Eric Willey

What it's like to work with AI: a 🧵

As AI enters the operating room, reports arise of botched surgeries and misidentified body parts Medical device makers have been rushing to add AI to their products. While proponents say the new technology will revolutionize medicine, regulators are receiving a rising number of claims of patient ...

Companies are putting us at risk with half-baked AI. Whether on the factory floor or the operating room, we all need guardrails on the use of AI systems to protect our jobs, our freedoms and our lives.

@altmetric.com: "Wikipedia tracking enhancement: supporting the Cite Q template" https://updates.altmetric.com/announcements/wikipedia-tracking-enhancement-supporting-the-cite-q-template

(via the @wikicite telegram channel)

#wikidata #wikicite #wikipedia

Please welcome Tasavvur Magazine to Bsky; they're one of South Asia's few dedicated sff magazines, and they do great work. There are barely any eyes on them here at the moment, so if you're interested in global sff (or a desi reader looking for more desi sff), please do read/follow them:

The Black Box Myth: What the Industry Pretends Not to Know About AI | TechPolicy.Press Tech Policy Press fellow Eryk Salvaggio says it's a problem is that those of us outside of the AI industry don’t know what rules they are following.

"This “sense” of how things work elevates statistical modeling to a sacrament. Lack of evidence becomes the source of faith. A lack of control over a system can, to some, feel like evidence of a divine intelligence: this is the start of myth."

I so needed wombat zoomies right now.

U.S. Employee Engagement Sinks to 10-Year Low Since last year, 3.2 million fewer employees felt enthusiastic about and involved in their work, matching a figure not seen in the U.S. since 2014.

The issues listed directly tie to the causes of burnout and a healthy workplace
www.gallup.com/workplace/65...

The ethical case for resisting AI I had my moment of reality regarding generative AI in my chemistry and physics classes this past term. I’ve always had the student frustration “I can’t Google your homework problems!” as a point of…

Just in time for the AI policies to go into my syllabi - I've cleared my brain of all of its frustrations regarding generative AI.

I think it's important to resist the stuff. For all of us. And for no greater reason than preserving our humanity.

I think the stakes are that high.

More Than Words A veteran writing teacher makes a “moving” (Rick Wormeli) argument that writing is a form of thinking and feeling and shows why it can’t be replaced by...

Large language models have near-zero utility when it comes to helping students develop as writers. I say this because of what I value in the writing experience: thinking, feeling, human exchange and communication with readers. I make the full case in my book www.hachettebookgroup.com/titles/john-...

626 authors. 23 listed as authors and linked to their Wikidata items. And 603 author name strings that don't link to anything. I'm going to start going through it with Author Disambiguator (I know, I lead a very exciting life), but it's gonna take a while, and it's got me thinking about priorities.

I get it, I also like adding new items to Wikidata. But there are a lot of existing items that could use improvement, and I'm starting to think cleanup projects are more valuable.

Check out this item for a scientific article.

Writing is thinking.

It’s not a part of the process that can be skipped; it’s the entire point.

I haven't, I usually start with the author and make sure they have an item and then change articles (if I'm reading it right this would start with an article and edit the name strings), but that's interesting. I'll think about it, thanks!

I love Mad Magazine

It's a fun (especially on a cold rainy night) cleanup project that helps improve Wikidata and I can do while I'm watching a movie, and while a lot people like creating new items there's a lot to be said for improving existing items.

Thanks to everyone who worked on author disambiguator and fin!

They can be graphed because the metadata for articles she co-authored are *linked* to her Wikidata item. Because author disambiguator replaced the text strings in the Wikidata item for the article with her Wikidata item. It put the "linked" in linked data.

And to paraphrase Greg Proops, that's when this shit takes off. Check out Karen Collins on Scholia now: scholia.toolforge.org/author/Q5967...

Those topics, venue stats, that big swirly co-author graph (people love the big swirly co-author graph) etc. aren't being pulled directly from her WD item.

