Matt Bury, MA Online & Distance Ed & AppLing 🇪🇸

YouTube video by IbanDlank How Transformers Work: A Detailed, Conceptual Explanation (No Coding / Math)

Idan Blank (UCLA, psych) makes the complex intuitive
if you want to learn how LLMs work, watch👇
newly posted to YouTube (no ads)
www.youtube.com/watch?v=cGMn...

Beyond belief: Reframing teaching as a science-based profession Some reflections on Douglas Carnine's classic article, "Why Education Experts Resist Effective Practices.”

My favourite read this month:

Hewitt, J., & Sachdeva, N. (2025, July 8). Beyond belief: Reframing teaching as a science-based profession [Substack newsletter]. The Science of Learning. scienceoflearning.substack.com/p/beyond-bel...

You've given me an idea for a pseudoscientific quack-therapy: Brain defragmentation. Reorganise the data in your brain to make it quicker to access!

Should come with ridiculous futuristic head-gear with wires & crystals & flashing lights coming out of it.

See: en.wikipedia.org/wiki/Defragm...

Cut the bullshit: why GenAI systems are neither collaborators nor tutors The rapid development of Generative AI (GenAI) technologies has led to widespread endorsement of GenAI systems serving as a ‘personal tutor’ and learning ‘collaborator’ in higher education. However...

Why we cannot trust LLMs.
Personally, I wouldn't change a word of this paper.

Cut the bullshit: why GenAI systems are neither collaborators nor tutors

Flenady, G., & Sparrow, R. (2026). Teaching in Higher Education, 31(1), 1–10. doi.org/10.1080/1356...

📣 "Linguistic Knowledge and Language Use" (CUP, 2024) now in Paperback!

The book is Open Access & can be downloaded for free from CUP's website. But if you prefer physical copies, this might interest you as it is 4x cheaper than the Hardback format!

Here: doi.org/10.1017/9781...

The Seductive Appeal of Discovery Learning - kirschner-ED The idea that children learn better if they discover something themselves is attractive for many different reasons, in spite of the fact that empirical research tells a different story.

The persistent and seductive appeal of discovery learning, despite its limitations and failures, reflects a mix of cognitive biases, ideological commitments, and cultural narratives rather than any strong empirical support. #EduSky
www.kirschnered.nl/2025/03/30/t...

If LLM-generated lesson plans or curricula are better or just equal to your own, it doesn't mean you should use them instead, it means you have some hard work ahead of you to improve your instructional design abilities.

Don't accept mediocrity.

Re: AI & LLMs in education, tech bros selling their products are like being on a date with someone who stalks you, is inappropriately intrusive, pushy, & manipulative, doesn't listen, & won't take no for an answer.

Pretty creepy when you think about it.

Learning as a Generative Activity | Cambridge University Press & Assessment

If you want to know why unaided independent studying, reading, & writing are so important for cognitive development & learning, this book will not only inform you but show you how to practise in optimal ways:

www.cambridge.org/gb/universit...

My favourite quote on GAI:
"One day, machines will exceed human intelligence." - Ray Kurtzweil
"Only if we meet them half-way." - Dave Snowden

another great paper from @mh-christiansen.bsky.social, showing that non-constituents* can be primed

It's more evidence that traditional linguists were mistaken to believe memory was in short supply:
Human memory is compressed, clustered, implicit and vast

Evidence for the representation of non-hierarchical structures in language - Nature Human Behaviour Language is often thought to be represented through hierarchically structured units. Nielsen and Christiansen find that non-hierarchical structures are present across reaction-time tasks, eye-tracked ...

@adelegoldberg.bsky.social Have you seen this?

Nielsen, Y. A., & Christiansen, M. H. (2026). Evidence for the representation of non-hierarchical structures in language. Nature Human Behaviour, 1–10. doi.org/10.1038/s415...

Yes, Tomasello has summarised & Nick C. Ellis continues to research the effects of frequency of exposure, complexity of forms, abstractness of meaning, & immediate salience to learners (prior knowledge & priming) in order to understand what "sticks" & what doesn't.

Industry Influence in High-Profile Social Media Research To what extent is social media research independent from industry influence? Leveraging openly available data, we show that half of the research published in top journals has disclosable ties to indus...

The pernicious intrusion into & effects on research of big tech is compromising our ability to understand what is really happening to all of us.

See: Industry Influence in High-Profile Social Media Research arxiv.org/abs/2601.11507

*How you really can't trust tech bros & their marketing teams*

ChatGPT is "hallucinating" again. I deleted my entire chat history a few days ago... & yet it tells me:

The explicit language teaching methodology "genre-based instruction" (Martin & Rose, 2012) succeeds in helping *all* students to develop mastery of the discourses of education because it is analytical rather than synthetic & puts meaning & purpose first.

In SLA we distinguish between a synthetic vs. an analytical curriculum.

Synthetic = Trying to reconstruct language from its constituent parts, i.e. syntax + lexis

Analytical = Treating texts as indivisible wholes & "zooming in" on its constituent levels of analysis (dual structure) in context

Ever wondered how the Dalmatian Dog effect works?

The human brain can quickly learn from a single experience and generalize it to related experiences — an impressive feat so far not matched by AI.

Our new paper in @natcomms.nature.com reveals how this works.
www.nature.com/articles/s41...

Project-Based Learning and Illusion of Learning TLDR: Project Based Learning (PBL) might satisfy the emotional and social aspects of being in a classroom but bypasses the cognitive conditions required for meaningful learning.

What Lev Vygotsky termed "empty verbalism" over a century ago & has been re-researched & repeated many times, & yet we're still encountering this highly problematic assumption, that doing = learning.

Article: umesko.substack.com/p/project-ba...

Well worth a read!

Doing things doesn't necessarily lead to learning things. Students may look like they're competent & knowledgeable but that's just an illusion; they may be merely following instructions &/or "aping" competent-like behaviour.

What You Might Not Know About Project Follow Through People often ask me to talk about Project Follow Through, a landmark federally funded study that compared major approaches to teaching young children at scale. They tell me they enjoy hearing …

A brief summary of how to create the necessary & sufficient conditions for teaching & learning to thrive: educationrickshaw.com/2026/01/31/w...

The automatically generated Google NotebookLM infographic has become the "word cloud" of the 2020s.

Event Self-Determination Theory (as much as I think it's overrated & misapplied) found that competence is the strongest motivator for learning. In other words, providing high-quality explicit instruction is the most effective way to increase learners self-efficacy and motivation.

RIP Sly Dunbar, 1952-2026. A legend & one of the most influential drummers of all time.

I think we're currently in the "radium skin cream" phase with AI.

When we discovered radioactivity, people tried to cram into everything, and it didn't end well.

AI is a discovery at least as important as radioactivity, and at least as dangerous.

I've had issues with taking self-report questionnaires at face value in education for some time now.

This study is just one example of how learning processes are frequently not available to the conscious awareness of learners & that we have to find more valid & reliable ways of gathering evidence

Quattrociocchi, W., Capraro, V., & Perc, M. (2025). Epistemological Fault Lines Between Human and Artificial Intelligence (No. c5gh8_v1). PsyArXiv. osf.io/preprints/ps...

References
Besigomwe, K. (2025). Algorithmic Self-Deception: How AI-Generated Feedback Skews Learners’ Self-Reflection. Cognizance Journal of Multidisciplinary Studies, 5(7), 333–342. cognizancejournal.com/vol5issue7/V...

"algorithmic self-deception" - an effect, "where AI-generated feedback’s vague positivity inflated learners’ self-perceptions and diminished critical reflection." (Besigomwe, 2025) ...