flock
2K posts
To be honest, it was quite clear at the time that the field was ripe for disruption and that there was enormous untapped potential. People's jaws would drop over, like, topic models. It was so early. I just didn't expect deep learning to be the thing that did the disrupting.
Andrew Gordon Wilson@andrewgwils
Sometimes I miss the days when people were passionately fighting about MCMC versus variational methods, or whether posterior tempering is problematic. We should have a nostalgia ICML 2010. You can submit AI slop, but expect a 2010 era reaction. What happened to our field?
English
@ashleevance @corememory What a bunch of silliness. Not one thing of substance in the whole clip
English
Alex Wang on the problems he found inside of Meta. #1, he says, was that the company lacked the AI religion of other frontier labs.
English
@DSBIC @alfairhall @SussilloDavid You misunderstand. Just because there is a journal with sys neuro in the title doesn’t make it representative of the field. She attempted to help you realize that. Cosyne, as she indicated, is the main conference that represents systems neuro. Look there: fMRI is all but verboten
English
@alfairhall @SussilloDavid My criticism isn’t of you specifically. It’s of your field. I would venture a guess that your field published equal numbers of fMRI and LFP papers over the past 5 years.
English
In systems neuroscience, this is simply not the case. Paradigms are challenged and updated all the time.
Yin lab@HenryYin19
Some branches of neuroscience (and science in general) are unlikely to see real progress for a long time. This stagnation stems from an academic echo chamber that very effectively silences dissent. young researchers introducing new ideas are often marginalized or driven out before they can establish themselves. There must be quite a few of these people who left academia this way and are now working in industry. The current institutional framework simply won't fund or validate anyone bold enough to challenge entrenched ideas. It's a conformist paradise gated by tenure. BTW these stagnant areas tend to have dominant senior people who have never been wrong and shall remain "authorities" forever.
English
flock retweetledi
This is your annual reminder that we don’t need to speculate about whether we will have a “theory of deep learning” and what form it might take, because we already have a basic understanding of generalization in deep learning: arxiv.org/abs/2503.02113
English
@DrMikeBrooks @anilkseth @michaelpollan @thenerve_news @GaryMarcus @carolecadwalla Agreed, but that makes the fight all the more important. Given that, historically, nearly every society believed in some sort of animism whereas now few in the West and other places do, that demonstrates the ability to overcome. Many other such examples
English
@DrMikeBrooks @anilkseth @michaelpollan @thenerve_news @GaryMarcus @carolecadwalla People believe in all sorts of ascientific things, like religion. That doesn’t mean science shouldn’t attempt to provide a more rigorous foundation and answer—in fact, it makes the matter all the more urgent
English
My take on the Dawkins/Claude/Claudia/Claudius kerfuffle in @thenerve_news - with many 🙏🏽to @GaryMarcus for reminding me how Dawkins wrote so powerfully about the weakness of arguments from personal incredulity. .@carolecadwalla thenerve.news/p/opinion-rich…
English
@SamuelFitouss10 @anilkseth @thenerve_news @GaryMarcus @carolecadwalla > His central point was this:
Claude demonstrates that it is possible to be intelligent without being conscious.
——
What Dawkins actually said:
> So my own position is: “If these machines are not conscious, what more could it possibly take to convince you that they are?”
English
I think you misunderstood Dawkins’ article. His central point was this:
Claude demonstrates that it is possible to be intelligent without being conscious.
Hence, he asks, why did natural selection never create intelligent yet non-conscious organisms? After all, such organisms could survive and reproduce. What evolutionary “function” does consciousness serve? Can we imagine, on another planet, intelligent beings devoid of subjective experience?
These are interesting questions.
Replying by saying that LLMs are not conscious misses the point entirely; Dawkins’ questions arise precisely from the fact that LLMs are (probably) not conscious.
English
@HydrogenGuide Look at me, I’m a total bullshit scientist who has no fucking idea what he’s saying!!!!
English
What really makes us gain or lose weight?
Dr. Steven Gundry, a gut health expert, has a surprising answer…
It all comes down to your gut type!
Take our FREE 30-second quiz to learn about yours!
Try Ultimate H2
English
@cremieuxrecueil You have an awful way of describing data. I bet it’s people like you I tear into as a reviewer. 1) Your data are incredibly old & narrow, all but worthless to draw such a broad conclusion. 2) Why “teachers” writ large? Secondary education nearly identical to computer programmers
English
Teachers average among the least intelligent university graduates. Little of what they 'learn' is even relevant to what they end up doing.
In a given year, the lowest-scoring groups on the GRE, SAT, and ACT are usually those pursuing degrees in education.
giardiniera gestator@just_riffing
Actually neither of us are qualified to homeschool children because we both have not obtained degrees in childhood education hope this helps, Allie.
English
@cremieuxrecueil Are there no error bars or are they just very small bc of large N?
