Stephen Downes

One of the more controversial decisions made by the developers at Bluesky in the early days was to make all data public. That is, your personal data store (PDS) could be read by anybody. In what they describe as "the biggest update to atproto since it launched", the managers of the specification behind Bluesky have created 'Atproto Spaces': "You can think of an atproto space as a miniature atproto network that can be gated so that only certain people and applications are able to access the data published in it." So, in other words, private gated personal data. Still not secure, though: "It's important to remember that spaces give you access control not confidentiality. The data in a space is readable by any user or application with access to that space, it's not encrypted." Here's an example: secretsky.

This article describes how a teacher revisited curricular materials that have been untouched for years, organized them, and compared them against the current standard for gaps, then rebuilding them. "Look at what actually got built here without touching a single student record: a dusty, sixteen-year-old archive rebuilt, a 297-page College Board framework read start to finish, a curriculum audit nobody had an afternoon to run, and a seven-week plan tied to a real school calendar." You probably don't need to read the article in detail to appreciate the point. Via Wes Fryer.

According to this and several otger articles, Divine has opened to the public and no longer requires invitation codes. Divine is a reboot of Vine, a short-video streaming app that was popular before the rise of TikTok and Loops and the like. The appeal of Divine (in theory at least) will be originality; videos will need to be created directly within the app, and AI is detected and blocked. "None of the stuff you see on Divine was ever on other platforms before, it was either on Vine originally or it's now on Divine. You're not going to see the same thing that you see on TikTok and Instagram." What's interesting is that Divine is based on Nostr, a federation protocol similar to ActivityPub and ATProtocol. There's also a story on BBC News which may or may not display for you - BBC is setting up geo-based walls, or at least, is trying to. See also Videoweek, Technology.org. Here's my one-video Divine.

This is a proposed set of standards (218 page PDF) from the Computer Science Teachers Association (CSTA) for computer science education for primary and secondary students (aka. K-12). It depicts computer science as "a scientific and creative discipline focused on understanding and designing algorithms, data, and computing systems," with an emphasis on such things as algorithmic thinking, collaboration, ethics and responsibility with the set defined as concepts and practices. There's far too much in here to summarize, but it appears well thought-out and comprehensive, including core elements of traditionall computr science as well as consideration of newer AI-based methods. Via the Association for Computing Machinery (ACM).

This article is kind of weird because it's written by Jim Highsmith and appears on Martin Fowler's website and says things like, "Martin Fowler writes this way. His best pieces don't build to the conclusion, they open with it," etc. Looking beyond that, the article purports to name something that hasn't been named, 'practitioner voice', that is, writing done by people who actually do the thing (which arguably forms the majority of the content in my newsletter). It's not a new idea - here's a reference to it from 2019. Here's one from 2024. And it doesn't follow any single form, much less the model proposed here ("Academic writing builds to the claim. Thought leadership makes the claim and explains it. Practitioner Voice starts with the claim..."). There isn't so much a 'claim' in practitioner writing as an 'experience'. Academics then 'discover' that experience, and influencers copy it and try to make it popular.

What I like about this article is that it gives us an extended extract from the beginning of Winter Takes One, "an 89,000-word upmarket thriller" that was written by an AI and "led to a four-way bidding war and a $3.5 million, two-book deal." It's an example of 'superslop', AI writing that actually fools professional editors. I also agree with the conclusions, two of them: "Commercial writing, at any level of quality, might be dead. It's already nearly impossible for a writer to earn a living on the quality of text"; and "Superslop will continue to evolve. It will defeat AI autograders. It will defeat human literary agents." Photo by me.