iiinova — where your digital self keeps evolving

Eduardo Leoni @domleoni.bsky.social · 04/10/2026

Norways continues to be the best/worst example of this.

010

Eduardo Leoni @domleoni.bsky.social · 23/09/2026

Not the same, of course. But it is a general phenomenon, it just happens to hit some authors really, really, hard.

010

Eduardo Leoni @domleoni.bsky.social · 23/09/2026

A professor of mine called this the "New York Times" effect.

100

Reposted by Eduardo Leoni

William B. Fuckley @opinionhaver.bsky.social · 16/09/2026

A minimum population size for local government in general would solve literally hundreds of problems. Lots of cities should be annexed or disincorporated.

461127136

Eduardo Leoni @domleoni.bsky.social · 05/09/2026

I got New York. I am actually Brazilian living in Brazil, but lived there for 8 years, so I think it got it right. Boston was high up, and I lived there as well.

000

Reposted by Eduardo Leoni

daniel fosco @danielfos.co · 26/08/2026

Your weekly reminder that every single paid speech-to-text app on MacOS is a scam – they all cost $0 to run because they're using a free local model. Download the free handy.computer or any of the many open source alternatives and go get a coffee.

handy.computer

Handy

Handy is a cross platform, open-source, speech-to-text application for your computer

715739

Reposted by Eduardo Leoni

Rodger Sherman @rodger.bsky.social · 15/08/2026

Lowkey excellent thing about Wikipedia: accurately describing the extremism and lies of right-wingers because its sole concern is provable truth, while the media tiptoes around accurately describing whats happening to avoid appearing “biased” against the extremists and liars

223391734

Reposted by Eduardo Leoni

Ian Hussey @ianhussey.mmmdata.io · 14/08/2026

I’ve been told for years that it’s legally and ethically impossible to share medical RCT data. In 2024, the BMJ made sharing anonymised data mandatory, and now lots* of folk do it. Interesting how the barriers evaporated like that, isn’t it? *not everyone, compliance problems exist.

medrxiv.org

Availability of clinical trial individual patient data in the BMJ before and after adoption of a stringent data-sharing policy, compared with recent rates at other major medical journals

Background In 2024, the BMJ updated its data-sharing policy for clinical trials, requiring open deposit of deidentified individual participant data (IPD) before publication. We considered whether data...

510040

Eduardo Leoni @domleoni.bsky.social · 12/08/2026

I just wrote to @gruber.foo . Expecting a reply? Not really. But I haven't felt this way since contacting professors about PhDs in the US at the turn of the century. It's about this snailapp.ink/manifesto.html

snailapp.ink

The Snail Manifesto

A manifesto for slow, turn-based correspondence — why a fixed schedule dissolves the guilt of the unanswered message.

000

Reposted by Eduardo Leoni

Carlos Scheidegger @cscheid.net · 09/08/2026

Nail this on the fucking doors of Those People. You know who they are.

A screenshot of an excerpt from the linked document. It reads:

Of course, LLM outputs are not entirely empty of meaning. They can reason through open math problems, after all, so synthetic text must convey some valuable suggestions and correct claims. But they often fail to comply with Grice’s rules of pragmatic communication because they are not engaged in a cooperative exchange. LLMs will generate extraneous filler, add irrelevant caveats, confabulate, and compress claims beyond easy recovery. To find a nugget of meaning, the user must sift through mounds of slop—and some simply won’t bother. They’ll serve the slop unfiltered to another party, who performs this interpretive labor themselves, only to discover that the underlying meaning does not exist. When someone expects me to perform their share of this interpretive labor, I get angry. Grice’s rules are, in essence, rules of etiquette—their unwarranted violation is rude.

0234

Eduardo Leoni @domleoni.bsky.social · 08/08/2026

"Of course! I used glm(formula = y ~ x, family = "binomial", data = survey, weights = w)"

110

Eduardo Leoni @domleoni.bsky.social · 01/08/2026

Brings me back memories of trying to reshape a dataset using f*ing spss in my master thesis in 1999. In the Phd reshape was a revelation, (a month of work in seconds) but I always, always, struggle a little bit to use it. Not a Stata user for the past 15 years, but this does seem nice.

010

Eduardo Leoni @domleoni.bsky.social · 16/07/2026

Different men for the same con.

010

Reposted by Eduardo Leoni

Daniel Benneworth-Gray @danielgray.com · 13/07/2026

Sam Neill on tackling depression and imposter syndrome when you’re between jobs.

