Friday · June 6, 2026 · ~5 min
Morning Edition
Day 47 of your magazine
Today, in AI safety — three things that mattered. The EU
moved on enforcement, Anthropic shipped a new interpretability paper, and a quiet funding
round hints at where the next wave is heading.
The Lead
EU AI Act: first enforcement actions land — what changes for foundation model providers
Reuters · 2h ago
The European Commission's AI Office issued its first binding guidance under Article 51…
Read original →
Anthropic's new interpretability paper on attention heads
Anthropic · 5h ago
The team shows a method for decomposing circuit-level behaviour in larger models…
Read original →
A $1.2B Series B lands for an open-weights lab
The Information · 8h ago
Investors are doubling down on post-training as the new moat…
Read original →
UK CMA opens review into AI partnerships
FT · 1d ago
The Competition and Markets Authority wants to understand cloud-model integrations…
Read original →