Comment Re:That's 1 Exabyte (Score 1) 25
CERN's main storage is using EOS:
CERN's main storage is using EOS:
Well, it's a causative effect and not a correlation, so it does not go 1:1 but more like through a chain of events.
Basically the less you care about economic viability and ecological impact, the more you can sell cheap useless junk or advertising for said shit, or use nasty chemicals to produce things even cheaper, or ideally you do a combination of all the above.
And then you can burn all your riches on important business flights and that kind of stuff.
I think this also shows how cheap it is to run things if you do not have investors, shareholders and rent seekers siphon off all the money.
The same project pitched by a private company to some government agency or investor would ask for magnitudes more money.
Could it be done cheaper? Yeah. I believe every slashdotter who says they could run it for 1/10th.
But would it be done cheaper?
As it runs now, some people can take care of all the minute shit I could not be bothered with (e.g. asking people for money) and well, paying the overhead cost is painful, but at least now the person who has to deal with those headaches gets some money in return.
In Germany we have around elections the "wahl-o-mat" which is some website where you can do that with parties.
It prompts you a bunch of questions to different points on the agendas and in the end you can weight some questions double or exclude questions from the evaluation.
Then you can compare how well your answers match with all the parties.
It is obviously not perfect and all the caveats that also apply to rankings apply also to such tools, but it has been done and can be presented in a way even genpop can interact with it.
Crows are amazing in all sorts of ways, like how you can train them to bring you items of interest (e.g. money) in exchange for reward or how some species have been observed tumbling down roofs to induce some kind of altered state. But "statistical logic" is kinda given just by the nature of neuronal networks, be it natural or artificial. It's the basis of its functionality and also the reason for all sorts of biases and malfunctions we develop, such as superstition and just learning correlations instead of causations.
It's nice to see some quantification of the granularity to which crows are able discern different probabilities, though.
On the other hand, the paper nicely demonstrates how long it takes to move away science from the very human-centric worldview of the darkages, where people would not even dare to attribute a modicum of intelligence to non-human lifeforms, even when we already do (or at least should) know such things for decades from basically petri dish experiements.
Also it seems to me they tried to be some kind of apple of ebikes.
With the exception of the pedal I think all parts where vanmoof exclusive, so it was basically vendor lock from day one.
Currently we seem to shift away from the the blue light theory, but in general the answer is "it is complicated", and e.g. the following review article arrives at the non-conclusive conclusion:
https://www.sciencedirect.com/...
Acute alerting effects of light: A systematic literature review
On the research side, there might be some confounding just by availability bias, as most current bright light sources have a high amount of blue light.
But you can also e.g. use red light to increase alertness in shift workers, just as well as blue light:
https://www.sciencedirect.com/...
Red light: A novel, non-pharmacological intervention to promote alertness in shift workers
While there is basically no doubt on **bright** lights, due to the prevalence of blue light filters we see a lot of anecdotal evidence for the blue light theeory, i.e. self reported "but filtering blue lights on my computer/phone in the evening helps with my sleep". But there we also have a causal behavorial/psychological explanations: Bright, vivid, colorful scenes (especially if it is only a very small bright colorful thing like a phone) are more engaging than dimmer, desaturated scenes and by reducing the blue light one also reduces the saliency. Since then it is less arousing and engaging, it is easier to fall asleep afterwards.
God of war basically already went the commercialization route, as the gameplay incentivising fun and good play in the first games moved to a long dragged cut scene with some button pushing inbetween in the latest part, where even simplest jumps are scripted. I don't think I could better describe it than underthemayo on his youtube channel, where he discusses the gameplay differences between the early games and the newest one in great detail.
And in some way it worked, as clearly, the game as piece of art is stunning and now it might even result in a movie. So, it already hasn't been left alone and this is only a logical next step.
The complex numbers C are not just a fancy way to express 2dimensional real space. It is more like a coincidence that we can visualize complex numbers this way, but actually they are still one dimensional - just in complex, not real dimensions and in fact have some surprising (at least at the time of their discoveries) consequences for the real numbers R.
