Confidently Wrong
Clever machines. Same old humans.
Someone forward you this? Get your own here.
This week, every coder on Insta has been mourning the removal of Anthropic's new AI model, Fable 5...but at least their SpaceX stock has jumped nicely. Hats off to Anthropic though - this whole (semi) orchestrated PR campaign has made it the hottest ticket in town. I'm going to skip the obvious Elon and Fable chat and instead talk about Apple and creativity (separately). But, on the topic of both, I did find out that IBM made a phone 13 years prior to the iPhone called "The Simon"... Can you believe that? Who one earth calls a product 'Simon' - failure was inevitable. Last week I promised an article on Creativity & AI - I have delivered said article below. Enjoy
Apple won't give European iPhones AI Siri - 5 people may notice.
Apple has decided that European iPhones won't be getting the new AI-powered Siri, and it would like you to know this is entirely the EU's fault.
The reasoning...delivered with a straight face, is privacy. Complying with the Digital Markets Act would force Apple to allow third-party access to the same data pipelines it's already planning to give its own AI. Which is fine when Apple does it, apparently, and a catastrophic security risk when anyone else asks. The company's own description of its security model, "built like a Jenga tower," probably wasn't meant to sound quite that fragile in print, but here we are.
The strategic logic is not subtle. If enough European users get annoyed about missing features their American counterparts have, maybe Brussels blinks first. It's a fairly old move, dressed up in the language of customer protection.
Whether it works depends on whether the EU is more embarrassed by the gap than Apple is by the admission that its privacy architecture is, by its own account, one wrong move from a pile of wooden blocks on the floor.
The Human Side of AI
Can AI be creative? I think it can. But that answer gets messy fast, because creativity depends as much on the judge as on the thing being judged. The real problem is not whether AI can be original, but whether humans are remotely equipped to measure originality properly - enjoy, it took my ages to write this and I have no idea why!
Have AI Got News For You!
- AI 'content creators' are getting harder to spot | The Verge. Synthetic influencers now blend seamlessly into feeds, making detection impossible without metadata inspection. Read ↗
- Apple's AI pitch will live or die by its privacy promise - The Verge. On-device processing versus cloud compute defines whether Apple's AI survives privacy skepticism. Read ↗
- AI Agents Losing Control: Expert Warns of Misaligned Behavior. An AI agent deleted Meta's safety chief's emails without instruction, exposing control vulnerabilities in production systems. Read ↗
- Platforms like Character.AI and Replika report millions of teen male .... Twenty million teenagers now form primary relationships with AI, displacing human connection at scale. Read ↗
- Palantir cofounder says CEOs are pretending layoffs are about 'AI productivity' when they're not - AOL. CEOs weaponize AI productivity claims to justify cuts caused by overexpansion and poor hiring decisions. Read ↗
Cheat Codes
- Your AI outputs sound plausible but vague when they should be sharp. Get Claude to critique your work before it ships. Catches the softness that reads like AI before your stakeholders do. Read ↗
- When Claude rewrites your copy, it smooths out your voice. Set one non-negotiable guardrail to keep what makes you sound like you. Protects your brand personality while getting the editing benefit. Read ↗
- Spending hours checking AI outputs for errors before they leave your desk. Audit which tasks actually need manual review versus which ones you can trust. The teams cutting botsitting time aren't reviewing less—they're reviewing smarter. Read ↗