Gregory's Server|
19 posts total
Self-driving cars are sorta not really all that tricky. Build out an infrastructure specifically for them, and we already have all of the technology to make them run safely and efficiently. The reason we don't already have widespread adoption of self-driving cars is that we don't have that infrastructure, and no one wants to build it. The real problem car companies are struggling with is designing self-driving cars that can navigate the existing infrastructure that's made for and used by people. It's possible to read people's enthusiasm for self-driving vehicles as a sublimated longing for the convenience and efficiency of really good public transportation—a longing that they can't quite reconcile with all of the associations they've built up around public transportation, or their willingness to give up the glamor of private car ownership. @lrhodes I think you have just created a new genre, the meta meme, and taken home the gold medal for it, all in one.  This is the most logical use of two memes I may have ever seen. May /ever/ see. :blobcatthinkOwO: The current wave of AI hype demonstrates of the problem of capacity creep. In some measurable regards, hardware keeps getting better, right? More powerful, more efficient, more affordable. But rather than letting those gains ameliorate the problems created by tech, tech companies see them as increased capacity, and look to fill that capacity. So while hardware improvements should be reducing the emissions caused by computing, their growing as companies look for ways to turn capacity into profit. It follows that assurances to the effect that the AI industry will eventually find ways to shrink the carbon footprint of their technology are entirely untrustworthy. Any improvements that might lower the emissions caused by AI processing are bound to be treated by the industry as excess capacity, which they will promptly find ways to fill, excusing the increasing size of the footprint the same way they excuse it now: by promising that they'll find ways to shrink it in the future. The big problem that the current trend in "AI" is meant to solve is that of too much opportunity. The web opened up all sorts of creative spaces and jobs to anyone who could navigate the technology. But it's hard to build an effective elite when the same opportunities are available to everyone. You have to counteract that leveling effect by redirecting all of that creative power to a relatively small pool of comparatively privileged people capable of functioning as a virtual aristocracy. Aristocracy is not a bad model for thinking about generative AI. If you think of websites as a form of virtual real estate, then the function of, say, LLM-powered search is to shift the value of other people's labor to the internet's biggest landholders. You pour your hard-won knowledge and creativity into your little digital plot of land, and an AI bot comes along to collect your harvest so that Google sell ads when they take it to market as their own. Somehow you work for them now. "Reverse chron IS an algorithm." Yes, but even more than that, we're all of us making our own custom algorithms by the way we arrange our accounts. The software and services we use provide us with particular tools for doing so. Reverse chron and follow are the basic underlying tools, but there's also block account, block domain, mute account, mute word, follow hashtag, mute replies from this account, and so on. Even choosing a host instance can be a means of shaping your algorithm. If Microsoft isn't your email provider, don't use Outlook. (Don't use it either way.) https://mailbox.org/en/post/warning-new-outlook-sends-passwords-mails-and-other-data-to-microsoft "Great Andamanese, it turns out, is exceptional among the world's languages in its anthropocentrism. […] Because no other known language has a grammar based on the human body or shares cognates […] with Great Andamanese, the language constitutes its own family." https://www.scientificamerican.com/article/this-ancient-language-has-the-only-grammar-based-entirely-on-the-human-body/ #language #linguistics #India Been daydreaming about a modular ActivityPub implementation. Each instance could decide which service types to offer, and each account could decide which to use. Accounts are held at the TLD level, services on subdomains. So maybe foo.bar offers micro.foo.bar, photo.foo.bar and checkin.foo.bar. You have a foo.bar account, but only connect to micro.foo.bar and photo.foo.bar. And I only follow you on photo.foo.bar. Don't know how feasible that is with AP. That maybe seems a little complicated from a UX perspective, but I think it could be made relatively seamless. There could be a standard post form, and the service/subdomain you're using changes depending on what options you invoke. Clicking "add image," for example, switches you over to photo.foo.bar. But the option to add image only shows up on your form if you connected to the photo service. Which would be a relatively straightfoward toggle in your account settings — if foo.bar offers photo. These are my attempt at a few fast-and-sloppy diagrams illustrating how #Mastodon and #bluesky handle two kinds of message, both in the simplest scenario and in the more complicated conditions they'll