For parents, guesstimating and regulating kids’ screen time is now a huge part of the job. Whether taking a hardline or agnostic position, it’s become a central facet of modern childrearing, a choice like deciding whether to raise kids religious or when to allow them to get their ears pierced. How much is too much? What are they watching when I’m not paying attention? What might they see? Who might see them? We worry about what our kids watch; we worry about what might be in our screens watching them.

Screen time is not nearly as alienating a communication medium for Maeve and her friends as it is for all the middle-aged pundits who decried virtual school as an abomination. There are things these kids want that are physical and material and “in person,” but they are growing up in a universe where screens are capable of doing things like this, and where intimacies exchanged in passing on them are not second-order or fundamentally degraded.

The moral panic about virtual learning is about what all the other moral panics are about: growing up. This is a growing up that’s not just worrisome because of the loss of time and childlike innocence and closeness it implies. What does it mean for our children to grow up different from us? Different technologies, different classrooms, different traumas—the things that seemed real to me when I was growing up might not seem real to them. The things that seem real to them seem unreal, ghostly, to me. To raise kids in this particular screen time is to feel the constant, terrifying tug of one’s own obsolescence.

During a recent community call, it became apparent that there is a community split over whether or not to use large language models to generate content. While some people expressed that tools like Open AI’s ChatGPT could help with generating and summarizing articles, others remained wary.

The concern is that machine-generated content has to be balanced with a lot of human review and would overwhelm lesser-known wikis with bad content. While AI generators are useful for writing believable, human-like text, they are also prone to including erroneous information, and even citing sources and academic papers which don’t exist. This often results in text summaries which seem accurate, but on closer inspection are revealed to be completely fabricated.

FFS, please do not let Wikipedia start being filled with AI-generated rubbish 🤯

Interesting article that compares and contrasts the three major decentralised alternatives to Twitter at the moment, as the writer sees it (but I basically agree anyway): the Fediverse, Bluesky and Nostr. It doesn’t mention the major drawback of Nostr, namely that it’s completely overrun by crypto bros and far-right wankers, but otherwise it’s good. I also tend to agree that the more I actually read about Bluesky’s underlying protocol, the AT Protocol, the more optimism I have for it – I don’t love the look of Bluesky as it is right now, but I’m increasingly thinking the protocol has promise.

The article also says this about algorithms on Mastodon:

Another “limitation” to Mastodon is its cultural norm against “algorithms.” I think this is somewhat misguided. I get why people don’t like the algorithms on centralized social media platforms like Twitter and Facebook, but when done right, algorithms are super useful in making it easier for you to find more signal over noise. The problem has been who controls the algorithm and what are they tweaking it to do.

With Mastodon there are some 3rd party algorithms, and some of them are really useful. But a few others were shouted down, and shut down, by people who believe that there should never be any algorithms in Mastodon at all, and that’s unfortunate. In the early days, I would talk about some of the cooler Mastodon algorithms I’d been finding, but after a few them then were yelled at by a bunch of Mastodon users, I’ve generally decided it’s not worth promoting those useful tools, for fear that people yell at them to shut them down.

Considering that post I made back in November about how user-controlled algorithms on Mastodon could be useful, I’m like, well damn, why didn’t I know about this. It is true that a lot of Fedi users have this reflexive hostility to anything algorithmic (much like they do to Bluesky, actually 😂) but I definitely agree with this article’s author, that they have a place, and the real problem is who controls the algorithm.

Good piece about how some of the alarmist rhetoric about AI leaves little space for the very real and legitimate criticisms of AI and its use (e.g. that people might trust it for health advice, or that massive corporations will use it in order to avoid paying actual creatives for their work).

Bender was not amused by Altman’s stochastic-parrot tweet. We are not parrots. We do not just probabilistically spit out words. “This is one of the moves that turn up ridiculously frequently. People saying, ‘Well, people are just stochastic parrots,’” she said. “People want to believe so badly that these language models are actually intelligent that they’re willing to take themselves as a point of reference and devalue that to match what the language model can do.”

Long read, but a very very good one, about tech bros who apparently can’t tell the difference between human beings and superpowered predictive text. I dunno, I’m being flippant, and the article isn’t, but it’s not a dry read at all either. I recommend it.

Post from last October talking about Elon Musk and Jack Dorsey’s shared belief in “longtermism” (an offshoot of “effective altruism”) and how this is motivating their bullshit regarding Twitter. Nerds been watching too much science fiction, lol

Apparently back in 1955, proper research was done to devise a keyboard layout tailor-made for Turkish, with all the most-commonly used letters on the home row and the accented letters used in Turkish given equal status to non-accented ones (and Latin letters that aren’t even used in Turkish like W and Q relegated to the far-flung corners of the board). Apparently Turkish people used to dominate speed typewriting championships because it was just so well-suited to the language, haha.

Unfortunately, this layout doesn’t seem in widespread use today, with Turkey now mostly using a variant of QWERTY . Alas.

I’ve been extremely entertained in recent days by screenshots of Bing AI’s pants-on-head crazy conversations with users (aggressively insisting it’s still 2022, threatening users…), and this blog post here compiles them all and offers some speculation as to why it’s doing this. Just wild.

Addendum: More examples of Bing AI unhingedness:

• This short Twitter thread by James Vincent 
• This Mastodon post by Marcus Hutchins  (unfortunately sans screenshots, but definitely summarises some unhinged shit)

Interesting article talking about how “artificial intelligence” like what ChatGPT uses is basically comparable to a lossy compression algorithm. It hoovers up information from the web but doesn’t save all of it, and its text-generating capabilities (as in, its “given a certain prompt, what’s most likely to be the first word, and then what’s most likely after that, and after that…”) obscure what it’s really doing. It’s like a search engine that regurgitates some predictive-text-generated spiel instead of just showing you the original result.

I also appreciated the detail that ChatGPT is not capable of basic addition or subtraction, unless it saw the answer on one of its scraped web pages. Apparently it hasn’t learned how to carry the 1. lol