AI with Kyle Daily Update 007
Grok 4 Consults Elon Before Answering + UK's $1M AI Embarrassment
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
🤖 Grok 4 Launches With Trust Problem
Elon's xAI just dropped Grok 4, and the benchmarks are solid. BUT there are been some launch week controversies.
The latest: it's automatically searching Twitter for Elon's opinions before answering controversial questions. Ask about Israel-Palestine, and it literally checks what Elon tweeted first.
Grok 4 decides what it thinks about Israel/Palestine by searching for Elon's thoughts. Not a confidence booster in "maximally truth seeking" behavior. h/t @catehall. Screenshots are mine.
— Ramez Naam (@ramez)
10:04 PM • Jul 10, 2025
Kyle's take: Yikes. This is proper dystopian stuff. Having one person’s (ANY person1) opinions shape an AI's worldview is terrifying. It’s the complete opposite of alignment.
No business in their right mind will build this into their systems - you're basically embedding Elon Musk into your customer service chatbot. The benchmarks don't matter if nobody trusts it.
Source: Techcrunch
🏥 NHS Doctors Using Unapproved AI Tools
Discussed at 16:08
UK doctors are using AI to transcribe patient meetings without approval, breaching data protection rules. The NHS is telling them off instead of providing proper solutions. GPs only get 10 minutes per patient - anything that gives them more face-to-face time should be welcomed…but needs to be done safely and whilst protecting the patient.
Kyle's take: This is shadow IT at its finest. Staff realise AI makes their jobs massively better, but leadership is stuck in bureaucracy mode. The NHS needs to get its act together and provide compliant tools instead of just saying "don't do that." And that needs to happen fast or otherwise medics will use the tools covertly. Much worse.
FYI: There's a huge opportunity here for anyone who understands both healthcare regulations and AI - these industries need consultants who can bridge that gap.
Source: Sky News
🇬🇧 UK Government's £1M AI "Investment" Is Embarrassing
Discussed at 22:07
Meta just gave the UK government £1 million for ten AI fellowships as part of their "massive AI adoption scheme." Meanwhile, Meta is dropping $100 million signing bonuses poaching engineers from other companies.
Kyle's take: This is embarrassing. Meta spends $100++ million on individual engineering talent then tosses the UK government pocket change and calls it an AI partnership. The fact that this made headlines shows how far behind the UK has fallen. We used to have DeepMind before Google purchased it - now we're celebrating corporate charity. The UK isn't even a player in this race anymore.
@iamkylebalmer Meta is donated $1M to the U.K. government to help with AI. And this is headline news? Meta have paid engineers $100M over the past few we... See more
🇨🇳 Nvidia's Creative China Workaround
Discussed at 25:49
Nvidia's planning a downgraded chip for China to bypass US tariffs, removing high-bandwidth memory and connection features. Plus that suspicious 28% of revenue going to Singapore - a tiny country that somehow needs massive amounts of AI chips.
Kyle's take: They're playing the long game here. Sell "weaker" chips to China, but maybe upgradeable ones? Or just sell ten times as many to make up the difference. That Singapore revenue is definitely, hmm, interesting- either they're running cloud services for Chinese companies or it's creative distribution via a middleman. Either way, Nvidia wants that Chinese market badly enough to risk annoying the US government. And now that it’s the largest company on earth maybe they are ready to throw some weight around.
Source: New chip to circumvent tariffs
📱 YouTube Kills AI Slop Monetisation
Discussed at 39:15
Starting July 15th, YouTube won't let you monetise "mass-produced and repetitive content" - basically all that AI-generated rubbish flooding the platform. Those weird Bigfoot and flying dog videos that somehow get millions of views? No more ad revenue for them.
Kyle's take: Good. YouTube's algorithm was being gamed by people pumping out hundreds of AI videos hoping one would go viral. This should clean up the platform and help real creators. Hopefully Instagram and TikTok follow suit - the internet's drowning in AI slop right now. And I say this as a pro-AI person.
Source: YouTube Partner Program updates
Member Question
From Paul: "You don't sound technical at all. Apparently you have a history degree."
Kyle's response: Yup! Ya got me! 😂 I'm not technical and nor do I pretend to be.
I've got a history degree and an MBA, not a computer science PhD.
But here's the thing: most people using AI aren't engineers either. Hell, even Google is 47% non-technical staff, and they need the same help as everyone else. My job isn't to build models - it's to explain to normal people why AI matters for their business without scaring them off with gradient descent and back propagation.
The world has enough technical experts who can't communicate with humans. What we need are people who can bridge that gap and help businesses actually use this stuff properly.
This, by the way is 1000% what I recommend my students do. Carve out a business niche for yourself teaching about AI in your industry. You speak the language and are far better positioned to assist.
This question was discussed on TikTok:
@iamkylebalmer I’m not an AI technical expert. Just got called out for this on a Live i did - but i’ve never purported to be. we don’t need more of thos... See more
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.