Toward a Definition of Open Source AI
David Wiley,
improving learning,
2024/05/17
Just a quick update from David Wiley on the Open Source Initiative (OSI) attempt to own, er, I mean, steward, the definition of 'open source AI'. Wiley notes "The definition is currently in its eighth draft, with the goal of finalizing the definition by October, 2024." Why am I so sceptical of the OSI initiative? Well, the first sentence contains the phrase "massive benefits accrue" and everything else is based on this, as though there would be no point to open AI if it didn't make money. Yet (ironically) there's no requirement in the four freedoms (use, study, modify, share) that these benefits be distributed equitably, or that they not deprive one of any existing freedom, or cause no harm in their application. Moreover, 'open source AI' is "made available", and not (say) developed openly, or developed by a community; it's as though it's some sort of creation that comes down from on high. Defining 'open' economically, and then cleaving it from social justice and any sort of community-based process, is the sort of 'open' we might find in Silicon Valley, but not really one we want in our own communities.
Web: [Direct Link] [This Post][Share]
Evaluating Neural Representations
Mazviita Chirimuuta,
The Brains Blog,
2024/05/17
OK, I haven't read this, but here's the whole book (open access publication) and I'll be spending the next little while dipping in and out of it (any interest in forming a reading circle?). For now, let me focus on this one blog post that extracts one little bit of 'The Brain Abstracted: Simplification in the History and Philosophy of Neuroscience'. Mazviita Chirimuuta writes, "the positing of neural representations abstracts away from the causal chain mediating the distal object and the firing it elicits, hence it is a simplifying strategy." Quite so. But - importantly - it follows that neural representations themselves have no causal power. You can't say 'his belief that P' caused 'his conclusion that Q' or 'his behaviour of R' (the map isn't the territory; Canada isn't bigger than the U.S. just because it's bigger on the map). Anyhow - I'm excited to see this book, and even more so that it is open access. It's like I can do philosophy again!
Web: [Direct Link] [This Post][Share]
Blog + Newsletter
Doc Searls,
Doc Searls Weblog,
2024/05/17
You'd think, with newsletters being "all the rage now", that I'd be seeing an increase in subscriptions. Why, early adopters "blogged here two years ago about the idea of writing a solo newsletter." But it's not so simple. What's actually happening is that a few platforms (most notably Substack) with some good marketing popularized the idea of bloggers and columnist using a 'newsletter platform' to write. There would be monetization, of course. You could also use WordPress (use the Noptin plugin) or Ghost but you'll have to set up or find a mail transport service (I use MailGun; I wrote my own client for it, but the WordPress plugin will do this for you). But hey - if you do think email newsletters are all the rage, you can still subscribe to OLDaily some 23 years after the first email left the server. Just saying.
Web: [Direct Link] [This Post][Share]
View of Alone in the Academic Ultraperiphery: Online Doctoral Candidates’ Quest to Belong, Thrive, and Succeed
Efrem Melián, Julio Meneses,
The International Review of Research in Open and Distributed Learning,
2024/05/17
"Online doctoral candidates constitute an ultraperipheral population in the academic landscape," argue the authors. "Online delivery was viewed as both a blessing for the accessibility it enabled and a curse due to pervasive feelings of isolation and virtually non-existent peer networks." I would suggest that these peer networks exist - just look at GO-GN - but the students' "strong drive for participation, sometimes matched by the supervisor but rarely supported by the institution" renders them invisible. In any case, it would take a study of more than 24 part-time online students to reveal anything in the way of trends (you can't say, as these authors do, that "they often feel invisible and neglected by the institution" - you don't know that and you haven't shown that). But to the extent that graduate students are sometimes not able to find online support networks, that is indeed something that should be addressed.
Web: [Direct Link] [This Post][Share]
Brain-computer interface experiments first to decode words 'spoken' entirely in the brain in real time
Bob Yirka,
Medical Xpress,
2024/05/17
The actual news here (because we've covered similar work in the past) is that the researchers probed "the supramarginal gyrus, a region of the brain never before tested with brain–computer interface (BCI) technology." The accuracy of the six-word test wasn't great - in one patient, barely above what would be accomplished by chance (23%, as compared to the one-in-six (15.5%) a 50-50 coin flip would produce). This part of the brain, also known as Brodmann area 40, and (per Wikipedia) "interprets tactile sensory data." Here's the full study by Sarah K. Wandelt and colleagues. Implications: first, that computers could (potentially) read thoughts; and second, what they 'read' is the words as they are 'spoken' in one's inner voice (which, I would suspect, more or less clearly articulated in different people).
Web: [Direct Link] [This Post][Share]
Archie, the Internet’s first search engine, is rescued and running
Kevin Purdy,
Ars Technica,
2024/05/17
"It's amazing," writes Kevin Purdy, "and a little sad, to think that something created in 1989 that changed how people used and viewed the then-nascent Internet had nearly vanished by 2024." The think in question was a search engine that allowed users to find documents made available through anonymous FTP services. But it has been rescued. "Not only did The Serial Port rescue the last working version of Archie (seemingly a 3.5 beta), but they posted its docs and now run an actual Archie server on an emulated Sun SPARCstation 5." Takes me back.
Web: [Direct Link] [This Post][Share]
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 2024 Stephen Downes Contact: stephen@downes.ca
This work is licensed under a Creative Commons License.