65107543680

Reposted by Eduardo Leoni

Dirk Eddelbuettel @eddelbuettel.com · 11/07/2026

Norm Matloff (1948-2026) @matloff.bsky.social I was made aware of the post below made by his daughter. We lost another friend and valued long-time contributor the #Rstats community, gone far too soon, and sure to be missed. His memory will be a blessing.

Screenshot of post on X in the account of Norm Matloff in which his daughter announces his passing. Gone too soon, and sure to be missed.

313038

Reposted by Eduardo Leoni

NANCY COMICS BY ERNIE BUSHMILLER @nancycomics.bsky.social · 06/07/2026

Nancy By Ernie Bushmiller June 04,1956

P1- NANCY IS RUNNING WHILE CARRYING HER CAMERA

P2- SUDDENLY NANCY STUMPED HER TOE ON A ROCK CAUSING HER TO DROP HER CAMERA

P3- NANCY HAS JUST LEFT A SHOP WHERE SHE HAD THE FILM IN HER CAMERA DEVELOPED , AND SHE IS PUZZLED BY A PHOTOGRAPH

P4- NANCY HAS TAKEN A CLOSE UP OF A VERY PLEASANT LITTLE SNAIL

231555281

Eduardo Leoni @domleoni.bsky.social · 29/06/2026

Make that a *knowledgeable* human.

010

Eduardo Leoni @domleoni.bsky.social · 28/06/2026

Cool! Claude modified the scripts to give a nice snapshot of some latinamerican countries since 1980.

000

Eduardo Leoni @domleoni.bsky.social · 22/06/2026

@cuducos.me Se tiver um tempinho, dê uma olhada.

000

Reposted by Eduardo Leoni

Justin Wolfers @justinwolfers.bsky.social · 22/06/2026

This is not the prediction markets future any of us were hoping for. www.cnn.com/2026/06/21/b...?

cnn.com

911821

Eduardo Leoni @domleoni.bsky.social · 22/06/2026

8/ If there's one person you've been meaning to write to properly for months, a parent, a kid, an old friend a few time zones over, snail might be for you. Duets or small circles, one letter at a time. Invite-only while it's small. Read more or get an invite: snailapp.ink 🐌

snailapp.ink

snail

100

Eduardo Leoni @domleoni.bsky.social · 22/06/2026

7/ Ly and I are a few letters in now. They're the right length: longer, more considered, not a text. They're not too frequent. But they're real, and they keep arriving. On their day, like clockwork. That's the whole thing I'd been missing.

100

Eduardo Leoni @domleoni.bsky.social · 22/06/2026

6/ And it keeps nothing. No "letters" database, no archive on my servers. Your letters live in Telegram, which serves as the infrasctructure. snail just schedules the rhythm.

100

Eduardo Leoni @domleoni.bsky.social · 22/06/2026

5/ That one little rule fixed something in my head. An unanswered text feels late. A snail letter is just waiting for its day. I never wonder if it's too soon to write again, or feel guilty for being slow. The rhythm carries it, so I don't have to.

100

Eduardo Leoni @domleoni.bsky.social · 22/06/2026

4/ So, being the kind of person who'd rather build a tool than do the work, I built snail 🐌. Two of us take turns: I write one week, Ly writes the next, each letter publishing on a fixed day we both know. No replies, no thread, no typing dots. The deadline belongs to the schedule, not to me.

100

Eduardo Leoni @domleoni.bsky.social · 22/06/2026

3/ Here's what I landed on: email freed us from waiting on the mail. Messaging freed us from waiting at all. But both lost the rhythm: the shared, expected moment, the way mail once came at a known day and hour, and you'd look forward to it. Yup, I'm that old.

100

Eduardo Leoni @domleoni.bsky.social · 22/06/2026

2/ It's not that I don't want to write to Ly. But I want to write properly, not a thumbs-up. And I don't want them to reply right away either. If they answer immediately, the pressure's back. If I answer immediately, I hand it back to them.

100

Eduardo Leoni @domleoni.bsky.social · 22/06/2026

1/ My kid Ly lives a long way from me. I live in Brazil, they live in the US. We love each other, but we're both bad at staying in touch. A message sits unanswered for days, sometimes weeks, and the longer it sits the heavier it gets.

100

Eduardo Leoni @domleoni.bsky.social · 14/06/2026

When I lived in the US, the greatest, by far, advantage of Telemundo is that it didn't have a fucking 20s delay to do what I do not know (cut expletives???). And this was in the pre-streaming years of this century. I fucking hate people cheering for goals 20s before me.

000

Eduardo Leoni @domleoni.bsky.social · 06/06/2026

I've done 24, 72, 168. Never experienced this clarity and focus that so many talk about. Didn't find it *that* hard either. Perhaps related, not sure. It's a waiting game, hunger comes in waves and passes. It's a good lesson.

