AI with Kyle Daily Update 013
UK's Pathetic AI Deal, EU Kills AI Startups, Baby Grok is Born
Watch this video on our dedicated watch page
Better viewing experience with related videos and full-screen player
The skinny on what's happening in AI - straight from the previous live session:
Highlights
🇬🇧 UK-OpenAI Deal: All Talk, No Substance
The UK government and OpenAI signed a "commitment" to use AI in public services.
OK? Sounds impressive until you realise it's a statement of intent, not a legally binding deal. But it’s still front page news?
Kyle's take: This is pure theatre. The UK economy is basically dead in the water - 0.1% growth is a rounding error. They're making massive noise about AI whilst doing absolutely nothing substantial.
I was at Google yesterday talking to entrepreneurs, and people are planning to bugger off to San Francisco because that's where the money actually is. We had the best AI talent with DeepMind, lost it all to America, and now we're pretending a staff of 100 people in London matters when Meta gives engineers $100 million bonuses over there.
Source: BBC News coverage
👶 Baby Grok: Elon's Kid-Friendly AI Experiment
Elon Musk announced Baby Grok - a standalone app for child-friendly AI content. This comes after their current AI avatar Ani is absolutely not safe for work, with sexually explicit conversations available. Baby Grok will supposedly have proper safeguards for children, potentially acting as a tutor and educational tool.
Kyle's take: I'm not convinced XAI and Elon are the best people to be bringing AI into the lives of the world's children, but they're the ones stepping up.
This could be brilliant - AI tutors that were previously only available to wealthy kids. But we're about to have the same debates we had about TV time, internet access, and smartphones all over again. When is the right age to give a child access to AI? Parents will have to work this out themselves, just like every other technology.
Source: Elon Musk X announcement
🇪🇺 EU Shoots Itself in Both Feet
Meta refused to sign the EU's AI Code of Practice, calling it legally uncertain and going beyond the EU AI Act scope. Meanwhile, the EU announced €20 billion for four AI gigafactories with 100,000 GPUs each. Problem is, that's already the size of existing American factories like Colossus. By the time the EU builds these in a couple of years, America will have five-gigawatt factories the size of Manhattan…
Kyle's take: The EU has completely removed itself from the AI race. They spent two years drafting regulations for technology that's changed massively since then. Now they're announcing they'll build what America has today but in two years' time. It's mental. Even Mistral, the leading EU AI company, is telling Brussels to back off. The uncertainty has driven all serious AI startups to San Francisco, and there's no catching up now. It's a two-horse race between America and China, and Europe isn't even at the starting line.
Source: WSJ coverage
Member Question from Ben: "Do I prefer O3 or Grok 4?"
Kyle's response: O3 for me, absolutely. xAI has done impressive work - going from zero to frontier model in just over two years is mental. But nobody I know actually uses Grok day-to-day.
Most people are still building with OpenAI, Claude, and Gemini. Grok is interesting but not particularly useful yet. For reasoning tasks, O3 wins hands down. Although that said I'm a Claude maximalist for most things, except coding where I use Gemini 2.5 Pro.
This question was discussed at [06:04] during the live session.
Member Question from ABC with Teacher Kim: "Will there be any apprenticeships in AI for London?"
Kyle's response: The scale of AI projects in the UK is tiny. Meta gave the UK government $1 million whilst they're signing $100 million bonuses for American engineers. They probably spend more on toilet paper than they gave our government. If there are apprenticeships, they'll be 30-50 people with big press releases and award ceremonies, but no meaningful impact. I hope I'm wrong, but so far everything AI-related here has been lip service with minimal scale…
Sorry!
This question was discussed at [11:50] 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 for the daily AI news live stream where we dig into the stories and you can ask questions directly.