- AI with Kyle
- Posts
- AI with Kyle Daily Update 022
AI with Kyle Daily Update 022
Today in AI: Public Figures get AI With-Hunted, Norman Reedus fools Youtube, Anthropic cut off OpenAI's Claude Code Access
Kyle Balmer
August 05, 2025
The skinny on what's happening in AI - straight from the previous live session:
Highlights
🎯 Public Figures Get AI Witch-Hunted
Two very different public figures got absolutely roasted for using ChatGPT. Sweden's Prime Minister Ulf Kristersson admitted he uses it for "second opinions" in government decision-making, whilst gaming YouTuber Markiplier used it in a comedy skit about almond farming.
Both faced massive backlash - Swedish voters calling their PM an "embarrassment" and Markiplier's artist fanbase accusing him of being "irresponsible."
Kyle's take: This is the new normal - public witch hunts for anyone honest about AI usage.
Sixty percent of Americans use ChatGPT, including politicians and CEOs, but most keep quiet still!
The Prime Minister wasn't uploading state secrets, Markiplier was literally taking the piss out of ChatGPT.
But as AI starts taking jobs, we'll see more anger against anyone who dares admit they use it. The real hypocrisy? Most of the people screaming about it are probably using it themselves….just quietly
Source: Swedish news coverage / YouTube drama
🎬 YouTube's AI Crackdown Gets Death Stranding'd
YouTube's rolling out AI age verification that can override what age you tell it based on your viewing habits. So you can’t just say “I’m over 18 chill…”
Watch too much Roblox content? Sorry, you'll get flagged as under-18 regardless of your actual age!
But people have already found the perfect workaround - Death Stranding 2’s photo mode. The photorealistic Norman Reedus character (Sam Bridges) can perform all the facial movements these systems require, completely fooling the verification.
Kyle's take: This is brilliant. Teenagers will always find ways around these systems - they've got time, motivation, and share solutions online. Throw any barriers in their way and they’ll get through.
Meanwhile, us silly adults are uploading passport photos to random third-party companies that'll sell our data and probably get end up getting hacked. The whole "protect the children" angle is just cover for increased online control.
And Norman Reedus/Sam Bridges inadvertently becoming the face of digital resistance? Absolutely mental.
Source: The Verge
🤖 OpenAI Gets Banned from Claude for Code Abuse
Anthropic cut off OpenAI's access to Claude Code after they violated terms of service by using it to build competing models - presumably GPT-5. OpenAI were among the heaviest users, racking up tens of thousands in model costs on $200/month plans by running it continuously…
Kyle's take: Even OpenAI thinks Claude Code is better than their own coding tools. I mean…it is! At least until GPT 5 in a couple of weeks?
That's a proper endorsement right there.
Getting caught and banned is very funny. It’s also standard practice - you can't use someone's AI to train your own AI. The real story is OpenAI developers clearly prefer Claude for coding work…
Member Question from Mark: "If someone can get ChatGPT to teach anything, how do you get people to buy your courses and programs?"
Kyle's response: Information has been free for a long time - MIT publishes entire computer science degrees online and no-one does them. Udemy completion rate is 5%.
The problem isn't access to information, it's action. If anything there’s too much information!
People need structure, accountability, community, and someone to hold them to actually doing the work. Online courses have had dismal completion rates for years, and that's not changing with AI. Focus not on selling information - but instead on frameworks that get your customers actually using what you teach.
This question was discussed at [55:00] during the live session
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.