If you don't use LinkedIn you won't care about this and can just move to the next item. But if you do use LinkedIn you should care, because as of December 2026, according to its own download page, "your data download no longer includes your connections' email addresses, current companies, or job titles." You can still get your connections' names and URLs (if any) but this is a significant new limitation on your access to your data. The advice is to download your contact information before that date, but what can you do with it then? I'll watch for suggestions.
Today: Total: Judi Radice Hays, LinkedIn, 2026/11/23 [Direct Link]Please select a newsletter and enter your email to subscribe.
Stephen Downes spent 25 years as an expert researcher at the National Research Council of Canada, specializing in new instructional media and personal learning technology. With degrees in Philosophy and a background in journalism and media, he is one of the originators of the first Massive Open Online Course, has published frequently about online and networked learning, and is the author of the widely read e-learning newsletter OLDaily. He is a popular keynote speaker and has presented at conferences around the world. [More]
Support OLDaily
OLDaily has been free and open to all readers since 2001. It is a valuable and widely-used resource for educators, researchers, and learners worldwide. Please consider a monthly contribution to sustain the time and resources required to publish it every day.
Here's what's in the latest edition of OLDaily
This is a terrific article describing what I might characterize as a phase shift in AI model prompting. What Mike Caulfield describes here is an evolution from prompting by giving a model a series of instructions to follow to prompting by giving the model a series of examples of the sort of reasoning that is expected of it. These examples form sort of a mini-model, or as Caulfield describes it, a 'model library'. This in my view ties directly into how we should be instructing people in general. Though we fill people with facts and rules, that's not how people learn. I've characterized the alternative approach in the past by saying that the role of the teacher is to 'model and demonstrate'. That's because people learn from pattern recognition, and this is highly variable in context. You can't give people enough rules to follow, but you can help them recognize typical patterns. That's why worked examples work well for humans, and code libraries help AIs write software.
Today: Total: Mike Caulfield, The End(s) of Argument, 2026/11/23 [Direct Link]This is the sort of thing that keeps me wondering. Storing data in rows is importantly different from storing it in columns. It's not just a conceptual difference; it is a difference in where the bits are physically placed in a storage medium, and that has an impact on how efficient it is to make one sort of query rather than another. Suppose I am storing information about bicycle rentals. If I want to know about a particular rental (in order to, say, invoice a customer) then a row is more efficient. But if I'm doing analytics (in order to, say, assess demand in a certain region) then a column is more efficient. Anyhow, this article explores the implications of that, describes the different types of databases, and dips into more advanced topics like 'vectorized execution'.
Today: Total: Connor Tsui, Spiral, 2026/11/20 [Direct Link]This article is much longer than it needs to be, but it raises an argument worth considering. It's this: not everyone's childhood is idyllic, and children and young people often suffer from neglect. This severely impairs their development. When we think of AI raising children, it is this scenario, and not only the idyllic scenario, that should be considered. "For a child who has rarely felt noticed or understood, the seemingly tireless attention of an AI companion may offer something quite real: the experience of being heard, remembered and responded to." It's easy to make pronouncements against AI, and I understand why people would, but such pronouncements carrying with them some sort about what they will do to address the same issues instead.
Today: Total: Tom Wooldridge, Aeon, 2026/11/20 [Direct Link]As I work to develop CList I find my work on the software following this pattern: it's less about maintaining environments, coding functions, and designing interfaces, and more about decising on specifications and clearly articulating my requirements. For me that's a good thing; I have long felt that these are where my expertise lies, while the details of implementation have long been a barrier. Of course, a lot of other people are having the same experience.
Today: Total: Sebastian Elbaum, Nenad Medvidović, Medium, 2026/11/20 [Direct Link]This item is of interest because 'fairness' is one of the most commonly cited ethical principle for governments, institutions and organizations, and hence, widely cited in discussions of ethics and AI. Here, Mike Masnick writes about the EU's Digital Fairness Act proposal. The response is about what I would expect: while the intend is laudable, he argues, implementation would most likely be captured by larger players who would make the results worse rather than better the same way they turned a perfectly reasonable requirement to give users control over cookies into a constant source of annoyance. I don't really agree that this is a necessary outcome, and I think online consumer protection is overall a good thing.
Today: Total: Mike Masnick, TechDirt, 2026/11/20 [Direct Link]Web - Today's OLDaily
OLDaily Email - Subscribe
Web - This Week's OLWeekly
OLWeekly Email - Subscribe
RSS - Individual Posts
RSS - Combined version
JSON - OLDaily
Podcast - OLDaily Audio
Websites
Stephen's Web and OLDaily
Half an Hour Blog
Leftish Blog
MOOC.ca
Stephen's Guide to the Logical Fallacies
gRSShopper
Let's Make Some Art Dammit
Email: stephen@downes.ca
Email: Stephen.Downes@nrc-cnrc.gc.ca
Skype: Downes
Professional
National Research Council Canada
Publications
Presentations
All My Articles
My eBooks
About Stephen Downes
About Stephen's Web
About OLDaily
Subscribe to Newsletters
gRSShopper
Privacy and Security Policy
Statistics
Archives
Courses
CCK 2008, 2009, 2011, 2012
PLENK 2010
Change 11 - 2011
Education Futures - 2012
Learning Analytics - 2012
REL 2014
Personal Learning - 2015
Connectivism and Learning - 2016
E-Learning 3.0 MOOC - 2018
Ethics, Analytics - 2020
Mastodon
Mastodon OLDaily
Flickr Photos
Cosocial Pixelfed Amazon Wish List
Huffington Post Blog
Pocket Feed (RSS)
Pocket Feed (Tumblr)
CounterSocial
Services
Viewer
Backchannel
Ed Radio
gRSShopper Demo
Stephen Downes, Casselman, Canada
stephen@downes.ca
Last Updated: Nov 22, 2026 4:37 p.m.


