HN user
Assuming this is actually a genuine question:
Here are the per country ratios of different VOC/VOI: https://covariants.org/per-country
It took a while until it became dominant in India, and then a similar pattern repeats in each new country.
According to this thread from another virus expert [1], you can integrate pretty much every class of viruses in the system the authors of this paper used.
[1] https://twitter.com/ArisKatzourakis/status/13904196191696732...
I implore everyone to read the full article and not only the headline. The paper that is being reported on was widely criticized in its scientific community for its weak evidence and should probably be seen (if you are very charitable) only as a starting point for other virus experts to jump into the discussion and do more experiments and not as something for general consumption.
That's a valid, but still weird take to me. Academic publishing takes a very long time. Every paper I was involved in took many months from submission to eventual publishing (completely ignoring the time it takes to prepare a submission).
Ioannidis published something extremely controversial (if not even flawed) and one of the main authors he attacked responded with a lengthy explanation so that this manuscript would not remain unchallenged. I found that aspect way more important than the venue of response. Would you prefer to leave Ioannidis' work unchallenged for potentially months instead?
I would advice to interpret the IFR reported by the Ioannidis paper with an extreme amount of caution. One of the authors criticized (which is quite an understatement) by Ioannidis went into a detailed rebuttal in [1]. A second thread [2] also gives a very detailed analysis of issues with the paper.
[1] https://twitter.com/GidMK/status/1376304539897237508 [2] https://twitter.com/AtomsksSanakan/status/137593538213983437...
If you can recompile but the code was written to only target AVX-512 you can use https://github.com/simd-everywhere/simde to near effortlessly map the intrinsics to AVX2 (or lower).
This still means that tools published in the last few years until now might just be gone soon. The people who uploaded the images might have graduated or moved on and none will be there to save the work.
This will be quite bad for reproducible science. Publishing bioinformatics tools as containers was becoming quite popular. Many of these tools have a tiny niche audience and when a scientist wants to try to reproduce some results from a paper published years ago with a specific version of a tool they might be out of luck.
My grandmother has nine children. Seven of those survived and two did not. This was very common in the area she lived then. To this day, however only in private moments, she talks about the two that did not make it. The loss of the two children many decades ago still brings her a lot of pain. She gave the name of one of the children that did not survive to a later child, however that never erased the suffering.
I think we should be more kind to our ancestors. Just because they lived in fucked up times, compared to our current standards, does not mean they experienced a different quality of suffering.
The message from the community was to ensure mask supply for medical staff first. They are both more at risk themselves and a risk for others due to the large numbers of contacts.
Once supply is secured everyone should be educated about correct use and be encouraged to wear them.
Most of Asia already has a culture of wearing masks and existing supply chains. Europe/US did not and it takes a while to establish.
It might not be surprising, but it is an extremely important point to keep reiterating.
People keep spreading the malthusian myth of overpopulation and shifting blame to the poorest. Blaming the poor while giving the rich a free pass is not only unfair, it’s cruel.
Once climate change denialism is not tenable anymore, climate change will be instrumentalized for atrocities.
I was looking for adding at least one Threadripper server to our local HPC system. I only found one ASRock motherboard for TR4, but none yet for sTRX4. I hope we will see 1U rackmounted sTRX4 systems soon.
Our new HPC system will mostly consist of Epyc 7742, but having one node with super high single-thread performance would be nice for less well parallelized applications.
Microsoft posted this blog post titled "Ambition is good; action is better: Making progress on our climate commitments" two days ago:
https://blogs.microsoft.com/on-the-issues/2019/09/22/ambitio...
The value of their actions seems pretty clear cut.
I did the calculation myself due to multiple conflicting numbers in this thread: I get 2.6 dollar per gallon of gas (8.65kg CO2 produced by complete combustion of 1 gallon gas):
300$/ton * 8.65kg = 2.6$ per gallon (0.6$ per liter)
100$/ton * 8.65kg = 0.87$ (0.23$ per liter)
30$/ton * 8.65kg = 0.26$ (0.07$ per liter)
7 cent per liter seems quite acceptable. And people are reporting that some companies are targeting that price point for sequestration. Please correct my if I miscalculated something.
