Stephen Downes

This is a very good description of how the ATmosphere protocol actually works, with the bulk of the focus on the personal data store (PDS) and how records are actually stored and named. Some of the most interesting bits are elided, like the central role of the Relay (and why it's a big problem if one relay 'throttles' another), as well as some of the intricacies of DID and how the hash plays a role in certifying data (hint: it looks a lot like blockchain). But that's just me complaining. I really like the way this article refocuses the reader from apps to files to records (and, implicitly, to data). 

This is another one of those items I've saving for myself for a time when I can settle in and enjoy it a bit. Hyperblam's slogan is, 'Make music not Javascript', which isn't a dilemma most of us face, but I'm sure it comes up when people want to use their computers to create music. "HYPERBLAM lets you make music with HTML. It's a declarative implementation of the Web Audio API and is completely dependency free. Create pedal boards, drum machines, sampled instruments. And don't write a single line of JavaScript in the process. Unless you really want to." Via Bionic Teaching.

I'm not going to dispute Ben Williamson's argument here. It is entirely plausible that much of the 'AI in Education' (AIED) research is, as he says, "junk science". What I will dispute the the suggestion - not made, but assumed, I guess - that this is new. While it is true that we are seeing "meta-analyses at an industrial rate and scale," this has been a trend for some time, certainly for as long as I have been a researcher, and the same case has been made - does the name Charles Ungerleider ring a bell? Did initiatives like the Campbell Collaboration solve the problem? No - here we are 23 years later, with the same old problems assessing the impact of a brand new technology, suffering from the same old delusion that we know what we want from education, as there is an all-seeing 'we' in that statement. Does a hammer produce better learning outcomes? That would depend very much on the discipline.

Reading Michael Feldstein's post I would have thought Satya Nadella's post was a tome, but in fact it's only a few well-chosen paragraphs long. We need to frame our understanding properly to even understand it. Feldstein compares it in a way to Ronald Coase's 1937 paper, The Nature of the Firm, where it is written, "that people join together in firms when the cost of coordinating as individual agents is too high." Now when Coase - or Nadella - are talking about people, they are not talking about you or I, they are talking about owners who run companies and enter into compacts. We are nothing more than "human capital", owned by them, alongside (writes Nadella) 'token capital', which "is the firm's AI capability it builds and owns". What Nadella sees is that there is no purpose to increasing token capital for its own sake; at most, it could eat entire segments of the economy, but to what end? But it's not clear he can see what more there is. "Our priority has to be building a frontier ecosystem, not just a frontier model, so value flows broadly across every company, every industry, and every country... When that happens, companies will create value for themselves and for the economy around them. Employees will see their expertise amplified and their judgment become part of systems that make it replicable and scalable."

Saving this one for later when I have a better room and a (much) larger screen. I do like the novel presentation. "Crash-land, gather, refine and fabricate your way off a dead moon. A card-game engine where verbs eat resources and transform them over time — the only way off this rock is the one you build." Completely unplayable where I'm sitting right now but I can imagine enjoying this. Via rmoff.

This report from the U.K.'s Information Commissioner's Office on their audit of EdTech companies operating in schools shows that they leave a lot to be desired. For example, they write, "We found that edtech providers need to assess whether reusing children's information for their own purposes complies with the law, before they do so." Additionally, "We found that edtech providers need greater control and oversight of their sub-processors. They must ensure due diligence checks and school authorisation are in place before granting access to children's information." The case studies included in the full length section though not naming ed tech companies nonetheless shows how urgent some of the concerns are.