[Home] [Top] [Archives] [About] [Options]

OLDaily

Welcome to Online Learning Daily, your best source for news and commentary about learning technology, new media, and related topics.
100% human-authored

Your data, your rules: Firefox’s privacy-first AI features you can trust
Jolie Huang, The Mozilla Blog, 2025/06/23


Icon

The best way to run AI in the browser is "by running AI models directly on your device, ensuring your sensitive data remains local." So I'm in alignment with Mozilla's main aim here. This post introduces some of those new AI tools in Firefox, including alt text, translation and tab groups. Ah, but there are questions. Can this all be done in the smaller computers Firefox is often run on? Can Mozilla afford to support AI development when it's revenue flow has come into question? Can I turn any of this off? "Smarter and more intuitive" isn't what I always want from my browser. I'm just happy if it shows me a web page as it was written but without the ads.

Web: [Direct Link] [This Post][Share]


PubPub
GitHub, 2025/06/23


Icon

I received in my email today a post from Knowledge Futures titled "Not Enough: Open Infrastructure Funding and the Future of Knowledge Futures". It doesn't appear to be on  the web anywhere. Anyhow, it states KF will be "responsibly sunsetting PubPub Legacy, pausing our hosted Platform service, and reducing staff." The explain, "the communities we served with PubPub Legacy, an easy-to-use, free academic site builder, largely did not have the resources to support its active development." I had my criticisms when PubPub was launched in 2018, noting that the Public Knowledge Project had been doing the same thing since 1998,albeit of course, in Canada, not at MIT. The current iteration of PubPub Platform seemed to me to be a last-ditch effort to make the project pay. This link is to the GitHub repository for the PubPub code, for those who are interested. End-of-life for the platform is the end of 2025 and for the legacy code the end of 2026.

Web: [Direct Link] [This Post][Share]


30 minutes with a stranger
Alvin Chang, The Pudding, 2025/06/23


Icon

The idea is simple: pay 1500 individuals $15 to have an online conversation for half an hour with a stranger (total cost for the 1700 conversations would be $51000, by my calculations - must be nice to have that kind of money). Also, collect some demographic information about them, track how they say their moods change, document it all in something called the 'CANDOR Corpus'. Alvin Change puts a really interesting gloss on the study in this blog post, though he is not one of the study authors. What to make of this? I wonder what the conversations would look like if the participants were not American, not paid, not in the middle of a pandemic, and not aware they are being watched. I wonder how honest they are about reporting their feelings (especially given the implicit 'conversations make you feel good' finding from the study). I wonder if this was used to train AI. I wonder how many of these participants are in favour of the war. Via Alan Levine.

Web: [Direct Link] [This Post][Share]


The Cognitive Turn: Locating Cognitive Difference in the Age of AI
J. Owen Matson, Intralation: Culture, Theory, Pedagogy, 2025/06/23


Icon

In this erudite, insightful and detailed article Owen Matson explains why why AI discourse needs Katherine Hayles's theory of cognition. I had to switch my brain into neutral to read this one; none of it is wrong in any sense, but a lot of it is written in a language I don't use in my own thinking on the subject. It's a lot of work but the writing rewards the effort. Here's the core: "Cognition becomes an emergent property of systems that interpret information in context and connect it to meaning—not through language or logic, but through modulation, selection, and adaptation." This takes a lot of unpacking, and Matson leads us through it, and out the other side we get what might be an embodied theory of cognition (where it's left unsaid that this is where the real difference between human and AI cognition will be found). My criticism? This: we could tell exactly the same story without the use of the term 'meaning'. "Meaning is just relevance in context," writes Matson. "It's the fact that the interpretation leads to something that makes a difference." But what does 'makes a difference' mean? Something merely physical, like 'results in a change of state?' Or something requiring a mythical 'eye in the sky', like 'furthers the survival of the species'? Via Graham Attwell.

Web: [Direct Link] [This Post][Share]


Bluesky is more open than you think
Reddit, 2025/06/23


The statement in the title is true only if you think Bluesky is not open at all. On the (very unoffocial) decentralization scoring system, Bluesky ranks 14/100 (compared to email, which is 98/100). But this article does make it's case well, though it's heavy going. It describes Bluesky (correctly) as having three main parts:

Techies will note the similarity to RSS. Anyone can set up a PDS and it's not hard (technical, yes, but not hard). A relay is more difficult to manage simply because of the volume of data, but in theory, given a list of PDSs, anyone could create one. This article points to https://atproto.africa/, a second full-network relay run by the blacksky team (and the only other one I know of). The article also points to AppViewLite, which can be used instead of bluesky.app. Also on Lemmy. See also: There is no bsky.social instance.

 

Web: [Direct Link] [This Post][Share]


We publish six to eight or so short posts every weekday linking to the best, most interesting and most important pieces of content in the field. Read more about what we cover. We also list papers and articles by Stephen Downes and his presentations from around the world.

There are many ways to read OLDaily; pick whatever works best for you:

This newsletter is sent only at the request of subscribers. If you would like to unsubscribe, Click here.

Know a friend who might enjoy this newsletter? Feel free to forward OLDaily to your colleagues. If you received this issue from a friend and would like a free subscription of your own, you can join our mailing list. Click here to subscribe.

Copyright 2025 Stephen Downes Contact: stephen@downes.ca

This work is licensed under a Creative Commons License.