In this context, normal people will believe the marketing and trust meta, and caveat emptor is a cop-out at best. I can sell you a basket of bread, and it’s privacy-preserving bread, but it will also punch you in the face if you don’t read every bit of the agreement. No one reads the agreement, and that’s what Meta runs on, plus skirting the law in every way they can possibly get away with.

In this context, Apple provided a feature that gives applications Full Disk Access. The journalist enabled it, and then had their face eaten by leopards after ignoring the warning.

Take Meta out of the equation here; any agent with FDA can do the exact same thing. Claude, Codex, DeepSeek, any of them. This is why Apple's response is a fix to their own software. The fix is a mea-culpa, Apple would not have to patch their OS if it was behaving exactly the way they wanted it to. Apple and the journalist made the biggest mistakes here.

Meta is a godawful company that should be regulated into the dirt until Zuck is left with nothing. Guess what? They're not to blame in this scenario, and your rabid attacks on lapcat (who is a reputable and generally impartial macOS developer) is unnecessarily hostile towards a perfectly normal observation. This brand of comment is so misleading, low-effort and harmful to good-faith discussion that I'm tempted to flag this whole thread for being founded on a misunderstanding. Your response has contributed to the derailing of this conversation by tribalist "Apple vs Meta" pundits who are drowning out a well-known and respected expert that has technically-salient architectural details to share. HN cannot foster intellectual gratification under these conditions.