[1] describes the initial elements of HTML. Almost all were derived from an inofficial SGML folklore tagset with the notable exception of <a> and http URIs.
The term "AI" has changed in recent years but if you mean classic game logic such as complex rules and combinatorial opponents then there's plenty of Prolog game code on github eg. for Poker and other card or board games. Prolog is also as natural a choice for adventure puzzles as it gets with repository items and complicated conditions to advance the game. In fact, Amzi! Prolog uses adventure game coding as a topic for its classic (1980s) introductory Prolog learning book Adventure in Prolog ([1]). Based on a cursory look, most code in that book should run just fine on a modern ISO Prolog engine ([2]) in your browser.
The difference is the "algorithmic" timeline (meaning ads) you get with Facebook, Insta, and co compared to the strictly chronological timeline you get on the Fediverse equivalents (Mastodon, Pixelfed). That it's less addictive, or at least not in the doom scrolling type of way, is more a consequence. Aka the enshitification argument.
Masto specifically is also a Twitter not Facebook replacement, with everyone soliloquizing past each other rather than holding a genuine conversation.
For the actual "good" Facebook use cases such as keeping in contact with school/uni veterans or other closed group, there's friendica, but it's nowhere near Fb in terms of volume.
Why are you leading your visitors to your channel on a monopolist site? To bring ad revenue? There's no need for video for your type of content in the first place.
I get it - a 2026 "hackers" campaign for binging yt. And in case you haven't noticed: appealing to the net neutrality debate of the last millenium is meaningless with just a bunch of monopolists left on the net profitting of vast public investments. The kind of thing traditionalist "hackers" in it for social recognition would be wasting their time on.
Because they're betting on the video finding its way onto people's feed, thus raising awareness among non-techy people. Hard to do that with a random website.
please. I don't understand how the fuck we still don't have p2p social networks and private sharing groups. The amount of possibilities to f* up any kind of control are massive - it's just that we end up writing some convoluted distributed mainframe when all people need is p2prss.
When you had gone to a site using a deep link, Safari insists on autocompleting any URL in the domain to that link you used even if you just want to go to the top-level URL/index of that site. You have to type out the entire URL and add a space at the end or something (and that still doesn't work sometimes) to stop iOS from doing that, which defeats the entire purpose of autocompletion. Btw switched off any autocorrection feature a long time ago. Still, I happen to mistype a lot compared to my old non-Apple phones (there was even an "it's not just you" article last year about it).
Apple needs to spend an entire release cycle to unfuck text entry and completion. However, with their qa lately (or lack thereof) they'd only manage to make it worse. The sad thing is they're still better than the alternatives, all things considered.
FIY the Wikipedia article rightfully says SGML CONCUR usage is uncommon, but compared to the stated alternatives for overlapping markup, it's basically the only one that is tolerable to use as actual markup language for use with a text editor. This is what it looks like:
<!doctype d -- element decls for a, b ... -->
<!doctype e -- element decls for a, x ... -->
<(d|e)a>
<(d)b>bla bla <(e)x>bla </(d)b> bla</(e)x>
</(d|e)a>
where the third "bla" span is marked up with overlap.
Basically, in case you've ever wondered, SGML CONCUR is the only reason that the element name in end-element tags needs to be specified. In strictly nested markup (XML) it always must refer to the most recently opened start-element tag hence it's redundant. SGML actually has "</>" but it didn't make it into XML.
Anything you'd like to share? I did some research within the realm of classic robotic-like planning ([1]) and the results were impressive with local LLMs already a year ago, to the point that obtaining textual descriptions for complex enough problems became the bottleneck, suggesting that prompting is of limited use when you could describe the problem in Prolog concisely and directly already, given Prolog's NLP roots and one-to-one mapping of simple English sentences. Hence that report isn't updated to GLM 4.7, Claude whatever, or other "frontier" models yet.
Opus 4.5 helped me implement a basic coding agent in a DSL built on top of Prolog: https://deepclause.substack.com/p/implementing-a-vibed-llm-c.... It worked surprisingly well. With a bit of context it was able to (almost) one-shot about 500 lines of code. With older models, I felt that they "never really got it".
[1]: https://info.cern.ch/hypertext/WWW/MarkUp/Tags.html
reply