OLDaily

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

Support OLDaily
2026/04/01

---

OLDaily has been free and open to all readers for almost 25 years. No paywalls, no ads, no sponsor copy. Keeping it that way takes real time — every weekday, year-round.

If OLDaily has been useful to you — saved you time, pointed you somewhere important, or helped you think through a hard problem — please consider a monthly contribution. Every donation helps keep OLDaily independent and free for everyone.

Funding for OLDaily is part of my retirement plan - see my Retirement FAQ for more details. To continue publishing OLDaily I have a target of 100 base-level subscriptions per month (or equivalent funding from any source). I'm currently at 3% of this target. To add to that support, visit this page

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

What AI Asks of Open Access
Alison Mudditt, The Scholarly Kitchen, 2026/04/01

---

Good article by the CEO the Public Library of Science (PLOS). "As AI systems increasingly reason from the scientific literature, the integrity signals that make research trustworthy - open data, structured metadata, robust retraction processes - matter more than ever." Some publishers say peer review makes these distinctions, but "not every publisher requires open data and open methods as a condition of publication," and these need to be open to make those distinctions auditable. Instead, some publishers choose to lock these behind a paywall - but "conditioning access on institutional control, demanding compensation is not just strategically self-defeating but a values failure." If we want trustworthy AI, we need open access publishing. It really is that simple.

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

Exploring AI Powered Newsletters
Miguel Guhlin, Another Think Coming, 2026/04/01

---

There's no danger of OLDaily becoming an AI powered newsletter. I'd rather shut it down entirely, because what matters in OLDaily is the voice. But that doesn't matter everywhere, and it's easy to think of circumstances where you want to keep people informed but don't need to do a lot beyond that - perhaps (hypothetically?) the 'Room 14 Weekly Update' by 3rd Grade teacher Ms. Rivera. Miguel Guhlin demonstrates just such an AI newsletter builder, and the results are, perhaps, good enough. I like this post and presentation - it isn't for me, but stuff like this could become commonplace. It depends on it being accurate, to the point, and just human enough to make the connection.

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

Measuring What Matters: From Blunt Sorting to Human Thriving Powered by Multimodal AI
Eric Tucker, Getting Smart, 2026/04/01

---

I have nothing to do with the event being advertised in this post and probably wouldn't recommend it. But the title prompted me to want to make what I think is a pretty important distinction (one I'm quite sure the authors had no intention of making). Eric Tucker writes, "Today, emerging breakthroughs offer an unprecedented opportunity to shatter traditional testing constraints... Advancing Multimodal AI, Measurement, and Assessment Innovation." Now I had just finished reading Julian Stodd's post describing a 'value architecture' and was thinking how in today's society everything is reduced to quantification by number (and hence, money). A Value(s) kind of thing. But - I was thinking as I saw this - AI doesn't measure the way (say) an instrument or meter of some sort measures. It senses (which made me think of Harold Jarche). It is at once much more fine-grained than any instrument and at the same time far more coarse. We think of 'machine assessment' as though a calculator were doing it. But human sensesation - a capacity to recognize - is probably a much better metaphor. 

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

The SOAP Model for Learning Strategy
2026/04/01

---

If you wanted to take the venerated OODA loop (Observe-Orient-Decide-Act) and create a business-focused version specifically for learning and development, complete with its own catchy acronym, you'd probably come up with SOAP, which stands for Situation-Objectives-Approach-Performance. And that's what WeLearn has done. The link here is to an 8 page PDF. I'd say more, but they've disabled 'select text' and I don't feel like retyping it.

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

What is inference engineering? Deepdive
Gergely Orosz, 2026/04/01

---

This is a really good article describing a new sort of job role called 'inference engineering'. This is the person who create the layer between the set of weights that constitute the 'model' in an AI system and the application surface where the output is intended to be applied. Here's a really simple of my own design illustrating this: a system, like ChatGPT can autocomplete sentences, but always gets this one wrong: "I'm on my way home from...". That's because it doesn't know where I am. An inference engineer would create the structure that lets ChatGPT figure out my location to fill in the sentence properly. In yesterday's newsletter we saw the (accidental) release of Claude Code's inference layer, including all the tools it can use to interact with the terminal and other systems the user is using. If this interests you at all, definitely read this article.

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

The Conviction That In The End Nothing Matters Except People
Josh Brake, The Absent-Minded Professor, 2026/04/01

---

This article takes us from the idea of the MOOC as " making high-quality lectures and course materials widely available" to an image of the university as a monastery. I think Josh Brake gets some things right along the way - criticizing xMOOC model (our MOOCs were never like this, of course) and pointing to the difference, raised by 'work-related' and 'control-related' technologies (which reminds me of my own distinction between free learning and control learning). And this is right: "we should think primarily not about how our students should directly use a given technology, but how that technology can be used to coordinate and facilitate the set of things that they are doing." But I think he he takes a wrong turn when says "what we want our students to do in our classes is not primarily to build a set of skills, but a certain set of character traits that enables them to push themselves to grow." Maybe I'm just misreading what he means by 'character traits' (but the monastic turn is not reassuring).

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

FoundationalASSIST: An Educational Dataset for Foundational Knowledge Tracing and Pedagogical Grounding of LLMs
Eamon Worden, Cristina Heffernan, Neil Heffernan, Shashank Sonkar, arXiv.org, 2026/04/01

---

I wonder whether this is a new type of open educational resource (OER). Composed of questions and answers from Illustrative Mathematics 6th - 8th grade math curriculum, FoundationalASSIST contains "full question text, actual student responses (not just right/wrong), records of which wrong answers students chose, and alignment to Common Core K-12 standards." There are 1.7 million interactions from 5,000 students. To access the database, submit a request here. The dataset is licensed under CC-BY-NC-4.0 but "you need to agree to share your contact information to access this dataset. This repository is publicly accessible, but you have to accept the conditions to access its files and content." E-Trials provides access to a number of similar datasets. Via Learning Engineering.

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

Potential Serendipity over Expectations of Gratitude
Alan Levine, unitwin-unoe, 2026/04/01

---

There's a thing called Postel's Law that states "Be liberal in what you accept, and conservative in what you send." The idea is that when we design systems we should expect and plan for variable (and sometimes plain wrong) input while at the same time being as stringent as possible when sending output. I take this to be Alan Levine's expression of the same principle when it comes to attributions: "We do not need mechanisms to deliver gratitude, we ought to make it our regular action. Putting aside expectations allows the serendipity of receiving unsolicited thanks much more powerful." So - similarly - I don't expect people to fall over themselves expressing gratitude. The idea - or the art - is what matters. At the same time, when I get something from someone, I typically use a 'via' attribution, so people know where I got it. And it's for the same reason as Alan Levine - I want to give people a route they can travel through to maybe get some serendipity. Following links is the soul of the web; being conservative on the web means being liberal with attributions.

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:

• Read in your web browser on this web page, updated weekdays
• Receive a daily or weekly email newsletter by subscribing here
• Subscribe to the RSS feed using your favourite feed reader
• Follow @OLDaily@mastodon.social on Mastodon or elsewhere in the fediverse
• Follow @OLDaily on Bluesky
• Read OLDaily as a LinkedIn Newsletter
• Integrate the open API using the JSON feed