Post

The type of programs malware and virus writers make are now being treated as though they're intelligence that respects requests and stay within limits.

Meta? Privacy? lol.

A permission prompt should name the data source and scope, then the agent should leave a receipt of what it actually accessed. “Notification banners” and “local Messages database” create wildly different expectations, whatever the disputed permission state turns out to be.

AI reflects the culture that produced it, one that really does not respect others boundaries

Zuckerberg's career began with software called Facemash that callously violated the privacy of Harvard students by encouraging viewers to rank students by their appearance, without their knowledge or consent. Nothing has changed. He makes $billions by exploiting private info and monetizing it.

I was very suspicious as soon as the “Meta released an AI and it’s actually really good!” chatter started and it is kind of satisfying to see that I was right to be suspicious lol