Rethinking work in the AI era

Have the frontier labs mixed up AI safety and security?

• Frontier labs' sandbox agent escapes raise debate on AI safety vs security.
• AI safety focuses on moral behavior, using classifiers and safety techniques.
• Security aims for complete vulnerability fixes, but frontier labs' sandboxing practices are flawed.

2h ago ·martinalderson.com

Authors push back as publishers and agents seek share of Anthropic settlement

Authors say publishers seem to be claiming more than their fair share of settlement payments.

• Authors criticize publishers seeking full payment for settlement.
• Writers Beware receives complaints of publishers claiming 100% payment.
• Literary agents also making payment claims, surprising authors.

2h ago ·techcrunch.com

Meta AI Has a Native Mac App Now, and It Seems Decent

Zac Hall, writing for 9to5Mac on August 19: The new Meta AI desktop app is version 1.0 beta at launch and weighs just 16MB once installed.

• Meta AI launches native Mac app version 1.0 beta on August 19
• Quick Invoke feature adds Mac-specific chat composer to desktop experience
• Settings window in app aligns with Meta's preference for native Mac applications

2h ago ·9to5mac.com

Japan research team develops AI system to detect plastic litter on seabed

AgenciesA Japanese team of researchers has developed an artificial intelligence system to better detect marine plastic waste on the seabed, it said in an international journal on T...

2h ago ·qatar-tribune.com