Interviews, insight, intelligence, and exclusive events for digital leaders.
Preparing for "a future where agents act on users’ behalf and need secure, scoped, and revocable access."
"Prompt injection attacks have become ... a systematic, category-wide vulnerability class," according to Noma Security.
“You get a good engineer and you also arm them with an LLM and Rust, you can vomit out really good code that just doesn't break”.
Future changes will be "fully and transparently opt‑in" says HubSpot after holiday weekend outrage.
"AI concentrates some risks at system level through shared models, infrastructure and cross-border dependencies..."
Last week, Sysinternals creator Mark Russinovich started experimenting with agentic coding. Now Process Explorer for Mac will be available within weeks, he tells us.
Vendors and users – and rest-of-world countries – need to take AI risks way more seriously, the 4,000 people hoping to chart its future hear in Geneva.
