Stephen Downes

In my work developing CList I want to enable content and messaging federation - so that people can follow and talk to each other - without being dependent on a specific platform. So I want people to be able to create and use DIDs to help them find and follow each other. But how?

This is what I would call a 'folk theory of understanding': "To understand something well enough to act on it requires three things to converge: reliable observations about what is happening, theory that can organise those observations into coherent patterns, and mechanisms that explain how and why." In other words: observations, ontologies, and causal principles. We spend a lot of time on this, and Jonathan Boymal takes time to describe "sense-making as a system of (six) interlocking components": "the information ecosystem, our natural and built surroundings, the institutions that produce and distribute knowledge, the cultures we grow up inside, the sensing and feeling bodies through which all experience is filtered, and the mental models we carry as frameworks." But (in my view) all of this is an artifice, constructed rather than discovered, and in a complex world no longer sufficient.

As Claire Grouze wrotes, "The semantic layer is presented as the holy grail to make your analytics agent reliable." This is the 'knowledge base' we suppose that intelligence and reasoning works from, filled with ontologies, causal principles, and the like. The stuff we think it takes to 'understand' a domain. But the semantic layer is brittle and context-dependent. When used on its own in this test, it refused to respond to most questions. Even optimized - at the cost of much more work and slower processing speed - it still answered few queries. And, as Grouze writes, it doesn't eliminate hallucinations - it just moves them. This sort of finding is why I think teaching recognition skills is more important that teaching facts and principles. Via Isin Pesch.

This short paper makes the point (and creates a framework of sorts) that offline AI can serve pedagogical purposes. "Offline systems make tradeoffs legible - coverage may be narrower and responses slower, but the system's boundedness encourages learners to treat AI output as a proposal to be tested rather than a conclusion to be accepted."

"I'm biased towards New_ Public's point of view: pro-social spaces, pro-democracy technology, and community as an ingredient for trust are all my jam," writes Ben Werdmuller, "But everything laid out in this presentation is already happening." Even more, a new pro-social technology that doesn't even have RSS is something I'm already questioning. If your collaboration is limited to email, LinkedIn and Bluesky, you haven't thought the subject through.

As Scott Leslie says, this is a thoughtful post on the Instructure hack. Like Phil Hill, D'Arcy Norman leads with Instructure's public communication regarding the incident, which was (and continues to be) not good, with most people interpreting their claim to have "reached an agreement with the unauthorized actor (ie, ShinyHunters) involved in this incident" to mean "they paid the ransom." Though, what else were they going to do? Along with many others, Norman points to the technology monoculture as a point of failure. "It looks like Instructure manages 3 different Canvas environments: 'production', 'beta', and 'test'. And all 8,800+ institutions appear to share those three environments." But the major lesson to be drawn, he writes, comes from UBC. "The UBC response is truly remarkable... (they) quickly put together a cohesive set of resources to support instructors in rapidly adopting different online platforms to meet the pedagogical needs in their courses." This is a lesson in maintaining expertise and developing resilience, even while working with external partners.