OLDaily

A First Pass At An ISTE Reflection
Tim Stahmer, Assorted Stuff, 2018/07/02

There's a little more from this year's ISTE. In this post, Tim Stahmer  argues that (at 15,000 participants) it's too big, and that it dominated by vendors. And he's still concerned about those smart badges. Gary Stager, also, raises concerns. "My greatest objection to being tagged like livestock was that it would only be a short matter of time before some bonehead referred to the fantabulous “Smart Badges” as educational technology," he writes. Of course, it already had been. And the key questions aren't being answered: who paid for the smart badges, and how much was paid? And who has the data

Web: [Direct Link] [This Post]

More States Opting To 'Robo-Grade' Student Essays By Computer
Tovia Smith, NPR, 2018/07/02

The thought of writing an essay that will be read by literally no one feels a bit odd to me, but I guess it's not a great stretch from those essays I wrote which were read by exactly one person. And "with computers already doing jobs as complicated and as fraught as driving cars, detecting cancer, and carrying on conversations, they can certainly handle grading students' essays," writes Tovia Smith. Essay-grading is at heart a categorization problem: does the candidate essay's feature set march most closely those of essays in class 'A', "b", 'C' or 'D'? The key is the feature set, which AI systems are taught (and preprogrammed) to recognize. And the test is in how well they grade - which, on balance, is pretty well. Good discussion of the issues, but I

Web: [Direct Link] [This Post]

500px photo site abandons freely shareable images with commercialization push
Stephen Shankland, Cnet, 2018/07/02

This is a warning sign for people sharing free and open resources: they can disappear in a flash. In this case, photo sharing site 500px deleted a million Creative Commons licensed photographs with almost no warning. As reported by Michael Zhang, "overnight, all of the CC photos that have been uploaded since 2012 have been nuked from 500px. Users can no longer choose a CC license during uploading, search for CC photos, or download them." Internet Archive volunteers "rushed at the last minute to preserve all the CC photos hosted on 500px, allowing 3 terabytes of photos to be saved." And as Creative Commons CEO Ryan Merkley said, "it's disingenuous to suggest CC works didn't fare well on the platform when they weren't given the same priority other platforms like Flickr give them." Hey, I like Flickr, but I keep a backup of every photo I have, just in case. Via Digital Koans.

Web: [Direct Link] [This Post]

Online & Blended Learning: Selections from the Field
Online Learning Consortium, Routledge, 2018/07/02

This eBook is actually a collection of excepts from other books. There isn't a direct PDF download - the PDF is "generated" when you click a button on the page (it's like publishers can never do anything in a straightforward manner - there always has to be something fishy or skeevy about the download). As I read through the eBook (which did not require that I sign in - yay) it felt odd to me thinking about all these concepts being read in books as opposed to the blog posts and white papers where they originated on the web (and originated there some number of years ago). Anyhow, there are interesting bits about the process of blending learning, about what you can do online that you can't do on campus, and about OERs and blended learning.

Web: [Direct Link] [This Post]

“I Was Devastated”: Tim Berners-Lee, the Man Who Created the World Wide Web, Has Some Regrets
Katrina Brooker, Vanity Fair, 2018/07/02

This is a superficial look at the effort to re-decentralize the web. It focuses mostly on Tim Berners-Lee and his Solid distributed web application. " The system aims to give users a platform by which they can control access to the data and content they generate on the Web. This way, users can choose how that data gets used rather than, say, Facebook and Google doing with it as they please." This project has been in the works for several years; I've covered it previously. The article also mentions some other decentralized web projects, such as Mastodon and Peertube.

Web: [Direct Link] [This Post]

Is strong AI inevitable?
Peter Sweeney, Towards Data Science, Medium, 2018/07/02

'Strong AI' in this article is characterized as AI that cannot only make predictions, as today's AI can, but as AI that can also create explanations. This is what is needed to progress from pre-scientific reasoning to scientific reasoning. The creation of explanations poses unique challenges to AI because it consists of things like interpretations (that is, things like models and world views) and formalisms (like math and language and other abstractions) along with predictions. Could AI produce these? Sure - humans do. But how likely is it? In my view, pretty likely. Interpretations and abstractions aren't magical things that appear from nowhere. They are the result of a knowable computational process. So I think computers will begin to be able to explain things. And as the clickbait headlines say, what they say will surprise you.

Web: [Direct Link] [This Post]