This item has caused ripples across educator social media (and increasing angst among English teachers). The story is in the headline (which is what a good headline should do). "People now appear unable to reliably distinguish human-out-of-the-loop AI-generated poetry from human-authored poetry written by well-known poets," write the authors, though of course people who have memorized Shakespeare can presumably tell the difference. "Furthermore, people prefer AI-generated poetry to human-authored poetry, consistently rating AI-generated poems more highly than the poems of well-known poets." In fairness, I'm no fan of the well-known poets; give me some contemporary Dylan or Swift any day.
Today: 319 Total: 319 Brian Porter, Edouard Machery, Nature, 2024/11/20 [Direct Link]Select a newsletter and enter your email to subscribe:
Stephen Downes works with the Digital Technologies Research Centre at the National Research Council of Canada specializing in new instructional media and personal learning technology. His degrees are in Philosophy, specializing in epistemology, philosophy of mind, and philosophy of science. He has taught for the University of Alberta, Athabasca University, Grand Prairie Regional College and Assiniboine Community College. His background includes expertise in journalism and media, both as a prominent blogger and as founder of the Moncton Free Press online news cooperative. He is one of the originators of the first Massive Open Online Course, has published frequently about online and networked learning, has authored learning management and content syndication software, and is the author of the widely read e-learning newsletter OLDaily. Downes is a member of NRC's Research Ethics Board. He is a popular keynote speaker and has spoken at conferences around the world.
Stephen Downes,
stephen@downes.ca,
Casselman
Canada
I'm sort of ambivalent about agents, mostly because I think they're really hard to get right. More optimistic people are predicting big things. "Going forward, what makes a System of Agents so effective at automating services is its ability to grasp context, apply reasoning, and generally function as a team of experts, each contributing unique knowledge and abilities. Software is no longer merely assisting humans. It is acting as an autonomous worker." Should such a future come to pass, I can only imagine the pushback as people argue it's dehumanizing to be forced to interact with a software agent. But hey, I just went to the ATM and then later bought concert tickets online. We'll adapt. Anyhow, this article offers a roadmap to future developments in software agents, and it's worth a read. See also: Salesforce's Agentforce.
Today: 294 Total: 294 Joanne Chen, Jaya Gupta, Foundation Capital, 2024/11/20 [Direct Link]It seems like half the learning and development community has started email newsletters on Substack, most of them engaged in the corporate sector. Substack filled a need, as the campaign against email spam has made it really difficult to run your own newsletter. It didn't hurt that you could charge people for subscriptions. But as Anil Dash warns, you don't really own your newsletter when you're on Substack. "Every single new feature Substack releases, from their social sharing to their mobile apps, is proprietary and locks you into their network." You'll also have some unsavory neighbours, warns Dash. "Here's how you can export your subscribers." He also points to a list of alternatives, but it's a Wired article behind a paywall. Here's the list: Ghost, Buttondown, Beehiiv.
Today: 212 Total: 344 Anil Dash, 2024/11/20 [Direct Link]I'm a little bit cynical about this one, since it leans into using AI for 'the basics' and also heavily into data collection; for example, they write "Eye-tracking, automated speech recognition, and brain-imaging capabilities can hone our measures of reading comprehension and fluency." For teachers, they're looking for systems that "can offer just-in-time suggestions and bite-sized research to help educators improve their teaching." Overall, it's the usual 'ed reform' agenda, only 'powered by AI'. But that said, it's a shopping list of projects that can be proposed and presumably funded, so if you're inclined that way, it's a useful report.
Today: 297 Total: 474 Kumar Garg, Renaissance Philanthropy, 2024/11/20 [Direct Link]I've referenced Tim Klapdor's discussion of learning types a couple times over the last few months. This post is the most complete description to date, framed within an open design system for learning. A design system, he writes, has three parts: principles (the motivations & drivers); language (words & visuals); and components (patterns & styles). The learning types typography is a part of the language being used. Klapdor "then built a library of common patterns related to each of the Learning Types" and applies them in a sample lesson. He and his colleagues have deployed them in "two fully online undergraduate programs consisting of 36 courses" with "3000 pages of content and 6084 hours of learning." There are videos and links to the learning types and learning patterns website available. What would be cool would then be tools for each pattern, maybe even a standard way to launch them.
Today: 88 Total: 460 Tim Klapdor, Heart Soul Machine, 2024/11/19 [Direct Link]Philip Kerr writes on the struggles of the Indian learening company Byju's. "Byju's is effectively 'worth zero', according to the founder, Byju Raveendran. Even before the World Cup, the company was on a losing streak, as financing dried up, and the bubble it had built was no longer sustainable. Key investors pulled out, and it was accused of ''toxic' targeting of low-income families, high-pressure sales, misleading advertising, and exacerbating existing educational inequalities'." Of course, the company's absurb valuation couldn't have helped. "It is a figure significantly greater than the entire Indian national budget for school and higher education." I'm sure Cory Doctorow would have a word for it, but to me it's just another example of the toxic influence of VC financing.
Today: 60 Total: 4907 Philip J. Kerr, Adaptive Learning in ELT, 2024/11/19 [Direct Link]Web - Today's OLDaily
OLDaily Email - Subscribe
Web - This Week's OLWeekly
OLWeekly Email - Subscribe
RSS - Individual Posts
RSS - Combined version
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
Stephen Downes, Casselman, Canada
stephen@downes.ca
Last Updated: Nov 20, 2024 11:37 p.m.