That's a great description of the boundary between logical deduction NLP and bullshitting NLP.

I still have hope for the former. In fact, I think I might have figured out how to make it happen. Of course, if it works, the result won't be stubborn and monotone..

The first premise you are avoiding is that a child can misuse a phone.

The second premise you are avoiding is that the government can define, for every child, what constitutes misuse.

You are advocating thought crime. You do not have my support.

My government cannot adequately manage responsibility for my cupboards. It therefore shall not have authority over them.

Your government does have various authorities over what you put in your cupboards though. like, you can't just put a gun in there (actually I don't know where you live but that's true for most countries). You can't just get in a car.

Anyway, ultimately it's best effort. No security is flawless, but if it stops 99% or more of cases it's better than 0%.

Cases of what? Beer in an unlocked cupboard? Porn on an unverified computer?

Do you also refuse to show id when buying alcohol because the gubbernment does not have authority over what you may buy?

That's how you sound.

No, because that's a public store. The government can go to the store. They can't go to my cupboards without a warrant. The same goes for my computer, and its connection to another computer.

I replied to the content of the article and HN comments, not what you think I should have replied to. If anything you even failed to notice that I expect parents to do some of the parenting and not expect an app to magically do it all for them.

The government already defines what misuse is both for children and adults, defines responsibility for a lot of things even in your cupboard, and has been doing so for as governments have been a thing. And I don’t think you understand what “thought crime” is.

You won’t hear me say this too often but next time use an LLM to write your comments, any LLM will do, can only get better.

Why would I want to write better? This is a comment on a website.

You replied to a subset of the topic, and that's the point I was making. I felt the conversation needed relevant details from outside that subset, so I provided them.

I was terse in my comment, because that's how I like comments: short and to the point. That makes them much easier to skim through.

The government doesn't enforce its rules by going through my cupboards. It doesn't put a lock on them. Instead, it tells me what the rules and consequences are, placing both authority and responsibility for the cupboards themselves into my hands.

This is the primary change we are taking about: allowing the government to introduce its own code (lock) into my private digital interactions. Why are you so intent on focusing the conversation on the mechanics of that lock? Is it really so unreasonable for me to ask you to think about the rest of the topic?

The surrounding context is that. Why else would you participate with a government in an age verification system?

Maybe your argument is that it's not a surveillance state because it is implemented with a 0 knowledge proof. Sure, the age verification is, but that is only part of the system we are talking about. The rest of the system is the demand that every adult play keep-away with their verification, and every host on the internet (that can be adequately threatened) play, too.

The only way for this to be anything else is if every participant can individually decide what should and should not be kept away from children. Such a premise is fundamentally incompatible.

Misalignment of trust is a category of bug in my book; right next to logic error.

You are giving those people the benefit of the doubt. It's been proven many many times that police will use "more information" to excuse their own decision to use violence. A decision that they already made well before the incident.

Title is missing the number 13, which makes it much harder to parse. Should be "Fedware: 13 Government Apps That Spy Harder Than the Apps They Ban"

I don't see how this provides any useful context.

The information for where nerve endings are likely to be will probably help surgeons give their patients a better outcome.

You understand you're replying in a thread about female genital mutilation?

Yes, I was being a bit terse with my language, which is why I clarified a bit in my last comment. Here's how I might have written it better:

> FGM reconstruction actually seems to have negative outcomes post-surgery. I'm surprised by this.

Surgery is essentially mutilation, just in the physical sense (you are cutting through healthy tissue), not a moral sense (the whole point is to make the body more healthy). The information gathered from mapping nerve endings in a clitoris will hopefully help surgeons perform reconstruction surgery with less damage to the body.

The whole premise of "correct" typing on a traditional keyboard is absurd. It's trying to force good ergonomics into a system that is simply incompatible with it. You're better off either making yourself compatible with the system, or vice versa.

I type "incorrectly" on traditional qwerty keyboards, too. I also type "correctly" on my split ergonomic keyboard, using the workman layout. As far as I can tell, I'm not any faster with either; but I definitely enjoy using the ergonomic keyboard more.

Sycophancy is not just a problem when you are asking for advice. Try to soundboard any new idea whatsoever, and it will just roll with everything you say, no matter how fallacious or absurd. If you ever manage to get an LLM to generate criticisms, they will be shallow and uninteresting.

And of course that is what it does, because there is no thinking involved! There is no logic. No consequence. No arithmetic. There is only continuation. An LLM can't continue a new idea, it can only continue a conversation about it.

An LLM does not have an opinion. Anything that looks like an opinion is just an emergent selection bias from its training corpus. LLMs are trained on what humans write, and human writing is kind and patient much more often than critical.

So what if we trained an LLM to be biased toward generating criticism? That would only replace the sycophant with a brick wall. What we really need is to find a way to bring logic and meaning into the system.