It's the em-dashes from a green account.

Meta spent 100 billion dollars on VR, what's a Zuckerbot or two?

You don't have to be irrational to not know things.

True, but isn't it irrational to continue operating something you know could cause harm to you when used wrongly, despite not knowing how to use it correctly?

The hypothetical person we're considering does have an entire life, too. Their rationale may have emerged from careful risk analysis and weighing of opportunity costs.

Even people who know it, don't think about it and don't connect it with the potential consequences of uploading a picture to a website. And why would they? It's not visible, there's no warning, it's just not something that's going to be top of mind.

So we should educate people about it. Don't you think that constantly coddling people about tech just breeds tech-illiterate people?

Wouldn't it be better if people were more tech-literate?

Coddling only works when those who are in charge of the tech play nice. But then breeds people who will more easily fall victim to the bad actors.

Agree. That's also the dilemma with asking the user for his permission, it is very difficult to frame a concise question and get an educated decision there. So, better to only ask if the App explicitly requests that permission sounds reasonable.

The prior threat-model was, that e.g. a camera/gallery app which may/may not have a permission to a users current location, also has access to the history of a users' locations just by scanning the images when showing the camera roll.

It frankly makes sense to create a separate permission just for this location metadata AND strip this data when no permission was granted, I believe everything else would be MUCH harder to explain the user...

I assume Google are very hesitant to add additional permissions, and any additions get very carefully thought about. Having too many prompts can lead to popup blindness, which defeats the entire purposr of the permission system in the first place.

I'm sure I recall much older Android versions presenting all of the app's permissions at install-time. I'm very willing to bet that most users didn't actually read any of it. Overall, it seems like a very interesting problem to solve.

yeah I don't hate LLM docs if they're labeled as such. but if someone wants me to use their code or read their README.md they are going to have to make it sound like a human cared about writing it, and right now Claude can't do that

If they're that hard to prompt maybe it's easier just to write the blog posts yourself.

They could write a post demonstrating that you can do that and surface the same bugs in the same codebases.

It would be way more informative than this one, which didn't do that.

"small models can do this if you scaffold them right" might be true, but it wasn't actually demonstrated in the post.

They could have linked their replication in this blog post, which we did all see, if they have one.

Right, but they didn't actually test that, did they?