both operate under when Bluesky goes into full production. Part of the point here is that the beta obscures some of the major differences between them. Seeing how Bluesky is intended to operate according to the spec makes it easier to assess its implication. Looking at the differences between multi-server fediverse relations and multi-server Bluesky relations ought to make it easier to see why Bluesky couldn't just lift Mastdon's moderation solutions wholesale. And it opens up a bundle of questions about how federation and moderation work on the Bluesky network that probably won't be clear outside of the company until the beta ends and third parties start federating their own servers into the network. If you're attracted to Bluesky by the promise of account portability, then I think it's important to understand what that entails. Because one of the claims they're making is that you'll be able to rescue your account in the event that your home server suddenly closes. But how would that work? How do you migrate an account from a server that's no longer accessible? And part of the answer has to be: Well, you don't. Because it lives somewhere else. 🧵 1/? The first thing to understand is that BlueSky servers aren't entirely analogous to fediverse servers. The protocol docs call them Personal Data Servers (PDS), and they handle only a portion of what fediverse servers would handle. Mostly they deliver personal interactions between accounts—"small-world" networking, in the jargon. The rest of your account—your network identity, really—is stored elsewhere. That has implications you should really pay attention to if you're thinking of joining. Some interesting discussion here about using Mastodon to build server-specific funding programs for artists and writers: https://mstdn.social/@Jonathanglick/110181080905371639 The potentials being discussed are due in no small part to the fact that we're not all on one big server, and can more easily build communities of alignment by leaning on the soft borders created by decentralizing social media across multiple servers, each with their own identity. Here's the latest Mastodon development roadmap: https://blog.joinmastodon.org/2021/12/roadmap/ Some interesting stuff in here, like reply controls, dynamic roles, and groups potentially replacing (or only supplementing?) the local timeline. Seems like some items have already been implemented in v4, though—like batch filter rules and post editing.
Show previous comments
A couple of people have pointed out that this page is a year old. Sorry for the mix up. It popped up in my RSS feed this morning, so I glossed right over the publication date at the bottom of the page. Make fun of John Mastodon all you want, but he gave an unflinching, nuanced performance in Goncharov (1973). In the past, I've had accounts that crossposted to Twitter, but I've ultimately either disconnected the crossposting or let the accounts go defunct. And gradually I've come to the conclusion that crossposting undermines the goal of building a better social media ecosystem. It provides people with an excuse for staying on corporate social media and encourages them to see it as your responsibility to deliver your posts to them wherever they already have an account. I'm not saying never crosspost. I'm willing to grant that there are probably some circumstances in which a limited degree of automated crossposting makes sense. But if your goal is to encourage people to migrate to a healthier social media environment, then it makes sense to be very thoughtful and restrained about how you mirror posts from one platform to another. And I don't think we're thoughtful enough about it. I certainly wasn't.  @lrhodes I agree. My current approach is to have specific and very limited accounts on corporate platforms for communicating with the people who use them (very specific to topic interests). By having a website and pointing to it as my "home" anyone interested in deeper connections can dig further and see how much more beyond the surface there is. @lrhodes so much this, honestly. People need to commit or it never works. Because maybe at the root of it... we don't all need social media. The threshold for discoverability on Mastodon is pretty high, so be generous about boosting other people. Sometimes you're the signal, and sometimes you're the path the signal takes through the noise. Maybe it's just my group, but a lesson I'm steadily learning is: If you've given your NPC a name, and you want your players to know it, introduce it early, because they absolutely will not ask what it is. The enlightened GM way to see it is: They're looking to you for cues as to which NPCs are significant enough to bother learning about, and providing a name upfront is a clear gesture in that direction. |
It's not that photographs CAN be fake. It's that there's always something outside the frame. That the image is captured from a particular place in relation to its subject. That all it reproduces the arrangement of light during a particular span of time, often measured in milliseconds, and that moment may not be representative of the milliseconds that came before or after. That the process and choice of equipment influence the result. That someone has reasons for what they've chosen to show you.