Algebraically it makes more sense to think of C as an extension of R, and not as a R^2, and one thinks of it more as C = R[i], so like, the extension of the real numbers by one additional algebraic thing i.
The main feature of course is that you can solve x^2 - 1 = 0 over C, but not in R, which is equivalent to saying the polynomial decomposes into linear factors, i.e. x^2 - 1 = (x - i)(x + i) = 0, so you can read off the solutions from the linear decomposition, and this is then the reason why one calls C also the algebraic closure of R.
While this first just seems to some algebraic nitpicking, this actually has
some real consequences.
For example in the case of power series, even though all
our formulae only involve real numbers, the function sometimes behaves
weirdly at points one would not expect.
Everyone favorites example for this would be the taylor series of
f(x) = 1/(1 + x^2), which obviously has no singularities in R and is a
very nice function, but the taylor series
sum n= 0 to infinity (-1)^n x^(2*n) only converges for |x|
But the answer is that in C there are singularities at +-i, and in
fact, any radius of convergence can actually be expressed as the radius of the
disk on which the function would be holomorphic, a complex property, if we
were to allow complex arguments.
This also also shows why holomorphic functions, so complex differentiable and
not only real differentiable, are much more restricted than real
differentiable functions and in some sense boring, but I guess I digress and
on the latter there are probably some people who might disagree.
Another fun thing is that in complex vector spaces, all matrices are
diagonizable, because eigenvalues are the roots of some polynomial and well,
we can find all those roots, but we can not do the same thing in real vector
spaces.
However, when we think about how complex multiplication works, multiplying by
i is the same as rotating by pi/2 on the plane represantation, so what we
actually figure out by this is that whenever we have complex eigenvalues,
there is some rotation going on if we think of it in the real world.
Starting from this thought then one can figure out how there are maybe some
ways to represent these complex numbers as real matrices (e.g. i as a rotation
by pi/2 can be very simply epxressed as a 2x2 matrix) and this leads you down
the path of things like Jordan Normal Forms.
I did check my numbers and I was indeed wrong. The half and half only holds for the elderly, which is what I had in mind, but those are only about 2/3 of those with covid symptoms in intensive care, which I did overestimate
To be more accurate, it is 40% vs 60% but there is some shadow number of those with partial vaccination, but which is roughly as effictive as a a JNJ "full" vaccination. The ones that received a JNJ vaccine are "vaccinated", the ones that did only receive one shot of the other vaccines do not count as "vaccinated" for these numbers. This trended up from somewhere in the 10/90 range a few months back.
Including the whole population it is a ratio of 1 : 2.
See https://www.rki.de/DE/Content/...
page 24 in the pdf
So, to be clear, I do not want to paint the picture "the vaccination does not work", but "even if we had a fully vaccinated population, we would need to stop a lot more infection".
As a German, I can attest that pretty much since about April/May 2020, the regularotry bodies and politicians have acted completely random and bogus and in pretty much the worst way, almost contrary against everything we already knew at that point, with all major decisions being too late.
Like a deer caught in the headlight, you can perfectly model the actions by "indecisiveness until it is too late", combined with a whole bunch of promises and strategical mistakes which were obviously bad long-term and not really beneficial short term.
The latest push for vaccination is a pretty lame attempt for victim blaming, because most of the non-vaccinated folks are people who either suffered more than most other people under the measurements of the lockdown, because they were forgotten and no one did care for them, making them easy victims for radicalization/conspiracy theories, or were just "forgotten" in the information campaigns, as they lack higher education and almost no attempt was made to provide easily accessible information about different vaccines and other safety measures, especially not in non-German language.
A lot of public available information was wrong or outdated, and the one very good public source (NRW corona podcast) was an excellent edutainment program for your average slashdot reader interested in some protein copying and RNA mechanics, with some statistical fun interspered, but maybe not the news source for at least my grandma or other people who did need convincing of the vaccine.