000

Eduardo Leoni @domleoni.bsky.social · 01/06/2026

"~CSS~[insert random tool here] is tricky and difficult to reason about, and trying to rush a fix is never the optimal outcome." Good stuff! And I should follow this rule sometimes.

010

Reposted by Eduardo Leoni

Anthony Moser @anthonymoser.com · 10/04/2026

and that feeling of "oh no" then visceral revulsion settling into deep distrust? that's how it feels when someone gives me slop. like you just handed me a cup of water, and i'm looking at the sink going "....where did you find that brush"

280044

Eduardo Leoni @domleoni.bsky.social · 29/05/2026

What about if you include fixed effect without interacting them with every other variable in the model? What happens then?

100

Eduardo Leoni @domleoni.bsky.social · 26/05/2026

It is "operação padrão" in Portuguese. The revenue service in Brazil (RFB) is very fond of the strategy.

120

Eduardo Leoni @domleoni.bsky.social · 25/05/2026

Me too! I rarely even glance on the comments, even when they are on, unless I need some info.

010

Reposted by Eduardo Leoni

rOpenSci @handle.invalid · 19/05/2026

📢 The call for submissions for #LatinR2026 has been extended! There’s still time to share your projects, experiences, and work with the R community in Latin America 💜 🗓️ New deadline: June 1 ✍️ openreview.net/group?id=LATIN-R.com/2026/Conference 🔗 latinr.org 📍 […] [Original post on hachyderm.io]

066

Reposted by Eduardo Leoni

R-Podcast (Eric) @rpodcast.bsky.social · 19/05/2026

The R/Pharma Hangouts are about to get Nix-ified! We have the amazing @brodriguesco.bsky.social joining myself and Jared on May 19th to share not only the power of {rix} & {rixpress}, but also the brand-new T language for polygot data science workflows. As a Nix addict, I'm really excited 😀

051

Eduardo Leoni @domleoni.bsky.social · 18/05/2026

I understand the outrage. But this fake hallucinations problem is destined to be solvable by AIs pretty soon. And even if it isn't, the incentive structure will solve it as soon as journals and peers follow this rule. And it wont mean much. Goodhart's law is a tough law to break.

120

Eduardo Leoni @domleoni.bsky.social · 18/05/2026

Honest question: should I be dead by now?

100

Eduardo Leoni @domleoni.bsky.social · 18/05/2026

You think people at stackoverflow were mean? There was a move to use stackoverflow by some the R community back in the day because users were intimidated by the r-help mailing list.

200

Reposted by Eduardo Leoni

F. Javier Rubio @fjrubio.bsky.social · 18/05/2026

This change went under my radar. Heads up for #rstats users: as of devtools 2.5.0 03/2026, install_github() and all other install_*() functions are deprecated in favour of pak. Full migration guide at: devtools.r-lib.org/reference/in... #github

devtools.r-lib.org

Deprecated package installation functions — install-deprecated

These functions have been deprecated in favor of pak, as that is what we now recommend for package installation. There are a few functions which have no pak equivalent, where you can instead call the ...

082

Eduardo Leoni @domleoni.bsky.social · 16/05/2026

simple fix: put butter, high heat pan, and wait for the butter to dry up. but do not put a f*g burger on it! just butter.

000

Eduardo Leoni @domleoni.bsky.social · 16/05/2026

Race condition, Heisenbug, Deadlock, Stack overflow, Off-by-one

000

Eduardo Leoni @domleoni.bsky.social · 14/05/2026

Where do I start? [I write mostly R, some python]. I write tests, but I know I should invest much more on the how, what, strategies, etc.

100

Reposted by Eduardo Leoni

Dr Di Cook @visnut.bsky.social · 11/05/2026

Super excited to get this at my door today! Ursula Laa and my book on exploring high-d data and models is in print! Book website is www.routledge.com/Interactivel... if interested I think my 30% discount code might work for you - msg me. Time to update dicook.github.io/mulgar_book/ too! #rstats

415832

Eduardo Leoni @domleoni.bsky.social · 08/05/2026

this intersection on the venn diagram (skilled and does what you tell) is not very big in general. [A skilled RA that does things differently than you tell her, however, can be a + ]

010

Eduardo Leoni @domleoni.bsky.social · 08/05/2026

Interestingly, the evidence for elected officers (which is strictly your point) is actually better, and the mechanism seems to be selection / competition. [Judiciary in Brazil is notoriously anti-competitive, so that fits!]

000