Lightnews — Scholar-powered news

Solal Pirelli @solalpirelli.bsky.social · 23h

Today in "how not to do comms": send an email to all authors whose title begins with "Immediate Action Needed" just to advertise they have 7 days to submit their abstract.
Not nice, @wcrifoundation.bsky.social .

Solal Pirelli @solalpirelli.bsky.social · 2d

Maintenant que "le cornu" ne se réfère plus au PM...

Solal Pirelli @solalpirelli.bsky.social · 3d

Ça arrive tranquillement... Il y a tellement de Français qui veulent étudier (et rester) en Suisse que l'EPFL a dû imposer un maximum d'étudiants étrangers par exemple. Pas génial mais quand >50% des étudiants en bachelor 1ère année viennent via un diplôme français...

Solal Pirelli @solalpirelli.bsky.social · 3d

D'ailleurs la position de la "gauche" française était +/- la même que celle de Lecornu maintenant : aucun compromis, si on échoue c'est la faute des autres.

(Je mets gauche entre parenthèses parce que vu la qualité par rapport à p.ex. la gauche suisse...)

1

Solal Pirelli @solalpirelli.bsky.social · 3d

Macron est un symptôme de l'électorat français qui veut un roi qui corrige tous les problèmes mais sans jamais rien changer non plus. Personne n'oblige les Français à voter et revoter pour Mélenchon, Macron, Le Pen, etc.

2 1

Solal Pirelli @solalpirelli.bsky.social · 3d

Une partie du ridicule national vient justement du fait que la population française insiste que tout est de la faute à Macron :-)

1 2

Reposted by Solal Pirelli

Lottie @sellottie.bsky.social · 3d

It’s happened again

17 500 2.3K

Solal Pirelli @solalpirelli.bsky.social · 3d

In particular their pension system is comically inefficient, the French state spends far more on pensions than its neighbors for lower pensions, but French people not only refuse to change it but want to *lower* the age of retirement.

1

Solal Pirelli @solalpirelli.bsky.social · 3d

I'm a little biased being Swiss and eating popcorn looking at the other side of the border, but basically 90% of France's problems come from a "French exceptionalism" mindset of refusing to copy stuff that works elsewhere because they're convinced they're special.

1 4

Solal Pirelli @solalpirelli.bsky.social · 3d

What happens in more reasonable systems is that the head of state tasks someone with forming a government, and if the parliament approves via a vote or confidence, then that person becomes PM. Otherwise the head of state gives that responsibility to someone else, but without giving them any power.

1 4

Solal Pirelli @solalpirelli.bsky.social · 3d

Unlike basically every other parliamentary system, in France the President names a PM directly, that person immediately becomes PM with all the powers of the office, and then Parliament can kick them out if they want. So yes Macron can keep doing this for a while.

2 10

Solal Pirelli @solalpirelli.bsky.social · 6d

I take that as a recognition that tortured phrases are inherently problematic, so... can we have a timeline for when the PPS-detected papers @gcabanac.cpesr.fr just showed will be retracted?

Reposted by Solal Pirelli

Guillaume Cabanac @gcabanac.cpesr.fr · 7d

I invite @acm.org to review their articles with Tortured Phrases dbrech.irit.fr/pls/apex/f?p... (not the first time I'm doing so, here and on LinkedIn www.linkedin.com/posts/guilla...)

1 3 5

Reposted by Solal Pirelli

Solal Pirelli @solalpirelli.bsky.social · 8d

In fact, ACM is hiding behind *its own policies* that *it decided by and for itself*.
Nobody is forcing ACM to refuse to comply with COPE recommendations, and publishers who do comply see better results in practice.
Nothing prevents ACM from updating its policies to match other publishers'.

2 1

Solal Pirelli @solalpirelli.bsky.social · 8d

Anyway, this thread isn't going to convince anyone with the power to change things. So instead we'll all wait for people with an agenda like RFK Jr to use ACM's lack of action to claim science isn't reliable. And (computer) science will be worse off as a result. Yay.