Publishing companies tend to create many journals for different (sub-)communities and also numerous other reasons. Nature Research (the company) has next to Nature (its most well known journal) many smaller ones, like Nature Biotechnology, Nature Communications or in this case Scientific Reports. Some are well regarded (e.g. Nature Biotechnology, Nature Communications). Scientific Reports has a comparatively low impact factor, which has furthermore dropped every year in the last few years. Impact factor measures the average number of citations a published paper receives. Its impact factor of 4 is not very impressive. The main Nature journal has an IF of 41 (Nat Biotechnol has 35 and Nat Comm has 11).
In this article the patent applications are disclosed in the Competing Interests section. The readers can draw their own conclusions from those. It would be very unusual for a journal to not demand a declaration of conflicting interests.
That has already happened. The NCBI maintains a taxonomy tree: https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?id=...
In this case human the identifier of human is 9606, all kinds of stuff about the human can be queried through this ID. All fungi are classified somewhere under the node 4751. (Even though this website looks awful, its one of the better and most useful resources of the NCBI.)
However this whole taxonomy business is a pretty difficult issue. According to the article, fungal research is (was?) wasting time on morphology. However, even though taxonomic classification through phylogeny is probably a massive improvement, it doesn't solve all issues. In this framework a species is usually defined by some sequence identity cutoff of ultra conserved proteins in the genome. This works probably pretty well for for most eukaryots, however becomes somewhat of an odd measure for bacteria, where two of the same bacteria can have a very low sequence identity of highly conserved proteins. Then you also have another issue, that once you want to classify a new organism, you might only find contaminated, miss classified or in any other way bad entries in the reference databases. Morphology as a sanity check probably makes sense.
Its a big field with a lot of different trade-offs trying to be united under one umbrella. There are no silver bullets.
Root-less installation of packages into user-directories.
I use it a lot in our scientific cluster environment. My User home directory contains a linuxbrew installation and is shared over NFS. Every cluster node can access all linuxbrew software.
It beats central coordination with the administrators, or writing module files for the cluster environment by a lot.
A cp -r $HOME/.linuxbrew $HOME/.linuxbrew.bak && brew potentially-destructive-operation is also super useful.
Packages are also very close to the latest version.
You get all the work people did got MacOS on Linux.
For package developers, their recipe system is really good and easy.
Of course, a transactional package Manager like nix or guix that can also run inside a user-directory would also give most of these advantages, but I am already used to brew from my MacBook, no additional learning effort required.
Depends heavily on the projected utilization. If you know your compute node is going to be computing for the next 3 years with at least medium utilization, then the self hosted metal is probably going to be quite a bit cheaper.
Its amazing how much hardware you can pack into a single machine for 10k€. Last year our group bought two additional high-memory (768GB) nodes for around that price each (including support for a couple years from the vendor).
A few years before we bought 40 nodes with 128GB RAM each, for a similar price to last years high-memory nodes (and a fast interconnect and a lot of storage).
If you are at a larger research institution, you probably also have an IT department that can co-locate your hardware for next to nothing (compared to cloud). There you also will save a lot of ingress/egress, storage, backup, etc. costs.
Regarding the per student costs, even with cloud instances I would consider running a traditional HPC job system (grid engine, lsf, torque, ...). The MIT had a nice solution with Starcluster [1] to easily deploy a SGE on AWS. It looks a bit dead now though.
AVX are vector instructions. They work on wide register (256bit) where you can stuff 8 32bit floats (or in later versions ints) or 4 (double, or longs not sure if AVX2 or later has support for that), or 16 chars (AVX2 iirc).
APL, as an array language, should, in theory at least, especially profit from automatic vectorization by the compiler. If the auto-vectorization works well, you get get a 4-16x speed-up for free (if you don't have data dependencies).
Likewise, most European countries do not fluoridate their water and the correlation doesn't seem to show any significant difference in dental health.
Just to expand on this point a book: In Germany you buy fluoridated salt (you can also buy non fluoridated salt, you have the choice, no idea why you would do that though). Children and people with dental problems also can get fluoridated toothpaste.
I think this fluoridation thing is a non-issue in Germany, never heard anyone talk about it at least . Also I think the salt solution is a lot more elegant than the tap water one in the US.
I can recommend Clion from Jetbrains. I am very happy with it. Although you have to be ready to drink the cmake kool aid. (I am using it on OSX, not Linux. When I edit something on a server I use vim + cmake + ninja build.)
These studies are usually done as cohort studies. Meaning the study group shares a defining trait (such as the same date of birth).
