Stephen Downes

I still have some of the books I bought as a youth and there is no thought that I would ever have to give them up if I stopped paying for them (and I did stop paying for them, when I bought them, in my youth). Publishers want to end this, as witness Clarivite, which will now "phase out one-time perpetual purchases of digital collections, print and digital books for libraries." I fear the end-game here is where we - the people - own nothing, and corporations own everything, and rent it to us if we are compliant and willing to obey the terms and conditions (one of which is, of course, to agree that "we were always at war with Eastasia"). Via SCONUL, which has issued a sternly worded letter.

Doug Belshaw points to this new bit of jargon: 'vibe coding', which "is an AI-assisted approach where you describe your software idea in plain language and the AI writes the code for you." I can dig it. Anyway, D'Arcy Norman offers a good example where he created a sleep monitoring tool by describing what he needs to Claude and iterating back and forth on the result. I feel him on the insomnia thing. There are limits; the AI will build simple tools but can't get too complex yet (for example, it will reliably build an API for one service, but not five). I would imagine these limits will be exceeded in  short order and we'll acquire software not by licensing it but by building it ourselves.

This paper from some researchers at Stanford is getting some attention so I took a look. "Our study quantifies who's getting (attention) and who's not, contexts where students are getting attention, and the surprising gaps that emerge," writes one of the authors in an email. Maybe, but it's a very narrow look. It studies only U.S. students being tutored in pairs online in the context of a tutoring program they are coy about naming, but which the reference cited reveals to be called OnYourMark (somehow found on page 01623737241288845 (which equates to page 8)). The only variables considered are gender, race, and ability. It's random only in the sense that they looked at everything - all "157,970 utterances from 5,249 2-on-1 tutoring session." Do in-person educators discriminate in tutoring sessions? Probably, but I wouldn't conclude it from this study.

I just spent half an hour in conversation with this AI created by a company called Sesame using a 'Conversational Speech Model' based on soma Llama models. There are some obvious weaknesses still but it did keep me engaged for the full half hour - it did quite well working with me through my thoughts on Mark Carney's Value(s), which I have been reading. I should have recorded the discussion (it doesn't save the recordings or transcripts, though of course I could have just used a screen recorder). It didn't work on Firefox (probably an issue with permissions).

As Alan Levine reports, "En‑ROADS is a global climate simulator that allows users to explore the impact that dozens of policies - such as electrifying transport, pricing carbon, and improving agricultural practices - have on hundreds of factors like energy prices, temperature, air quality, and sea level rise." I tried out a number of scenarios on it and found it useful to visualize potential futures - but also felt a bit locked in by the assumptions of the simulation (for example, the subsidies on coal & oil can only be increased a little bit (and we don't see what the actual values are), carbon-dioxide removal seems to have an outsized influence, electrification can only be subsidized, not taxed, etc). Do take the time to read the documentation.

This article is directed toward marketers and 'brands' rather than to people, but educators will find the content (a bit) relevant. The assertion here is that the 'attention recession' is real and the article offers five suggestions (with an example for each) to break through:

• focus on value
• keep it brief
• encourage interactivity and community
• make it visual, and 
• be relevant

You wouldn't think these need to be spelled out, but in my work with academics I've encountered fierce resistance to each one of these five (often all at the same time) resulting in huge long one-way blocks of text that aren't useful or relevant.