Notices by Boris Barbour (borisbarbour@mastodon.social)

Ce qui frappe c'est le caractère peu impressionant des harceleurs une fois démasqués et isolés. Même constat pour les émeuteurs lors les "Farage riots" au UK. Rendre la justice à leur niveau c''est certes nécessaire, mais pas du tout efficace - ils sont numbreux, dispensables, rapidement renouvelés, et la correction est bien trop lente et éparse.

https://www.mediapart.fr/journal/france/220125/au-tribunal-la-streameuse-ultia-face-ses-cyberharceleurs-je-suis-epuisee-j-aimerais-que-ca-s-arrete

@pluralistic

^ Recommend the linked blog by @matthew_d_green

@glynmoody

It does depend a bit on what is meant by "AI". There are some successes that could probably be built upon: translation, alphafold, weather forecasting, medical diagnostics.

The (welcome) focus on public ownership could reduce the price and precipitate the bubble bursting?

But it does feel like they bought the snake oil. And it's desperately disappointing not to see the same sense of urgency for the real social and climate emergencies.

@glynmoody @wikipedia

Same reaction. However, I suspect the broader aim is to polarise Wikipedia and create distrust of it amongst those he can influence.

@pluralistic

LOL

@pluralistic

^ Some good news from the EU:

@jwildeboer

Unsurprising. I believe Inserm (FR medical research org) has always lost money on patents, while the CNRS (general science) did make money for a while but because of single patent (Taxol).

The world leader in technology transfer is MIT. They do make money, but still only 1-2% of their total budget. (Numbers from memory.)

Hardly worth the effort.

@glynmoody

I wonder what Putin would want. Iran is a key proxy and currently arms supplier.

@brembs @pluralistic

Not difficult to imagine Bluesky (and Threads) going the way of Twitter - an algorithmic swing to the right - shortly before the next elections in the US. Once people have been locked in, start boiling those frogs again.

@brembs @pluralistic

No billionaires, weak lock-in, no algorithm, no ads. And editing.

First big question. How did nobody in a lab operating in this way for 26 years and producing 800 papers raise an alert? If your data are altered, you notice. Were junior members systematically intimidated? Throughout their subsequent careers?

His lab has a systemic problem and it undermines confidence in all his work: people don't cheat if it's unnecessary. For the innocent in the audience, to prove me wrong, you will need to check the original data. Good luck with that under NIH's risible 4-year retention rule!

By all measures, Masliah is totally central to research on neurodegeneration. A very successful career at UCSD with 800 publications and 18k citations, then occupying the hugely influential position of the head of the Division of Neuroscience of the National Institute on Aging.

A few comments on the Masliah affair revealed by Charles Piller and colleagues in @sciencemagazine It's a genuine scoop.

https://www.science.org/content/article/research-misconduct-finding-neuroscientist-eliezer-masliah-papers-under-suspicion

Yet the investigation reports that a huge fraction of Masliah's papers contain image manipulations. So far 1 in 6 of his papers spanning 27 years have been flagged, and many have not yet been checked.

His work also generated treatment strategies that are the subject of multiple clinical trials, with large sums of money at stake. It is likely that Masliah is by now a wealthy man and, at 65, came close to wrapping up a career without being inconvenienced.

@futurebird @medley56 @wmd @alexwild

Have you installed the PubPeer browser and Zotero extensions?

https://pubpeer.com/static/extensions

@Gargron

Exactly this.

@Gargron I am a native speaker and I just had to look up the second meaning...