 million signing bonuses around to hoover up talent, so this won't be the last poaching we see. The AI talent war is getting vicious.
Source: AI News coverage of poaching
❓️ Member Question
From Blench Photography: "We'd like to start using agentic AI to improve our customer experience. What service is best?"
Kyle's response: First start with the problem you're trying to solve, not the sexy buzzword.
"Agentic AI" is often just marketing speak. You might solve your issue with normal (non-agentic) AI, or even traditional automation without AI at all. That said, if you do need something more sophisticated, n8n is your best starting point. It does both simple automations and more complex agentic systems, so you can start simple and expand if needed. Just don’t jump straight to the complicated stuff - it's got more ways to go horribly wrong!
This question was discussed at 10:31 during the live session.
Want to submit a question? Drop it below this video and I'll cover it in a future live.
Want the full unfiltered discussion? Join me tomorrow for the daily AI news live stream where we dig into the stories and you can ask questions directly.