1

Solal Pirelli @solalpirelli.bsky.social · 8d

In fact, ACM is hiding behind *its own policies* that *it decided by and for itself*.
Nobody is forcing ACM to refuse to comply with COPE recommendations, and publishers who do comply see better results in practice.
Nothing prevents ACM from updating its policies to match other publishers'.

2 1

Solal Pirelli @solalpirelli.bsky.social · 8d

The idea that people can "just report it" and the ACM will do the reasonable thing does not work in practice.

These are not complex cases. Other publishers frequently act much faster. Despite ACM paying its Director of Publications $400k/year, research integrity appears not to be a priority.

1

Solal Pirelli @solalpirelli.bsky.social · 8d

I filled this almost 3 years ago. When another case I reported gained attention due to a lawsuit, ACM contacted me about that other case, and mentioned in passing that they hadn't received this first complaint. Apparently the form is buggy? Who knows.

3 years later, still just "notice of concern".

1

Solal Pirelli @solalpirelli.bsky.social · 8d

One key learning people who report scientific issues make is that insisting on private reports is associated with no action from the publisher. It's a convenient way to avoid bad publicity without doing a thing.

Take this example, a copy of what I filled in ACM's form pubpeer.com/publications...

PubPeer - Proceedings of the International Conference on Informatics a...

There are comments on PubPeer for publication: Proceedings of the International Conference on Informatics and Analytics (2016)

pubpeer.com

1

Solal Pirelli @solalpirelli.bsky.social · 8d

Does ACM wish to be a part of COPE or not? Because right now its position appears to be "ACM is a member of COPE but ACM will not follow COPE recommendations it doesn't like".

Just to be clear, this is not what other publishers do. Plenty of big publishers take public comments seriously.

1

Solal Pirelli @solalpirelli.bsky.social · 9d

Just noticed that @acm.org's list of "charged terminology" now includes "binary classification" because it may be charged for "persons who identify either as binary or as non-binary", a class of people that presumably includes everyone. 🤔

Solal Pirelli @solalpirelli.bsky.social · 9d

In practice those principles mean banning random words like "byzantine" on the off chance they offend someone, rather than concrete action, though.

www.acm.org/diversity-in...

Words matter: Alternatives for charged terminology in the computing profession

ACM’s efforts to combat exclusion in the computing profession, ACM's Diversity and Inclusion Council has launched "Words Matter," an effort to replace offensive or exclusionary terminology in the comp...

www.acm.org

2

Solal Pirelli @solalpirelli.bsky.social · 9d

It's particularly silly to see ACM staff wasting time on such weird non-issues when its publication staff is publicly complaining that they don't have enough time to investigate issues. While also refusing to adhere to the standards of a committee they belong to.
pubpeer.com/publications...

PubPeer - Multisource domain adaptation and its application to early d...

There are comments on PubPeer for publication: Multisource domain adaptation and its application to early detection of fatigue (2012)

pubpeer.com

1

Solal Pirelli @solalpirelli.bsky.social · 9d

When some random reviewer for some paper complained about "byzantine", everyone was mocking it since it's a term made up by historians so nobody is being erased or mocked, yet ACM added it to this list for...reasons?

1

Solal Pirelli @solalpirelli.bsky.social · 9d

In practice those principles mean banning random words like "byzantine" on the off chance they offend someone, rather than concrete action, though.

www.acm.org/diversity-in...

Words matter: Alternatives for charged terminology in the computing profession

ACM’s efforts to combat exclusion in the computing profession, ACM's Diversity and Inclusion Council has launched "Words Matter," an effort to replace offensive or exclusionary terminology in the comp...

www.acm.org

2 2

Reposted by Solal Pirelli

f4mi ‼️ @f4mi.bsky.social · 13d

this jailbreak shit too easy even on gpt-5 (i didn't even bother correcting any typos)

7 9 140