English
@emollick but also
x.com/karpathy/statu…
Andrej Karpathy@karpathy
Don't think of LLMs as entities but as simulators. For example, when exploring a topic, don't ask: "What do you think about xyz"? There is no "you". Next time try: "What would be a good group of people to explore xyz? What would they say?" The LLM can channel/simulate many perspectives but it hasn't "thought about" xyz for a while and over time and formed its own opinions in the way we're used to. If you force it via the use of "you", it will give you something by adopting a personality embedding vector implied by the statistics of its finetuning data and then simulate that. It's fine to do, but there is a lot less mystique to it than I find people naively attribute to "asking an AI".
English
We tested one of the most common prompting techniques: giving the AI a persona to make it more accurate
We found that telling the AI "you are a great physicist" doesn't make it significantly more accurate at answering physics questions, nor does "you are a lawyer" make it worse.
English
flock retweetledi
This could have covered the entire budget of the National Science Foundation for 10 years.
Instead, Trump wants to reduce the NSF budget by 50% ($5B a year instead of $9B), which would decimate the American scientific research ecosystem, dramatically reduce the number of PhD graduates, and destroy the technology innovation flywheel.
English
@prathoshap @ben_golub @pangramlabs Thanks for the response. To people like us here in the US it sounds like garbage LinkedIn cringe cranked up to 11. You can ask your chatbot why that’s the case.
English
@flock49600209 @ben_golub @pangramlabs About 60 % (of the language). It neither sounded cringe nor painful to me. I don't honestly understand what's this ado about nothing. The intent and message is what that matters.
English
They say academia is a "thankless" job, but then you get an email like this on a Sunday evening.
A 3rd-year student from a college in Kolkata wrote to me. He felt "disconnected" and "trapped" in a coding-centric major, missing the mathematical rigor he loved until he found my NPTEL course.
He didn’t just want to learn ML; he wanted to feel the philosophy and the probability behind it. He wanted to understand why, not just how.
Moments like this make the "Chalk and Talk" sessions and the late-night prep worth every second. In a world obsessed with "shortcuts" and "quick certificates," finding a student hungry for first principles is the ultimate reward.
It reminds me that our real job isn't just to transfer information, it’s to provide the ignition for a mind that was almost ready to give up.
This is why we teach. This is what makes this profession truly invaluable.
English
@prathoshap @ben_golub @pangramlabs Yes, do give the actual number please. But it certainly reads like 100% slop. Truly cringe, painful to read.
English
@ben_golub @pangramlabs Why should that be the case? The intent is 100% mine, but the actual text is edited with AI assistance. Why should that be a taboo? By the way, the tool is wrong, it's not 100% AI generated. I can give you the actual number if you care ;)
English
flock retweetledi
flock retweetledi
Shooting oneself in the foot?
Nope.
Shooting oneself in the prefrontal cortex.
News from Science@NewsfromScience
U.S. President Donald Trump has fired all 24 members of the National Science Board, the body that oversees the National Science Foundation. Many science advocates see it as the latest step by his administration to erode—some would say destroy—the independence of the 76-year-old research agency. scim.ag/4eGM0YS
English
flock retweetledi
A single GPU can now calculate hundreds of global weather scenarios in under 60 seconds. The exact same task requires a supercomputer and hours of brute-force physics.
Google DeepMind recently released WeatherNext 2. The model beats the previous state-of-the-art system on 99.9% of weather variables across a 15-day forecast window. It achieves this massive jump in accuracy using a new modelling approach called a Functional Generative Network.
Meteorologists categorise weather data into two buckets:
1. Marginals are isolated data points, like the precise temperature at a specific location or the wind speed at a certain altitude.
2. Joints are the massive, interconnected systems that form when all those individual elements interact.
The researchers hid the joint systems from the model during training. They only taught it the isolated marginals. When they turned it on, the model skillfully predicted the massive, complex systems anyway.
The architecture forces an 87-million-dimensional output distribution through a 32-dimensional mathematical bottleneck. To survive this severe constraint and still produce accurate individual data points, the neural network has no choice but to learn the underlying physics linking everything together. It figures out the weather because that’s the most efficient way to solve the maths.
The practical results are immediate. The model gives forecasters a full 24-hour advantage in tropical cyclone tracking compared to the previous leading system. It maps extreme wind speeds and heatwaves with unprecedented precision.
We’re watching a pretty big shift in predictive capabilities. The machine is deducing the structural reality of planetary weather from isolated fragments of data.
English
Iran says Hormuz is open as long as the ceasefire with Lebanon holds.
I think they may have stumbled upon something. Use Hormuz as a tool to put pressure on Israel specifically. Concede to America, and try to create division between Israel and the US.
English
Dugin tells Sneako and "Professor Jiang" that Trump may have been opposed by the "Deep State," but he was chosen by the "Deeper State" composed of "hyperglobalists."
Their stupid conspiracy theories fail to explain the world, so they concoct even more complicated conspiracies.
English
@AlexTran677026 Wait what?? 500 yards is just a shade iver a quarter mile, not “just shy of half a mile”
English
One of the scenes that impressed me the most in cinema is this scene.
The result of 20 years of perseverance of a prisoner. It symbolizes hope and patience.
The Shawshank Redemption
Love Classical Music and Movies 🎺🎻💖🎥🎬@AlexTran677026
What is the most impressive moment in your movie?
English