iiinova — where your digital self keeps evolving

Reposted by Peter Johnson

aj-boston.neocities.org @aj-boston.bsky.social · 29/09/2026

[*a horrifying Eldritch creature ushers me into a literal Borgesian library*] Me: "This cannot be." Creature: "You find this unfathomable, don't you?" Me: "Was this built before 1990?" Creature: "It is timeless." Me: "I don't think it's exempt. You're gonna have to make this ADA accessible."

1187

Reposted by Peter Johnson

Alice Lastname @rbdfwkedkegfkt.bsky.social · 27/09/2026

This and the math thing in a way are good news. At the end of the day, LLMs and all their harnesses and bells and whistles still can't go beyond the distribution. They are great if you want to get to the mean. If you are out of the distribution, they can't replace you. They won't even know you.

065

Reposted by Peter Johnson

Ian Hussey @ianhussey.mmmdata.io · 25/09/2026

"Systematic reviews are only as credible as the studies they include" INSPECT-SR was published yesterday, giving concrete guidance on how to exclude untrustworthy studies. Cochrane's statement: www.cochrane.org/about-us/new... Article: www.cochrane.org/about-us/new... Guidance: inspect-sr.com

cochrane.org

New tool helps researchers spot problematic trials before they distort evidence | Cochrane

26031

Reposted by Peter Johnson

Lembas Eater @waybread.bsky.social · 24/09/2026

The problem with education is that it's a solved problem. Students need small classes, individual attention, lots of writing, lots of feedback, access to a well-stocked library, time to think, and time to talk with their peers. We just don't like the answer, so we keep trying to do other stuff.

10410139

Peter Johnson @libraryvines.bsky.social · 24/09/2026

Not even inadequate safety guardrails - generative AI inherently behaves in non-deterministic patterns. The seeming randomness of locating security vulnerabilities without (or in spite of) direction is always going to be present in this tech. It's what makes it both potentially useful and dangerous.

040

Reposted by Peter Johnson

Mike Rugnetta @mikerugnetta.com · 22/09/2026

"Looks like they know how to pronounce words in Tolkien books privilege."

1152

Peter Johnson @libraryvines.bsky.social · 17/09/2026

#medlibs "Engineers, data scientists, researchers, startups, and non-profits are invited to build AI systems that improve biomedical literature and clinical research data discovery across PubMed/PMC" www.nlm.nih.gov/pubs/techbul...

static.klipy.com

Alt: skeptical Fry meme

170

Peter Johnson @libraryvines.bsky.social · 16/09/2026

accounting for Boolean order of operation differences across different bibliographic databases is an underrated aspect of effective searching and @kastrel.bsky.social hits the nail on the head of the negative impact on retrevial of relevant results for systematic reviews (and other forms of KS/ES)

040

Peter Johnson @libraryvines.bsky.social · 16/09/2026

Updated to macOS Golden Gate, and I noticed the system is using AI to suggest titles to TextEdit files that I haven't yet applied a title to. The suggested titles are fairly accurate, but I also feel some discomfort at the intrusion. They are suggestions, but save as the file title if left unchanged

110

Reposted by Peter Johnson

bakindeb3.bsky.social @bakindeb3.bsky.social · 15/09/2026

In case you missed it, librarians won on one of America's biggest stages. A documentary about librarians standing up against book bans and defending the freedom to read was honored at the Emmys. via @everylibrary.bsky.social action.everylibrary.org/emmys

action.everylibrary.org

Librarians Won the Emmys!

We need to turn this moment of recognition into a moment of action.

0152

Reposted by Peter Johnson

Jason Kander @jasonkander.bsky.social · 11/09/2026

Raise your hand if today feels like the 25th anniversary of the day your childhood totally ended. It’s just a strange anniversary. Personally, I was a college student in DC who decided that day I would join the military and volunteer to go to war.

11797

Peter Johnson @libraryvines.bsky.social · 10/09/2026

Same. Always possible that someone has gone in so deep on their work that they can no longer check their own biases, and that their concerns do have some validity. Ironically, the typical hyperbolic nature of these claims usually doesn't help their case.

010

Peter Johnson @libraryvines.bsky.social · 09/09/2026

#medlibs - Registration is open for the MCMLA's one day remote 2026 Annual Meeting, "Spaces, Places, and People in Health Science Libraries" on Oct 1. Registration is free for MCMLA members ($30 annually, free for LIS students) and $30 for non-members. mcmla.org/2026-Registr...

mcmla.org

Midcontinental Chapter of the Medical Library Association - 2026 Registration

241

Reposted by Peter Johnson

N Sims @copyrightlibn.bsky.social · 08/09/2026

Re: possibility of Navier-Stokes proof scoop I have been saying for a while that I don't trust the purported "closed-loop" tools to actually be closed-loop, unless there is a real, -auditable-, guarantee on that.

2212

Peter Johnson @libraryvines.bsky.social · 06/09/2026

#medlibs - I’m excited to see and register for @nerdyicarly.bsky.social’s presentation, “Teaching AI in Hospital Libraries: Strategies, Critical Appraisal, and Practical Approaches”! www.nnlm.gov/training/cla...

nnlm.gov

Region 7 Presents - Teaching AI in Hospital Libraries: Strategies, Critical Appraisal, and Practical Approaches

141

Reposted by Peter Johnson

👻madeline bOoOoOoOodent👻 @oldenoughtosay.com · 25/08/2026

if you're in line to donate money to the Imagination Library, please stay in line

91542432

Reposted by Peter Johnson

Kate Saylor @ksaylor.bsky.social · 25/08/2026

If you are thinking about suicide, worried about a loved one, would like emotional support. Help is available 24/7 across the US. Call 988 988lifeline.org/get-help/

988lifeline.org

Get Help

Using the 988 Lifeline is free. When you call, text, or chat the 988 Lifeline, your conversation is confidential. The 988 Lifeline provides you judgment-free care. Talking with someone can help save…

012

Peter Johnson @libraryvines.bsky.social · 24/08/2026

“Media quiteracy and the role of AI refusal in information literacy programs” repository.ifla.org/rest/api/cor...

repository.ifla.org

020

Reposted by Peter Johnson

Lisa No Thank You @ohseafarer.bsky.social · 21/08/2026

Frankly the biggest barrier to my considering AI as a valid tool isn't the ecological or community impacts or even the blatant content theft but the way some people take its output as more valid than what an actual human expert says to their face when asked the same question.

26622

Peter Johnson @libraryvines.bsky.social · 20/08/2026

I have seen many faculty and academic leadership wanting to jump to the tool stage without considering foundational knowledge or competency around assessing or utilizing AI generated content in learning. They want practical applications, understandably, but also underestimate literacy needs.

060

Reposted by Peter Johnson

katiehoskins.bsky.social @katiehoskins.bsky.social · 19/08/2026

I used their guide to submit a comment on the proposed NSF policy to #defendresearch. Although #medlibs may not be submitting NSF proposals, we support researchers that do, and our institutions are impacted by grant funding.

012

Reposted by Peter Johnson

Anna Holmes @annabookwriter.bsky.social · 17/08/2026

Hey, guys, when the librarians get pissed about the destruction of books, it’s time to put those listening ears on. Because we are very comfortable with the idea that books are tools that can be retired. What’s happening right now is not that.

2540121399

Peter Johnson @libraryvines.bsky.social · 17/08/2026

I appreciate @librarian.bsky.social's recent post on "How to disable or avoid intrusive AI." Settings guidance is challenging to maintain, but it's worth struggling for. We should all have the right to avoid usage of tools we find undesirable. www.librarian.net/notoai/

librarian.net

How to disable or avoid intrusive AI

One of the biggest questions I get at Drop-In Time at the library (besides "what is taking up all my cloud storage?") is how to disable or avoid intrusive AI that shows up where people don't want it. ...

130

Peter Johnson @libraryvines.bsky.social · 14/08/2026

I appreciate @carrieprice.bsky.social’s thoughtful and responsible approach to experimenting with genAI in order to educate others on evidence synthesis reporting best practices and improve research quality. And of course I’ll be testing it out with my own search documentation 😉

130

Peter Johnson @libraryvines.bsky.social · 12/08/2026

"Examples of inadequate use of PRISMA 2020." Perhaps not the impact the different review teams were anticipating

120

Peter Johnson @libraryvines.bsky.social · 11/08/2026

#medlibs - To me these kinds of issuances represent one current challenge with AI governance. An overly generalized "we allow AI for searching" is as useful as saying "we allow computers for searching," and leads to inconsistent understanding and conduct for researchers.

060

Peter Johnson @libraryvines.bsky.social · 11/08/2026

#medlibs “As part of that shift, the OSF Projects workflow – the part of the OSF used for project collaboration, file management, and storage – is being phased out as an active workspace. This change does not affect persistence and accessibility of existing public projects.” www.cos.io/osf-changes

cos.io

OSF Changes | Center for Open Science

We are preparing substantial changes to OSF that will reduce its functionality, focus OSF on its unique strengths, and move toward an integrated model with complementary services. As part of that shif...

172

Reposted by Peter Johnson

Annie the Book @anniethebook.bsky.social · 10/08/2026

I'm grading my library research class's final papers & am delighted at how many students reported that they discovered that their topics were more complicated than they thought once they took a look at the literature. This is what college is about! The world is bigger and more nuanced than we knew!

1131

Reposted by Peter Johnson

Bill Grueskin @bgrueskin.bsky.social · 04/08/2026

Gotta love the “or stocks” part. Like Gemini realized its first answer was nuts so it tried to wedge in a better one.

Al Overview
DSt +2
"Buying the dip" means purchasing an investment like Dean's French Onion Dip or stocks after a price drop. ® TD Bank +2

551565225

Peter Johnson @libraryvines.bsky.social · 03/08/2026

#medlibs - this was a generated response to a query I have often used as one test to see how the commercially available LLMs are dealing with less straightforward MeSH terms, and the Claude (Sonnet 5) output got it in one. Need to find new test terms 🙂

an output from the AI model Claude that accurately generates a correct MeSH term for the concept of "hand surgery"

231

Peter Johnson @libraryvines.bsky.social · 30/07/2026

one way or another the road is going to end on this, and the economic results are likely to be devastating. I don't think AI related technology will go away as a result, but the current landscape isn't going to hold.

030

Reposted by Peter Johnson

New-Cleckit Dominie @ncdominie.bsky.social · 28/07/2026

This is just to say I have OCRed the text thet was in tne arch:ve and whicb yov were prabaldy seamhing for "breahtast" Fargme ms fheg mem debo:w5 sb ?mccf mmb 5c oc1b

7778512570

Peter Johnson @libraryvines.bsky.social · 28/07/2026

a permanent challenge with LLMs are their lack of transparency: you can only verify what actions are taken by their outputs. Because of that they are inherently insecure, and should never be used with PII or any item you wouldn't want publicly known. Also applies to workplace provided tools, frankly

030

Peter Johnson @libraryvines.bsky.social · 28/07/2026

Most of the criteria is meant for umbrella reviewers to aid in consistently including higher quality #systematicreviews, but this is also a useful measure for SR planning and evaluation: teams could use this list to identify gaps in their knowledge and planning, and increase the quality of their SRs

021

Reposted by Peter Johnson

Anna Holmes @annabookwriter.bsky.social · 27/07/2026

You guys. You can hold two feelings at once. An adaptation can be a good time and also Not rise to the themes of source material

1442752

Peter Johnson @libraryvines.bsky.social · 27/07/2026

a core of education is agency, and libraries are at the center of helping our users choose where and how they want to engage or not engage with AI. Especially in lieu of inconsistent governance I’m not surprised that many want to opt out techcrunch.com/2026/07/25/l...

techcrunch.com

Librarians are hosting viral 'Avoiding AI' workshops for people who are fed up with Big Tech | TechCrunch

At libraries around the country, "Avoiding AI" workshops have elicited unprecedented demand.

252

Reposted by Peter Johnson

ISSG Search Filters Resource @issg-filters.bsky.social · 27/07/2026

#medlibs we added a paper by Gagen et al. on development and validation of #searchfilters for #Transgender and #GenderDiverse Populations in #Ovid #MEDLINE and Ovid APA #PsycInfo sites.google.com/a/york.ac.uk...

0147

Reposted by Peter Johnson

Librarian by Day 🏳️‍🌈 @librarianbyday.bsky.social · 24/07/2026

Fellow professors, how do you have your students address you, and why? Title & last name? title and first name? First name? Something else? #edusky #academicsky #SLISsky

1242

Peter Johnson @libraryvines.bsky.social · 23/07/2026

I was today years old when I realized that in addition to showing an unlocked lock, the commonly used #openaccess logo contains both an uppercase "O" and a lowercase "a"

050

Peter Johnson @libraryvines.bsky.social · 22/07/2026

#medlibs - Reflecting on the potential for AI with cataloging and article indexing, I was reminded of my favorite PubMed mis-indexing example.

the metadata of a PubMed article with a title of "Bulls in a China Shop: Narcissism, Intragroup Conflict, and Task Cohesion" that has MeSH terms including "Animals," "Cattle," and "China" (the country)

4201

Peter Johnson @libraryvines.bsky.social · 21/07/2026

An interesting consideration of the benefits and limits of agentic AI systems in research. It left me with more questions about effective validation and oversight, but I appreciated the emphasis on the need for shared standards and operating environments media.oregonstate.edu/playlist/ded...

media.oregonstate.edu

OSU MediaSpace

000

Peter Johnson @libraryvines.bsky.social · 21/07/2026

one of the reasons we deaccessioned many of our library's physical titles was due to their age and the potential of containing out-of-date health information and guidance. Scary to think of them as "new" potential sources of AI training data.

25934