hatzka

@notyourfanboy@kolektiva.social @Elindur@retro.pizza The first episode that came up on YouTube when I searched was How People Make Crayons. It's about half an hour long, and has amassed 12 thousand views in the less than a day it's been online. On YouTube, they can do that for a maximum cost of zero dollars and zero cents. If it was hosted by Cloudflare Stream, this single episode would already have cost them 360 dollars or so, assuming they had managed to get such a large audience to visit their website in the first place. Building bespoke infrastructure just to host Mister Rogers' Neighbourhood and nothing else would cost enormously more. And the bigger their audience grows, the more expensive self-hosting becomes, while YouTube's cost to the uploader holds steady at absolutely nothing. Don't get me wrong. I really, really wish they could. But the fact is, nobody has found a way to make online video work at scale yet without either charging the audience directly or throwing ads at them.

@lil5@social.last.nl @hgrsd@hachyderm.io @slava@mathstodon.xyz Zap is just one project written in Zig, I don't see why Zig should be blamed for what Zap does

@chaucerburnt Not to mention, the purpose of asking a person something like that is to get them to understand what they did wrong, so they can learn from it. LLMs, meanwhile, do not learn from their users' logs, at least not in any way that persists beyond the chat session. The fact that they don't is crucial to their appeal to bosses.

@mdhughes@appdot.net @arclight@oldbytes.space That seems overly pessimistic. I can't imagine you grow your own food, built your own house, synthesize your own medicine. For all these things you rely on others, even though the stakes are higher than for most software. I think the difference is the lack of oversight. There are enforced regulations governing how each of these things are done, so that (at least ideally) you can rely on them being of a certain level of quality. Software is pretty unique in not having that yet, at least at a societal level.

@LinuxAndYarn @futurebird @ingrid GrapheneOS's goal is to be as secure and private as possible. To the extent that it's also degoogled, it's because accepting Google's control over Android creates problems for security and privacy. If you think avoiding specifically Google at the expense of being much more vulnerable to everybody else is worthwhile, you're free to use another Android variant, but I honestly find it petty and demanding to be mad at a noncommercial project for not putting your preferences above theirs. But, for the record, they have said they're working with another OEM (exactly who hasn't been announced) to make devices that meet their hardware security standards. This isn't a change in goal, of course—it was prompted by Google moving towards restricting third-party OSes on Pixels. You can't protect people's security and privacy if they can't actually run your software.

@whitequark@social.treehouse.systems well of course you play Doom on it, that's why it's called a Doom Processing Unit

@AutumnDork@meow.social you can travel the world looking for cheap ram

@Marijke@mastodon.green @maxleibman@beige.party Because they're businesses. Water cooling means moving heat from the computers to the water, which leaves you with a bunch of hot water to deal with. Unfortunately, according to a quick google, the cheapest thing to do at that point is just let it evaporate.

@frela @0xabad1dea Japanese does this too, but even worse: mēru can mean email or texting. the story I've heard is that in the past, they preferred to send short emails from their phones instead of SMS, and then when they switched they kept using the same word for it. nowadays they prefer internet-based messengers, just like in the West, but if you do want to talk about SMS, mēru is still the colloquial word.

@bascule@mas.to I'm hopeful, but they tried once already and failed. We shouldn't celebrate as if the deal is through, not until they actually control the site itself.

@mos_8502@studio8502.ca I keep thinking about this, even though I don't have nearly the EE chops to realistically do it. By "a design which exposes the full system bus", do you mean that it must be possible to connect off-chip (and potentially off-board) peripherals directly to the same wires used by RAM? It's been decades since anything remotely mainstream has worked that way. I get that something like PCIe is complex, but purely in terms of expandability, is it really insufficient?