AI Reading for Saturday June 14

Hiring sprees for AI geniuses, early retirement for everyone else?

Management talks even scarier about it non-publicly, they got AI fever and the only cure is large scale AI-driven headcount reductions.

A lot of robotic systems will fail spectacularly in a truly adversarial environment. A smart adaptive attacker can generally get inside its (nonexistent) OODA loop and figure out an adversarial way to make it defeat itself. Success kind of depends on society using laws and enforcement to provide very orderly environments.

Robotaxis gonna have a tough time in New Delhi, might work in NYC with well protected infrastructure and enforcement. People will just walk in front of them if they know they will always stop, or otherwise screw with them, unless it can e.g. take their picture and call the cops.

Follow the latest AI headlines via SkynetAndChill.com on Bluesky