Or, sometimes they just stopped the vaccination before everyone got a chance, and when it became clear that the JNJ vaccine is after one shot less useful than the other vaccines after two shots, it still took months until the corresponding governing body gave its OK to vaccinate JNJ receivers. One might almost think because then more people show up as "fully vaccinated" in the statistics, but of course no one in government would ever think that. Anyway, this reluctance now lead to the situation that in one state about half the police force is sick with covid and unable to work.
Of course there a still a few nuts you really can not convince, but most people have consistent reasons to not be vaccinated.
To make things worse, we reduced the number of available beds massively (e.g. in my town we went down from 300 available intensive care beds to 200), while giving pretty much no support to healthcare workers, so we are now in a state where our healthcare system is weakened, the healthcare workers that have not quit yet are on the brink of collapse.
And while about half the people with covid in intensive care are unvaccinated, we tend to ignore the other half, which is actually vaccinated, becaue we opened up things too much and are still hesistating to impose even minor restrictions. Because, you know, tHE UnVAcCiNaTeD.
Then there is this issue with schools, where over the summer of 2020, nothing happened to make schools a safer place and reduce infections there, we did totally learn from our mistake and again let the summer of 2021 pass without making any progress on this issue, but also making sure masks were not used in as many classes as possible.
Pretty much every decision we made since the quick and strict lockdown in 2020 was wrong, and not only wrong, but maximally wrong and diametrical to our proclaimed goals, considering what we knew at the point and what we did up to that point.
There were many warnings by all the people with at least a modicum of brains, in public, in private and in pretty much all places, with enough time to act accordingly, but none of those warnings was recognized.
And then we can sprinkle on top some nice corruption scandals and
tl;dr: It's the fault of the unvaccinated. Our government is perfect and did everything right and they are stupid and should have no right to anything.
I want to defend the accusative tone a bit, because actually we did know the importance of airborne transmission (or whatever you want to call it when tiny droplets are circulating through the air, actual airborne transmissions are even worse) since about april, through some serious case studies, e.g. some starbucks in an asian country where there was a transmission via the air ventilation system or several cases where cafetarias spreaded the virus across vast distances. At latest in May the situation was pretty clear that this is the main way of transmission.
And yet it was almost impossibly hard to convince people of this new evidence. People put up tiny plexi glas shields and sometimes local laws even allowed you to take off your mask behind these shields.
Also face shields were a thing for too long of a time.
Of course, I don't offer a solution. But this was known for long but people chose to ignore new evidence or just did not understand that our first guess was clearly not as good or even people started doubting science because we changed our mind *once* on the main source of transmission.
Higher level languages are actually more efficient in the grand scheme of things, because ordinary mortals can write working code with it that does not suddenly consume all the resources of the world.
You smart people all have created these wonderful algorithms accomplishing things in a ridiculously efficient way (anyone ever tried to beat this c++ boost library on any task?).
Also we can not require only really smart people to write code or have anyone do a five year cs degree (or whatever you deem a good standardizable base) with lot's of coding practice, since, on one hand, this is obviously not something that is going to happen, and on the other hand, such a commitment of time is better spend on the actual problems people are trying to solve there. And this is really crucial, because the worst code is the code that should never have been written in the first case, because it does not solve a thing.
So, moving away from the status quo and away from "less efficient higher level languages", you are either ending up with many people writing bad c/c++/rust code that is significantly slower than boost (to use boost you already need some amount of proficiency) or numpy and has far more problems than say the python requests library, or, of course, you end up with far less people actually writing code. Which would be really bad, because it would throw us back in the stoneages of the early 2010s regarding data evaluation in science.
On the other hand, the constant suffering that is brought to uso in form of desktop applications and even more central system components by people that obviously have no business doing anything on that level, certainly does give some appeal to the idea that fewer people should code, but I suspect this would boil down to remove a few very specific individuals and, as mentioned before, often is code that should never have been written, but people where too enthusiastic about coding and not too concerned with solving actual problems with their code.
Though this is one of these jokes that might not age well...
The test of intelligent tinkering is to save all the parts. -- Aldo Leopold