Hacker Newsnew | past | comments | ask | show | jobs | submit | xethos's commentslogin

The tools we use are not neutral. A sword can be made to work like an axe, but we use axes for chopping wood because a sword makes a shitty axe. A sword is designed to kill people. The handle, the mass, the weight distribution, and every other aspect I am not qualified to get in to, means swords are designed to kill. They are a tool, and their use is not neutral.

This is a clear example, but I don't believe any tools are neutral. Your immediate fallback was to a hammer, not a mouse, with the obvious corrollary being to bludgeon, but the same line applies. Tools are not neutral, and that's why when you looked for something that causes harm, you grabbed something that's objectively been serving a dual-purpose for hundreds of years. Nobody's using a computer mouse to bludgeon someone to death; it makes a shitty bludgeon, and the design of the tool reflects that.

That's also why these comparisons always fall back to knives, or hammers, or the AK-47: they are dangerous tools that are designed to make killing easier. Nobody is making these comparisons to more benign tools, like desk lamps, coffee cups, or car stereos, and it's because tools are not neutral, and none of my examples are designed to make direct, bodily harm, easier.



My larger point is that nobody - nobody - defaults to telling us the coffee mug is unregulated, as AI allegedly ought to be. They always compare it to something much more commonly used as a weapon; something that, when asked to name a household object likely to be used as a weapon, the average person would guess.

Your point is that people make a stronger argument even when a weaker one would be sufficient?

Instead of comparing AI to any other tool, especially one closer to "useful with a computer", the common comparison is always a weapon of some kind.

If the design of tools are neutral, one tool should do as well as another in this common comparison. But the useful application of tools is inherent in their design.

If tools were neutral, as so many on this site claim, why is AI only ever compared to knives and hammers?

Parent has lots of links to other common objects causing harm, why are they never used as the example when tools are allegedly neutral? That would be a stronger argument opposing AI regulation - ethernet has less regulations that knives, but can still be used as a murder weapon


Hammers are kind of just the prototypical tool, but I've definitely also seen comparisons to keyboards, paintbrushes, and traditional digital tools.

> why are they never used as the example when tools are allegedly neutral? That would be a stronger argument opposing AI regulation

The argument is strongest if pointing to tools that have larger potential impact yet are still widely considered neutral and not/loosely regulated.

"We should consider AI a neutral tool and not heavily regulate it because we do the same for drink coasters" is not convincing, because there's not all that much you can do with a coaster.


The fact that you had to find an article from three decades ago for an instance of killing with a keyboard is telling. All the others aren’t exactly that recent and are mostly isolated cases. Meanwhile, on gun related deaths, there are entire Wikipedia pages for it:

https://en.wikipedia.org/wiki/List_of_countries_by_firearm-r...

https://en.wikipedia.org/wiki/Lists_of_mass_shootings_in_the...

There are more mass shootings in the US per year than there are days in a year. It’s so bad they need pages for each individual year.

https://en.wikipedia.org/wiki/List_of_mass_shootings_in_the_...

Meanwhile, pages of deaths perpetrated with household items are curiosities. You parent comment stands: tools are designed for specific purposes and are used for those purposes.


>The fact that you had to find an article from three decades ago for an instance of killing with a keyboard is telling

Yeah, is telling that modern keyboards weight a lot less nowadays, and nobody would use one as a weapon to hit someone else. ;)

The original IBM Model M was 2.3 Kg.


Thank you for the laugh, wasn’t expecting that. Though I have a modern external Apple keyboard, which is not that weighty but it is metal and fairly thin with sharp corners. It could do some damage.

Plex's outrageously heavy Android client (on my admittedly older phone) was actually what pushed me the rest of the way over to Jellyfin

They support (or did previously support) more platforms, but Jellyfind is what runs at a reasonable speed on my phone, so that's what'll get used


Ouch. I wonder if this is related to when Plex tried to unify development of their mobile app in late 2025 so they’d have a single shared codebase for both Android and iOS?

