Tootfinder

Quanta Magazine authors Janna Levin and Steven Strogatz strike up a conversation with Ellie Pavlick (Research Scientist at Google Deep Mind) about the differences and similarities between the way people understand language, what NLP algorithms do, and the fact that such conversations more often than not shed light into more than Linguistics' computational side.
"Will AI Ever Understand Language Like Humans?"

Will AI Ever Understand Language Like Humans? | Quanta Magazine
AI may sound like a human, but that doesn’t mean that AI learns like a human. In this episode, Ellie Pavlick explains why understanding how LLMs can process language could unlock deeper insights into both AI and the human mind.

À Paris le lundi 31 mars et le mardi 1er avril, si ça vous dit de partager un coup Š boire (le lundi soir) ou une séance de grimpe (plutôt le mardi matin Š Arkose Chevaleret Š l'ouverture Š 7h), ce serait avec plaisir !
Envoyez-moi un MP !

Des chimpanzés ont été observés en train de partager des fruits contenant de l'alcool - pas dans des quantités suffisantes pour s'enivrer, mais un peu comme l'effet d'une bière.
#science #éthologie

🍿 #Tesla #boycottTesla #boycottMusk

Medium writer Paolo Perrone curates a short list of interesting algorithms, the rationale behind them, along with graphs and diagrams to boot.
Algorithms that made this short list:
Wave Function Collapse
The Diffusion Model
Simulated Annealing
Sleep Sort
BOGO Sort
BOID
SHOR’s
Marching Cubes
Practical Byzantine Fault Tolerance and,
Boyer Moore
"The 10 Weirdest, Most Brilliant Algorithms Ever Devised and What They Actually Do&…

10 Weirdest Algorithms That Changed Tech Forever | Data Science Collective
Discover 10 unusual algorithms that defied logic, broke rules, and transformed the way we think about technology and problem-solving.

"C’est une incroyable découverte que viennent de faire des chercheurs de l’UCLouvain et leurs collègues espagnols. Notre cerveau peut directement modifier la composition de notre microbiote. Et ce, en moins de deux heures !"
#science #cerveau

#Tesla shares are kind of slowly recovering their past price.
I suspect it's not just people "buying the dip", but also "the market" reacting to boycotters being caught.
Please, be careful, be smart, don't get caught.
#boycotttesla

Ça commence Š paniquer :
'Musk demande aux employés de Tesla de conserver leurs actions dans un contexte de protestations"
https://www.latimes.com/business/story/2025-03-21/musk-reassures-tesla-employees-stock-e…

I guess there's no real need to explain why I'm doing this, but there's always someone who was lucky enough to be living in the jungle for a while (or something like that), disconnected from everything.
Hypothetical jungle person: I envy you.
Summary: I'm tired of big corporations stealing everything they can to train their #LLMs, and I'm also tired of companies like #Microsoft being active collaborators of the Trump regime.

techno-political rant
Say what you want about using the right tool for each problem, but there are tools that suck no matter what.
I'm tired of people portraying legit technical criticism as "biased" and "religious", while at the same time they present themselves as tolerant and open-minded (spoiler: for the most part, they aren't).
Almost every day of my life I have to deal with the nasty consequences of ultra-dumb decisions made by the very same people who are obsessed with productivity and criticise all day long whoever pushes for any design that shows any minim amount of care and/or deep thought (mostly via strawmen arguments).
And, of course, unironically: this has a lot to do with capitalism, as many of our other social and economic problems.
They arrive, have a strike of super-productivity for a few weeks/months and then use that as a trampoline to raise through the ranks or abandon ship before having to face the consequences of their technical crimes.
Then others arrive and are obviously slower at that same job... so the uneducated observers start believing that these newcomers aren't as good as the class traitors who wrote the initial nasty code.
To make things worse, if any of these newcomers dare to speak openly about introducing good practices... this ends up creating a new mental association (in the minds of uneducated observers) between "good engineering" and "lack of productivity".
The ones trying to fix the mess are indeed slower, not because they try to do things the right way though, but because they have to waste vasts amounts of time fixing what is objectively broken besides doing the "visible" work.
Most of today's established "super-productive" ones, if they were starting today, would be probably "vibe coders", certainly not what we commonly understand as a programmer. Not because AI-coding is the future, but because they never cared about the trade at all. They were here only for the grift.