I love the internet archive but yeah, there was just no way this wasn’t going to backfire. And by handling things the way they did they damaged the reasonable defense of archivist (not only for themselves) because publishers and others often cite that archival and backups are just “pseudonyms” “synonymous” for piracy.

They aren’t but the way this was handled made it impossible for them to argue otherwise and it also creates a legal precedent for lawsuits and judgments by publishers against others who are doing such work.

It’s a fine line. If they’re working on them reporting the issue before it’s resolved increases the risk somebody can use this as a kind of todo list of social and technical engineering weakpoints to get at other user data.

On the one hand that’s ridiculous but it’s nice to see that they didn’t throw the prior owner under the bus or give out their info.

Offloading the data to the cloud and making it accessible on other devices no longer signed into iCloud.

That is so much worse if true.

Good point.

I wonder if they’re doing that to reduce the write cycles on the cells and since they’re “encrypting” the contents of the cells they figure the overall IO flag of the data being deleted is “good enough”.

So, in a perfect world, when you wipe the phone it’s basically just trashing the encryption key and so it’s useless data.

That’s all assuming that the encryption method/keys are foolproof which is always a bad bet.

And, this here makes me wonder how effective that is.

And a person claimed in a later post that “around 300” of their old pictures, some of which were “revealing,” appeared on an iPad they’d wiped per Apple’s guidelines and sold to a friend.

That’s a huge issue. Not just for photos but also files for sensitive data, secrets, etc. this, if true, is a massive issue overall since it even happening at all shouldn’t be possible.

Here’s the ELI5.

Imagine there’s a set of lockers in a school.

When a student leaves the school or changes lockers they remove the label on the locker but don’t empty it.

A TRIM, however, means that they not only remove the label from the locker by also clean out its contents.

Exactly. Even for mobile use besides accessing your home resources you can avoid your cellular provider monitoring/hijacking your traffic.

Of course self hosting means you’re still sending that info from your home network over your ISP.

So it’s a trade off there but depending on your ISP vs your cellular network makes sense.

What happens when you take the output and make it the new input with “more passionate”?

E.g. how far can you push it? Or even the other way.

That’s what I’ve been playing with. Cool stuff even though it’s limited because of my 8GB nvidia card

There’s a real risk of survivorship bias here. Somebody asking about a car gets that and thinks nothing of it. A privacy minded person, however, would find it odd. And being the kind of person concerned about what could have been the cause considered the prior conversation.

I’m not saying its an unreasonable concern or technically not feasible. It’s just not how the LLM’s tend to work.

Id consider it more likely to be a bug, or general inquiries like you said, or that SO had a bunch of documents locally that reference privacy or browsing history (anytime really) that MS could have used as a kind of “here’s more about the person asking you a question”

Oogabooga makes running one locally pretty easy. At least for me with an 8GB nvidia GPU so models like llama and such fit.