svantana

30 ms·

391.
by svantana 8y ago
I never said it wasn't. But I'm not so sure you can buy status without putting in some effort on your own part. Maybe it's a bit like physical fitness -- you can pay for a nice gym, a personal trainer, etc, but you still have
392.
by svantana 8y ago
I've long wondered whether it's possible to buy social status. It's tricky because having wealth gives an automatic status increase in most places. Of course, people need to know about the wealth, hence the prevalence of fanc
393.
by svantana 8y ago
Wouldn't that just put you in the "no fingerprint" category? Presumably fairly few people use such techniques, so couldn't that make you more trackable, not less?
394.
by svantana 8y ago
I find that modern c++ compilers can do a pretty good job of auto-vectorization, given that the code is "vector-friendly" to begin with. But writing code in that style is a bit of black magic with lots of trial and error. I wish c
395.
by svantana 8y ago
Are you implying there's something wrong with that?
396.
by svantana 8y ago
Some quantities are non-negative in their nature, such as energy, intensity, number of occurrences etc. When modeling data that is assumed to be a sum of such non-negative terms (i.e. the terms cannot cancel out), it can make sense to enfor
397.
by svantana 8y ago
Wasn't javascript the main way to get moving elements in webpages? Here they're using css animations, which definitely weren't around in those days. So I guess the style is throwback, not the implementation...
398.
by svantana 8y ago
Easily disproven statement by looking at real stats [1], where most western european countries are ahead of the US. If you're suggesting those immigrants are not "poor", consider the Syrian crisis, Germany and Sweden were by
399.
by svantana 8y ago
> The only drawback is you have to crack someone's skull to do it... Couldn't the recent work in implanting mesh electronics via syringe "solve" this problem?
400.
by svantana 8y ago
8 bytes??? Are you sure you don't mean 8 * sizeof(class)? That's just ridiculous. Regardless it would be nice if one could specify the block size as an argument. I suppose I could write my own allocator, but it's just too muc
401.
by svantana 8y ago
Nice work, though I didn't find any actual surprises in the list. Also, looking at the data source (spotifycharts.com), they offer csv downloads of all the charts, which makes it a weird choice to showcase this apparent html scraper li
402.
by svantana 8y ago
I wouldn't call that comparable -- Adam gets 99.99% of the way, and this person is wondering why it doesn't go all the way. The answer of course is that it wasn't designed to. In this case XGBoost fails to do anything at all,
403.
by svantana 8y ago
The individual modules (the "operators") are about the same size as the PO line products.
404.
by svantana 8y ago
Very nice! Did you get the chord data from hooktheory or somewhere else? It seems to me that scraping is the name of the game in this field, but it would be nice if there was some "standard" dataset to work with.
405.
by svantana 8y ago
> as soon as anyone mentions unions, it's all over Isn't there a technical solution to this? Something like an outwardly anonymous social network with "proof" of employment from LinkedIn, where you can pledge to union
406.
by svantana 8y ago
I thought so too at first, but here's my new thinking: The most efficient way to learn real world tasks (for say a robot) is to learn in a realistic simulated environment, which ideally can be controlled by the agent, e.g. rewinding st
407.
by svantana 8y ago
The momentum is slowing on the basic research side, not as many outrageous improvements to benchmarks coming out these days. At the same time, a lot of non-trivial tasks are close to being automated using ML methods, including high-paying t
408.
by svantana 8y ago
In general the trend is toward more complex architectures, which increasingly blurs the line between modular nets and sparsely connected parts of a single model. IMO it's only meaningful to talk about modularity if the modules can be s
409.
by svantana 8y ago
I think the main performance bottleneck is that I'm adding to a submatrix. Which seems to be a big performance hit in basically all high level languages.
410.
by svantana 8y ago
One major issue with Julia is that it only recently reached 1.0, with a lot of breaking changes, that make a lot of libraries incompatible. Another issue is that it's not always that fast, for a recent project I never managed to exce
411.
by svantana 8y ago
This is looking very nice, can't wait to try it out. But I feel it's a missed opportunity to not provide some simple examples [1], as well as a benchmark to quickly estimate the amount of flops that can be expected. The main reaso
412.
by svantana 8y ago
The nice thing about using likelihood as the training criterion, you don't even need an extra category, you can just combine the outputs with basic decision theory, like so: thres = wrong_cost / (wrong_cost+right_gain) respo
413.
by svantana 8y ago
> Apple ditched the 3.5 mm headphone jack in the iPhone 7 for a Lightning connector If this had only been the case, I would have been happy. Then we would have had two lightning ports on the phone, which would have been nice. No, all the
414.
by svantana 8y ago
I was also at Burning Man 2013 (when this seems to have taken place), and some of my female friends were approached by middle aged men who offered them massages. When trying to decline politely, these men used a sort of "uptight-shamin
415.
by svantana 8y ago
> a functional democratic society will stomp a dictatorial one on pretty much any axis you care to think of That's correlation, not causation. A lot of the "best" democracies were functioning well before democracy arrived.
416.
by svantana 8y ago
The cross-platform audio app library JUCE has a benchmark on mobile audio quality published for a lot of phones. Basically only the Pixel and Nexus phones are even close to Apple devices, and for most android phones, the situation is abys
417.
by svantana 8y ago
That's a common technique in math proofs, to disprove a statement by providing a counterexample. But there probably are improvable statements that are not self-referential, e.g. certain statements about infinite sets. But the grander p
418.
by svantana 8y ago
> If you've got a 5% monthly churn rate then at 1000 customers you'll lose 50 customers/month. At 2000 customers you'll be losing 100 customers/month - and all of a sudden your 100 new customers a month will net
419.
by svantana 8y ago
In reality, all societies have exceptions to free speech, such as: threats, libel, inciting riots, hate speech, asking for bribes, etc. Discussing other potential exceptions is hardly fascism.
420.
by svantana 8y ago
Arguably it's not a pure ML problem where you can just learn what to do from human drivers. You need to know what the alternatives are, which actions are potentially dangerous etc. So you run your "beta" system and record whe

More ›