Obviously this is not as great as a double-blind study. But you can't let one group do real exercise and the other one placebo exercise. So the cohort study is probably the best possible study design in this case.
The people doing these studies try very hard to minimize biases (sometimes failing, but ruling out every possible biases is hard).
That idea seems to be discredited: https://en.wikipedia.org/wiki/Ural%E2%80%93Altaic_languages
The remaining connections of Korean still fascinated me when I first heard about it: https://en.wikipedia.org/wiki/Altaic_languages
Various versions included the Turkic, Mongolic, Tungusic and sometimes the Korean and Japonic languages.
I was surprised of a kind of "UEH-ll" sound that you can hear in both modern turkish (their goodbye) and Korean language (don't have an example) and wondered if they had some kind of connection. Apparently they do!
That is just the interconnect. Grid jobs are usually very compute heavy: "[...] includes the majority of the 100,000 processing cores in the CE." That is the expensive part. We bought p2.16xlarge equivalent nodes for just about 1/20 of the "3 Yr All Upfront Reserved" cost of EC2.
Depending on the utilization (i.e. somewhat low utilization) you could save some money in the cloud if you compare it to the usual 5 years depreciation cycle.
But for scientific computing you usually still use your computing nodes longer than the 5 years of depreciation/support contract (often until they are close to fail).
Another reason why cloud is not very attractive is with how budgeting rules work in those old institutions. You propose a cluster and get the money, all calculated from the start. On-demand cloud pricing will be very hard to get approval for.
[...] other costs like networking and storage [...] I am currently using the equivalent 20k dollar in yearly EBS costs on our scientific computing cluster. You can buy a lot of storage for that. And the internal IT department that exists anyway will manage it.
Cloud can be a very hard sell to scientist.
Edit (a point I forgot): Cloud only makes sense if you can use spot market capacity extensively, then it can become very competitive with local installations (again depending on required utilization).
Landauer's principle would come into effect if you had a theoretical "Computorium" material that could perfectly turn energy into computation. Then the computation you could produce per volume of space would be limited by it.
Our brains are (very) far from being the perfect computing material (so are our chips).
The theoretical minimum is only useful here to show that our brains cannot do free work. (Irreversible) computation always has a cost.
The 20 Watts are the important number. Our brains have a certain energy budget and have to perform within that constrain.
Like the sibling post to this I also have a bioinformatics background and awk is one of the most important tools in my toolkit. It is just so extremely fast. A few years ago I was laughing about a friend that did a lot of simple data transformation in awk and told him to save his time use python instead. He challenged me to do the same task in python. It took me longer to write and it was magnitudes slower. That was eye opening.
It is useful to be aware about the different awk implementations. For example mawk. That awk version is stupidly fast. GNU awk is already extremely fast, but if I work on a > 200gb file, I just prefix the m and the task will be done even faster. It however does not have all the features of gawk. The differences in regex support are the most painful.
A very useful feature of the GNU version of awk is the debugger. Super useful to be able to step line wise through a awk script.
If I ever have too much time I would like to take a look at that experimental llvm-awk (lawk) someone started and get it production ready. Sadly is seems abandoned currently.
Bill Gates is a good example for OP's lot and luck point:
"In 1980, [Mary Maxwell Gates] discussed her son's company with John Opel, a fellow committee member and the chairman of International Business Machines Corporation (IBM). Opel, by some accounts, mentioned Mrs. Gates to other IBM executives. A few weeks later, IBM took a chance by hiring Microsoft, then a small software firm, to develop an operating system for its first personal computer." [1]
Just an anecdote, but I like OP's point with lot and luck.
I was there! About 8 years ago during a school trip. The church lies next to a picturesque small lake, and once inside its simply stunning.
I regularly took part in tours of the local super computing center (Munich), but the BSC is really something else. Would definitiv recommend going to see it (book in advance!).
If the speech Facebook is forced to delete is illegal, there must be criminal investigation, by law.
I agree with that. Facebook should not be the instance deciding what is to be deleted or not. The german judicial system should be involved and offenders be prosecuted within the judicial system.
Please do not assume that US-American ideals are universal.
Germany has a complicated history, which it has and will always struggle with. Restricting hate speech and symbols is completely legal and also seen as an important and good thing. The modern German constitution was built especially to defend its democracy even against a majority hostile towards it (Artikel 20). Germany has learned its own lessons from history.