Stephen Downes

I've discussed AI agents and related protocols (MCP, A2A) in previous posts, but unlike David Ross, I don't actually see them being applied to classroom learning. And I consider the recommendations to be misguided (the number one priority in the table illustrated is "invest in adaptive learning pilots", which has been the same recommendation coming from this crowd for decades, and has never been useful). If AI does anything, it will free students from teaching, adaptive or otherwise, and allow them to learn by creating and doing.

How often have you heard this about AI? "It's just a next-token predictor." Strictly speaking, that's true. But that understantes what it is. As Carlo Iacono explains, "The critique of AI as 'just prediction' makes the mistake of focusing only on the mechanism rather than what emerges from that mechanism at scale. It's like saying humans are 'just neurons firing' - technically true but missing everything meaningful about human experience and capability."

A "rough sketch of the near future" that seems to me a lot more plausible than a lot of what I've read in tech media. "Even if it goes as well as possible, make no mistake: AI agents will involve human beings taking their hands off the wheel of the economy to at least some extent. Most of the thinking and doing... will soon be done by machines, not people." Ball adds, "Epoch's Ege Erdil and Matthew Barnett published a piece with a somewhat similar thesis." Also worth reading.

This article introduces the Learners' Online Discussions Analysis Model (LODAM) to analyse learners' discussions in MOOCs. "When discussing in MOOC forums," we are told, "learners explain and clarify concepts, answer questions, and negotiate meaning." The authors are trying to resolve what appears to be a contradiction: "Participation in MOOC forums is often correlated with course completion and higher learning gains; yet, studies also report limited knowledge construction." This changes if you recognize that "studies analysing MOOC forum textual data have focused on discussion topics rather than the concepts themselves." That's what the model corrects. "LODAM does not classify learners' discussions into fixed, discrete categories. Instead, our novel model aims to highlight the spectrum of potential dialogues that may emerge from learners' socio-cognitive engagement in discussions." 

This article considers the arguments for and against AI by first discarding the most popular artuments for each position - "depriving each side of its best weapon" - on the grounds that it is unconvincing to the other side. So, no 'analogy with humans' for the pro side, no 'it violates copyright' for the con side. The result is an argument between 'it may result in great scientific advances' for the pro side, and 'it may put the entire creative industry out of work' on the con side. For me, the scenario not considered is far more interesting: what if it were a good thing to replace our entire creative industry with AI? People complain about how much energy AI consumes; consider how much more energy all the people in the creative industries consume! Consider how much we have to pay for all manner of creative content! Why must artistic genius be rare, when we could produce it for pennies every day? What would all these people do instead? I don't know - something useful? I jest, of course, but only a little.

Miguel Guhlin thinks about making a backup of Wikipedia and other similar sites and reports on Kiwix: "you can run it from any storage drive, making it great for placement on a USB drive after you've downloaded (faster) the libraries you want to safeguard for democracy." It's not just for Wikipedia: for example, you can also download RationalWiki (RW), "a wiki-based community working together to explore and provide information about a range of topics centered around science, skepticism, and critical thinking."