*And* author information is now linked because articles are now linked to the author items and any statements they include. Prior to this article items were probably linked to a journal title, but that was probably about it for linked data. Now it's part of a big juicy metadata web.

Here's a Scholia link to an article I used author disambig to replace name strings with Wikidata items for: scholia.toolforge.org/work/Q117422...

Which, okay, that's nice enough. But the point to me isn't the individual article I'm working from. It's that I'm also catching a lot of other articles.

Then you select the articles by the author you're working on, select the author (it provides a list of similar names) and it generates statements to automatically replace the text string in the article Wikidata item with the Wikidata item for the author. And your linked data is now more linked.

Things like co-authors, journal published in, etc. If a name keeps showing up in the same journal with the same co-authors on the same topic, it's likely it's the same person. And you can always check to make sure (links to ORCIDs are great for this, or going to the article for the author info).

Enter author disambiguator. You plug in a name, and it looks for name strings and clusters them (this is all explained fully with examples in the link above). You identify (or create) a Wikidata item for an author, and select articles with that name based on provided contextual metadata.

In other words for items about articles where author names are text strings, that "linked data" ain't linked because the import process doesn't know which "Jane Smith" Wikidata item to use. And a lot of people doing batch imports don't take the time to go back and do anything about that.

But when you pull in the author names it's going to be just a text string not linked to the authors' Wikidata items (if they even have one). It's better than nothing, but since you can automate import of text string names, there are a LOT of articles with text strings for author names.

One of my favorite Wikidata tools is the author disambiguator.

www.wikidata.org/wiki/Wikidat...

It's relatively easy to batch/automatically import some metadata for a scholarly article: title, issue, volume, date of publication, etc. especially if the journal has a DOI or other stable URL.

A screenshot of PayPal's terms of service. Big wall of legalese text, which I will be pasting in below. You don't need to read it, here's the important part: "The key update to the Privacy Statement explains how we will share information with merchants to personalize your shopping experience and recommend our services to you. Personal information we disclose includes, for example, products, preferences, sizes, and styles we think you’ll like. Information gathered about you after the effective date of our updated Privacy Statement, November 27, 2024, will be shared with participating stores where you shop." Here's the whole paragraph: We are updating our Privacy Statement to explain how, starting early Summer 2025, we will share information to help improve your shopping experience and make it more personalized for you. The key update to the Privacy Statement explains how we will share information with merchants to personalize your shopping experience and recommend our services to you. Personal information we disclose includes, for example, products, preferences, sizes, and styles we think you’ll like. Information gathered about you after the effective date of our updated Privacy Statement, November 27, 2024, will be shared with participating stores where you shop, unless you live in California, North Dakota, or Vermont. For PayPal customers in California, North Dakota, or Vermont, we’ll only share your information with those merchants if you tell us to do so. No matter where you live, you’ll always be able to exercise your right to opt out of this data sharing by updating your preference settings in your account under “Data and Privacy.”

In 2025 (less than 30 days away), PayPal will start selling your transaction history for targeted advertising.

I very highly recommend logging into your account and going to Settings > Data & Privacy > Personalized Shopping.

If you're reading this, turn that off RIGHT NOW before you forget.

Practical Solutions for Surviving Burnout, Iowa Library Association 2024 Today, I presented at the Iowa Library Association Conference on Burnout. It was my first time speaking about burnout, and exploring the research and popular books on burnout recovery was enlighten…

So many libraries are set up to burn out staff. Librarians - burnout should not be the expectation!

Burnout isn’t simply about feeling overworked or tired. It’s a sustained response to unmanaged workplace stress librarianbyday.net/2024/10/10/p...
#skybrarians

Remember when Archer gave those aliens that used people as baby incubators the complete works of Shakespeare? How much more fun would that episode have been if he gave them the complete works of Mel Brooks?

Eric Willey
I dunno, the Klingons almost start a war over spaghetti too.