svara

24 ms·

361.
by svara 2y ago
*Scientific Reports article
362.
by svara 2y ago
The alternatives he proposes have their problems too. Just plotting points will lead to saturation in high density areas that depends on point size and opacity. Making bin color proportional to point density will require normalization to ma
363.
by svara 2y ago
I'm not necessarily disagreeing, but, given enough capital, a lot of wild sounding things can become real. Hype is a great tool to attract capital. Clearly Musk understands this very well and plays that game expertly. It's really
364.
by svara 2y ago
I sort of see where you're coming from, and it's probably correct that processes that rely on complex spreadsheets would often be better served by well engineered and designed software. But that is often not the realistic alternat
365.
by svara 2y ago
Honestly this is something you say as a person who doesn't really run into the use cases for spreadsheets. I've been writing code since I was a kid, but there are jobs a spreadsheet is just the right tool for. Almost anything that
366.
by svara 2y ago
> Note that french people saying "inshallah" say it as a joke (they use this word for fun effect), To be fair that is how a lot of new vocabulary ends up in the standard lexicon. Total tangent there's this thing in Japanes
367.
by svara 2y ago
There's a long-standing academic tradition of teaching researchers, and frankly that makes a lot of sense and is a really important part of scientific education. Otherwise scientific education becomes like school teaching, and you end
368.
by svara 2y ago
That doesn't sound right at all. Improving generalization in deep learning is a big deal. The phenomenon is academically interesting either way, but e.g. making sota nets more training data economical seems like a practical result that
369.
by svara 2y ago
Grokking is certainly an interesting phenomenon, but have practical applications of it been discovered yet? I remember seeing grokking demonstrated for MNIST (are there any other non synthetic datasets for which it has been shown?), but the
370.
by svara 2y ago
You're thinking simple image augmentations. These nets learn much more complex invariants. Basically to isolate concepts from irrelevant context. The point is you can't remove that image from the training data (not practically) an
371.
by svara 2y ago
It has learnt all sorts of invariances, almost certainly also that. I've gotten some very weird results with 4o on images, it seems entirely possible to me that it would go off the rails if the image wasn't in the training data. F
372.
by svara 2y ago
This is incredible, (3) is literally the better iPad ad. Really goes to show how poorly thought out the original was.
373.
by svara 3y ago
https://eic.ec.europa.eu/eic-2023-work-programme_en
374.
by svara 3y ago
That's a risk, but really in Europe the alternative often isn't to find private investment somewhere else, but to continue working for some big company or university or whatever. The EU runs the programs it does because otherwise
375.
by svara 3y ago
> There's no way a single study can prove causality Why? If you're talking about epidemiological studies looking for associations between variables in big datasets, okay. But that's just one type of study, and it doesn&#x2
376.
by svara 3y ago
That's not the only problem with my logic though ;) It's in fact the smaller one, since the rare blank surname are likely to go extinct given enough time.
377.
by svara 3y ago
Everyone has a surname, and as long as there is more than one surname, each surname has a non zero chance of going extinct. Therefore, in the long run, we'll all share the same surname.
378.
by svara 3y ago
Yes but the question is whether there are repeating non empty patterns. I don't think you've answered this? I don't know but I think that's a fascinating question.
379.
by svara 3y ago
I'm sure this year is going to be the year of the Linux desktop.
380.
by svara 3y ago
Try to do something slightly complicated and have GPT4 explain how to do it efficiently in Excel! That's how I do it and it works great. I've gained a new appreciation for Excel.
381.
by svara 3y ago
I'm worried about that too, but it does seem like one of the things that can be moved to clean energy the easiest. You can route requests to the datacenters where energy is cheap, i.e. where generation from renewables is currently high
382.
by svara 3y ago
It's way more than hundreds of millions, it's now well over a billion and most of it by a large margin goes into failed programs. Though it does seem to be getting better (or getting worse more slowly?) recently. https:/&#x2
383.
by svara 3y ago
Why though? If you squint a little, a filesystem and a database are the same thing and filesystem semantics are useful and well understood. Performance is "just" an engineering issue.
384.
by svara 3y ago
In biology, being amenable to reductionist explanations is an evolvability trait. Repeating building blocks can be re-used, combined and repurposed with small changes. That's the fundamental reason why so much in biology does make some
385.
by svara 3y ago
Regarding your point (2), it's not that simple, since written / literary language uses a larger vocabulary, making homonyms more of a problem. In fact if you search a Japanese dictionary in hiragana with a combination of two reaso
386.
by svara 3y ago
As far as entertainment goes, Reddit and TikTok are relatively efficient. As far as financial instruments go, Bitcoin is fantastically, comically inefficient, and by design.
387.
by svara 3y ago
It's aimed at OpenAI's moat. Making sure they don't accumulate too much of one. No one actually has to use this, it just needs to be clear that LLM as a service won't be super high margin because competition can simply s
388.
by svara 3y ago
Cynicism may sometimes seem smart, in this case it leads you to stupid conclusions. While another war between France and Germany would have been unlikely for the reasons you state, it absolutely wasn't a given that the countries would
389.
by svara 3y ago
That France and Germany are now good neighbors is a miracle. It's possible because wise humans on both sides realized that the law of retaliation would cause a never ending cycle. I worry that this sort of wisdom might be in short supp
390.
by svara 3y ago
Yes, that's what I was referring to. If you have an edge flickering at a high frequency and perform a saccade (fast eye movement) over the edge, your retina will be exposed to the edge at regularly spaced intervals. So the motion will

More ›