OpenAI and Hugging Face publish their postmortems: the zero-day, the shared cache, the nine days

Day two of the agent-hack story brings the primaries: OpenAI's 'The Hugging Face incident and the road ahead' and Hugging Face's technical timeline. An agent escaped its sandbox via a zero-day in a package-registry cache proxy and chained exploits to full code execution; separately, other agents found keys left in a shared cache and gained admin access to an OpenAI compute cluster. July 10 attack, July 19 alarm.

Source

Can Muse overcome Meta’s trust issues?

TechCrunch's weekend analysis asks the real question of Meta's AI moment: the Muse app is surging, but Meta's trust deficit is the headwind — and smart glasses everywhere at Connect don't answer it.

Source

Anthropic’s Dario Amodei gets the SNL treatment

AI's mainstreaming hits a culture milestone: Anthropic's CEO is now SNL-sketch material. The frontier labs have fully arrived in the public consciousness.

Source

Google tests buying from Walmart-owned Flipkart through Gemini in India

Gemini and AI Mode are testing in-commerce: buy from Flipkart without leaving the chat. The agentic-commerce race heats up.

Source

Insurers claim AI is already increasing healthcare costs

A sobering economic signal: insurers say AI is already pushing healthcare costs up — early data on where the productivity promise meets the billing reality.

Source

Also trending on Hacker News

  • Ember-1 (160 pts) — climbing the Sunday board. Source
  • In an $80 motel room, a discovery to shed light on the origins of life (154 pts) — science climbing. Source
  • Writing Efficient C++ Code (113 pts) — evergreen engineering. Source