1/ The UK wants to ban under-16s from certain social media platforms from Spring 2027.
Protecting kids online is the easy part to agree on.
The hard part: how do you prove someone is under 16 without turning age checks into identity checks for everyone?
5/ Child safety rules need privacy-by-design from day one.
No giant ID databases.
No casual face-scan normalization.
No “trust us bro” compliance theater.
Protect kids, yes. But don’t accidentally build the internet’s bouncer layer for everyone else.
What’s the right balance?
1/ Anthropic just gave us a preview of the next AI fight:
not “which model is smarter?”
but “who is allowed to use the smartest models at all?”
On June 9, Anthropic launched Claude Fable 5 and Claude Mythos 5.
7/ My take: this is where the “just ship better models” era gets complicated.
It’ll be access rules, trust programs, government pressure, and who gets left outside the gate.
What do you think: necessary guardrails or the start of AI gatekeeping?