I’ve heard from a lot of people who accidentally upgraded to the latest version of Plex on their iPads that it’s a shadow of its former self: tons of features are missing, SRT are unreliable, picture-in-picture flakes out, etc etc.


It's possible, because I used the Plex app on that same hardware for years. It certainly wasn't Google's software gettng heavier (it was already RAM and battery-hungry enough I'd dropped it), but trying to use Plex became an exercise in frustration around that time

Jellyfin's outrageously heavy Android client was what made me switch to Plex

> I can't think of a lot of crimes whose metadata warrants being killed for personally

You're (literally) missing links then. If A is a high-value target that we look at closely (because they're a high-value target), what if B frequently contacts A? If C, D, and E always recieve messages from B immediately following A messaging B?

What about times? Is B messaging F at a consistant time, and never outside of that? Is A only messaging G, at a set time, with G's phone immediately being put into (ineffective) airplane mode immediately before and after?

Facebook built their business on the social graph, but the CIA's been at this for decades


Thanks for explaining. I guess we are talking about espionage or something like that. I've been so focused on the rise of domestic surveillance lately that I forgot about the noncitizen aspects. Which is ridiculous but at the same time, it does seem like a trillion dollar focus lately.

My examples are all based on the CIA and NSA playbook though, as it was the NSA director that said the quiet part out loud, explicitly, in front of Congress. The NSA is effectively America's red team, an offensive arm, meaning they (should be) focused on threats (percieved or otherwise) outside the country

The FBI has been much quieter about this though - there has yet to be a Snowden-for-the-FBI, though they would be one of the agencies I would fully expect to be doing similar work domestically.

As this becomes more well-known, I would expect state and county police to start looking into data and metadata as well. In some cases, they already are [0] - even if some aspects of that case are less relevant today (Google Maps no longer uploads location history, though cell tower trilateration is getting more accurate, not less).

It's far more prevalent than most people realize, though I invite you to consider which you'd rather have when building a second-by-second profile of a person's life: the message contents, or the metadata?

[0] https://www.wired.com/story/find-my-iphone-arson-case/


Metadata would be more powerful in 9 out of 10 cases. Message contents could be invaluable in some cases too. Interesting to think about

One thing I hope we've all discovered by now is that, if Stallman hasn't been proven right at the present moment, on any topic that touches on libre computing, is that it's only a matter of time until he is

That's because having to pay the large fine does not deter crime, and bumping the price does not have a major affect. Increasing the odds of getting caught is much more effective. [0] shows states this outright in the abstract

[0] https://www.economicstrategygroup.org/wp-content/uploads/202...


"The United States switched to Metric to show smaller numbers at the pump", due to a surge in gas prices, caused by a war in the Middle East started for fabricated reasons, would be the most on-brand thing for America I can think of

Between the US and Canada, Canada (with it's population the size of California) has three out of four of the highest-ridership light rail systems.

Blaming sprawl or population count, while being outshone by Canada, means it's neither of the above. Perhaps we can move on to the auto companies pushing out light rail in California in the fifties to bump their own profits, or accept that it's the American people and their ethos that has left the automobile as the claimed only option.


I actually went to call you on this before double-checking just to be sure. No, you were right - 80 characters exactly. Couldn't even fit the apostrophe for "reporter's"

> EVs were sold in the US as muscle cars for rich people

Yeah, the Nissan Leaf was a high-torque monster. Though to describe the BMW i3 as a muscle car is... not the descriptor I would use.

EVs were not sold by every OEM as high-power drag-strip rock stars - that's just what it took to get Americans to pay attention


Fennec, for Android too. The unfortunate part is that it doesn't (by default, on F-Droid) use Firefox Beta - meaning custom extension packs can't be used

This matters for things like Redirector (www.reddit -> old.reddit), Greasemonkey (hckrnews dark theme), and (for my keyboard-equipped Android) Vimium


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: