iiinova — where your digital self keeps evolving

Caspar van Lissa 🟥 @cjvanlissa.bsky.social · 09/09/2026

Why not a dozen 1-piece meals?

010

Reposted by Lisa DeBruine

`webexercises` creates interactive exercises for pkgdown, Rmarkdown, and with this update, Quarto (books), Latex output, and #R package vignettes! My hope is that this will support open educational materials, especially in R-package documentation. cran.r-project.org/web/packages...

cran.r-project.org

webexercises: Create Interactive Web Exercises

Functions for easily creating interactive web exercises in 'R Markdown', 'Quarto', and package vignettes that students can use in self-guided learning.

1104

I love the effort you put into this! I also had the same hypothesis about 3 being just the sum of 1 and 2, but 6 makes no sense.

141

Nope! They always ask me what pieces I want, and will let me order three 2-pieces instead of a 6-piece.

130

I like this clear overview of ai use in preparing a resource (by @hadley.nz) but am really struck by the difference in language used to talk about tasks done with Claude versus other programs. How does this anthropomorphism change how we take responsibility? tidydesign.substack.com/p/y-code-whe...

1. I started by asking Claude to download the video from youtube, since I can never remember how to do it myself.

2. Then I used MacWhisper to create a transcript, and saved the slides from my keynote presentation as individual PNGs.

3. Next I did a few rounds of cleanup with Claude, first asking it to read my slides (so it know what terms I’m saying), then asking it to fix transcription errors, break it into paragraphs, and remove disfluencies (like “ah”, “um”, etc). I used git and committed each step so I could see exactly what it was doing.

4. Then I turned the plaintext transcript in to markdown, enriching it with links and asking it to insert the slides at the appropriate place (which it did remarkably well).

5. At this point, I realised I gave a live demo, so there was a long chunk of text with no slides. So I asked Claude to extract screenshots from the video, based on the original timestamps from the transcript.

6. I finished it off with a thorough personal read, lightly tweaking the text to ensure it captured what I actually meant not what I said.

0100

My local chicken place’s chaotic pricing scheme winds me up an unreasonable amount.

Menu with per-piece prices written in manually

Chicken
1 piece - £1.90/per - £1.90 total
2 piece - 1.75 - £3.50
3 piece - 1.80 - £5.40
6 piece - 1.85 - £11.10
12 piece - 1.17 - £14.00

109312

I had it hidden recently so I had to do a few extra steps to get into it, but I unhid it because I was having useful academic discussions I didn’t want to miss. Probably time to hide it again!

100

I was pleased when most of my cool academic colleagues got on here and there was more interesting content, but lately every time I log on, I end up depressed about the world and contemplating just giving up and joining a commune with my extended polycule.

5220

Giles @gdeejay.bsky.social · 07/09/2026

I’m envious, I planned brief oral exams for my data skills students to explain their project code to me, but turns out it isn’t possible within my uni regulations.

110

Reposted by Lisa DeBruine

I heard someone say "Glad I didn't bother learning to program now that LLMs can do it." And my first thought was "Wait, wouldn't that same argument apply to learning to read too?" 🤨

3141

Institute for Replication @i4replication.bsky.social · 08/09/2026

I guess that’s what a no-reply originating address is meant to prevent. (And the barrage of OoO auto-replies when you email academics in August.)

130

Reposted by Lisa DeBruine

1/ What is the prevalence of personally identifiable information (PII) in author-collected data publicly shared in the social sciences? We audited 327 replication packages from 11 leading journals in economics, political science, and psychology Answer: surprisingly high 🧵

25537

Are they taking nominations? It would be funny to flood them with nominations for @datacolada.bsky.social

0120

Are allergy shots useful for cats now? I started them in the 80s and they told me it was pointless and just to avoid cats. But my parents got this wee guy last week and it would be nice to be able to stay with them when I visit.

130

Lisa DeBruine @debruine.bsky.social · 07/09/2026

That gives me a sort of proprioceptive imagery of how it feels to walk in snow so cold it squeaks, even though I haven’t had to in those temperatures since the early 2000s.

040

That’s what I mean. Among those who have *really* thought it through, “wants” is just shorthand for “what is selected for by evolutionary processes under these circumstances”. But you can’t use that kind of shorthand in writing or teaching. And it’s probably dangerous to thinking even for experts.

021

Lisa DeBruine @debruine.bsky.social · 07/09/2026

We have a similar problem in evolutionary sciences. It’s really hard to talk about a non-conscious process without anthropomorphising it. Among experts, it’s metaphorical and shorthand, but too likely to mislead non-experts.

131

Lisa DeBruine @debruine.bsky.social · 07/09/2026

I love the concrete examples of how to avoid anthropomorphism in this article: artificial intelligence → probabilistic automation image recognition → image labeling speech recognition → automatic transcription shows bias → reflects bias hallucination → undesirable output goal → success condition

39237

Reposted by Lisa DeBruine

Q. McLamore @ghawinriver.bsky.social · 06/09/2026

This is, incidentally, ALWAYS where the anti-trans playbook was going to lead, because they fundamentally conceptualized being trans as having a mental disability, and every last scrap of ableism they were covert with there, they're mask off on now. They must be rejected--in whole. Hobbes is right.

347990

Reposted by Lisa DeBruine

Trevon Logan @trevondlogan.bsky.social · 05/09/2026

If you know the details of the Claudine Gay (Harvard) and Jason Arday (Cambridge) controversies but little about the Marc Tessier-Lavigne (Stanford) controversy, you should ask yourself why. [And, yes, the answer is as simple as you think it is. This isn’t 3-dimensional chess...]

411329

Lisa DeBruine @debruine.bsky.social · 05/09/2026

And if you have robust estimates for these parameters and the effect size of interest, why do you need to collect more data?

031

Lisa DeBruine @debruine.bsky.social · 05/09/2026

And if you’re calculating power for anything more complicated than a t-test, you probably don’t have a great source of justification for most of the parameter assumptions.

2111

I got small town southern Ontario, which I guess is pretty accurate as I lived in Hamilton, ON for three years.

360

My PhD student is prepping a submission to CHI and they have the same requirement. But he and I aren’t eligible to review (having never had a CHI paper before), so his CS supervisor has to take on all 4 reviews.

100

But what is the carrot? Money might motivate ECRs, but as you get more senior, your time is worth more than the money they could pay reviewers. How do we fairly make this a part of the social contract of doing research? The amount of money publishers are making makes that really hard.

220

I did the distributed reviewing for the MetaScience ESRC grant call and found it reasonable and educational, but also had a team to divide the 7 reviews among (we were given 8 but had one conflict). This policy would absolutely make it way more difficult for a single applicant.

000

I would really worry that this sort of policy would encourage AI-generated reviews or very low quality reviews by people with insufficient expertise, just to fill the requirement.

340

But then I guess the argument is that the people most likely to be taking advantage of the system will also be likely to lie about the reasons they couldn’t review?

210

Jenny Bryan @jennybryan.bsky.social · 02/09/2026

I could support a requirement that your co-authorship team will provide the same number of reviews your paper requires (presumably 2) over the next year, and that if you don’t do this you need to explain in the cover letter of your next submission (with a policy to only punish egregious abuses)

331

Reposted by Lisa DeBruine

👉 "But in this new era, we must delete code even harder, because if we don’t constantly remove complexity, we are going to drown in generated kudzu slop that overtakes codebases." 🌿

04710

I’m not clear what that entails. Did they assign an agent to correspond with you about the conference?

110

Some potentially useful resources; #PsyTeachR section on naming things: psyteachr.github.io/ads-v3/02-re... @djnavarro.net’s project structure resource: djnavarro.net/slides-proje... @jennybryan.bsky.social’s talk on naming things: speakerdeck.com/jennybc/how-...

041

So much this! How do we solve it? Is there an existing good explainer (for Mac and PC) especially for new coders? Should we make one if there isn’t? This is how we address it in the #PsyTeachR Applied Data Skills book: psyteachr.github.io/ads-v3/02-re...

282

I have seen way too many students who are having odd errors that are explained by the fact that there are several versions of the file across several folders and they end up working on an arbitrary version each time they open it. R projects is now one of the first things I teach in data skills.

160

I’m so grateful for their contributions, and I would not have had time recently to do the coding, but I also don’t really have time to address the tech debt and worry it will eventually harm the project.

120

Yes yes this! — I am also looking at a mountain of tech debt for very useful new functions contributed by collaborators, but that it will take me as long to understand and refactor as it probably would have to code them myself (which would have been more fun).

150

Wow. This is incredibly convincing evidence of falsification with clear intent. Huge props to Klaus Wertenbroch, who is probably having a pretty stressful time.

0244