mindviews

I got one of these $25 kits in a physics class that gives you the eyeball version of the fancy digital spectrometer: https://shop.sciencefirst.com/starlab/kits/5800-cardboard-sp...

Here's a link to the paper: https://www.medrxiv.org/content/10.1101/2020.05.06.20092999v...

This is interesting data. I want to see Vitamin D status included in a large population study like this because I've been following two smaller studies covering about a thousand cases total that shows Vitamin D deficiency has a risk ratio of 10 to 20 (more even than being age 80+ in the above study). The studies:

[1] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3585561

[2] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3571484

Table 1 in each of those papers show Vitamin D status vs Outcomes. The correlation between Vitamin D status, where Normal is >30 ng/ml (or 75 nmol/L), and death rates is stark.

From [1] (which had n=780 cases) here is the punchline: "98.9% of Vitamin D deficient cases died while only 1.1% of them were active cases. 87.8% of Vitamin D insufficient cases died while only 12.2% of them were active cases. Only 4.1% of cases with normal Vitamin D levels died while 95.9% of them were active cases."

From [2] (which had n=212 cases) here is the punchline: "Of the 212 (100.0%) cases of Covid-2019, 49 (23.1%) were identified mild, 59 (27.8%) were ordinary, 56 (26.4%) were severe, and 48 (22.6%) were critical (Table 1). Mean serum 25(OH)D level was 23.8 ng/ml. Serum 25(OH)D level of cases with mild outcome was 31.2 ng/ml, 27.4 ng/ml for ordinary, 21.2 ng/ml for severe, and 17.1 ng/ml for critical."

Note: the classification for outcomes was "(1) mild – mild clinical features without pneumonia diagnosis, (2) ordinary – confirmed pneumonia in chest computer tomography with fever and other respiratory symptoms, (3) severe – hypoxia (at most 93% oxygen saturation) and respiratory distress or abnormal blood gas analysis results (PaCO2 >50 mm Hg or PaO2 < 0 mm Hg), and (4) critical – respiratory failure requiring intensive case monitoring."

I want to see a dozen more studies like [1] and [2] to see if this holds up to replication with larger populations.

Here's the paper mentioned in (a) above that's the source of the claim: [1] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3585561

Here is another paper with similarly stark data: [2] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3571484

Table 1 in each of those papers show Vitamin D status vs Outcomes. The correlation between Vitamin D status, where Normal is >30 ng/ml (or 75 nmol/L), and death rates is stark.

From [1] (which had n=780 cases) here is the punchline: "98.9% of Vitamin D deficient cases died while only 1.1% of them were active cases. 87.8% of Vitamin D insufficient cases died while only 12.2% of them were active cases. Only 4.1% of cases with normal Vitamin D levels died while 95.9% of them were active cases."

From [2] (which had n=212 cases) here is the punchline: "Of the 212 (100.0%) cases of Covid-2019, 49 (23.1%) were identified mild, 59 (27.8%) were ordinary, 56 (26.4%) were severe, and 48 (22.6%) were critical (Table 1). Mean serum 25(OH)D level was 23.8 ng/ml. Serum 25(OH)D level of cases with mild outcome was 31.2 ng/ml, 27.4 ng/ml for ordinary, 21.2 ng/ml for severe, and 17.1 ng/ml for critical."

Note: the classification for outcomes was "(1) mild – mild clinical features without pneumonia diagnosis, (2) ordinary – confirmed pneumonia in chest computer tomography with fever and other respiratory symptoms, (3) severe – hypoxia (at most 93% oxygen saturation) and respiratory distress or abnormal blood gas analysis results (PaCO2 >50 mm Hg or PaO2 < 0 mm Hg), and (4) critical – respiratory failure requiring intensive case monitoring."

I want to see a dozen more studies like [1] and [2] to see if this holds up to replication.

The show notes for the parent video are quite comprehensive with links to references. The summary above leaves out the detailed discussion of the interplay between Vitamin D, the renin-angiotensin-system, the ACE2 receptor, and SARS-CoV-2: [3] https://www.foundmyfitness.com/episodes/vitamin-d-covid-19

Here's a description of a "press-pulse" protocol that includes fasting, calorie restriction, and ketogenic diet scenarios (main goal seems to be to limit glucose intake as part of the treatment as part of an effort at stressing the energy systems of the cancer cells to cause cell death): https://nutritionandmetabolism.biomedcentral.com/articles/10... Check the references for some other papers on fasting and chemotherapy. Also, here's a related human case study: https://www.ncbi.nlm.nih.gov/pubmed/29651419

A couple of interesting publications on Glioblastoma treatment that might be pointing in the right direction for targeting glucose and glutamine starting with mouse models: https://www.nature.com/articles/s42003-019-0455-x and progressing to human case study: https://www.ncbi.nlm.nih.gov/pubmed/29651419

"The Grover search as a naturally occurring phenomenon" arXiv link: https://arxiv.org/abs/1908.11213

I've used it on Linux (Raspberry Pi for my case) just fine. Guides here if you need it: https://www.libcinder.org/docs/guides/linux-notes/index.html

While lacking in lots of built-in UI widgets, I use Cinder like a UI library. https://libcinder.org/ Perhaps "graphics library" would be a better descriptor and explains why I didn't find it on the list. But with mouse/touch, images, audio, video, etc. support, "graphics library" seems too narrow of a description. Maybe it's better to think of as a GUI library without a lot of commitments to the usual set of GUI widgets.

I suppose a bit more context would be explanatory. My real interest is not the free-space performance of the antenna, but the eventual installed performance. In the video describing the setup he says "I talked with a couple of engineers, and some other guys, and we've come to the conclusion that if I put 100 watts into it, it might radiate a milliwatt. And not all that well." I'm not sure if they were thinking of it as a magnetic loop antenna the way you described or coming up with the numbers some other way. But then at the end mentions, "I'm gonna set it on the top part of that air conditioner and put it right in the window." How much of his success is a result of sticking the antenna on what may be effectively a (admittedly electrically small compared to 20m) metal box? That's what I think is interesting to simulate.

Can anyone simulate this light bulb setup in HFSS or another antenna simulation software? I'm always on the lookout for comparisons on weird cases like this for the simulation code I've written. If I can find material property data on custard as easily as for tungsten (watch the temperature dependence!) I might give that a shot next.

>Adjusted for inflation, the US economy has more than doubled in real terms since 1975. How much of that growth has gone to the average person? According to many economists, the answer is close to zero.

Simple question to ask yourself and anyone you know: would you rather be alive in your income bracket (inflation adjusted, etc.) today or 30 years ago?

I keep asking this question to people I've met and have yet to have any takers for the 30 years ago option. Clearly these types of economic measurements are missing something important. Deflationary technology improvements not being properly taken into account? Something else?

I'm worried based on some of the comments that people might be reading only the Impartial view and not also the Left-bias and Right-bias versions of the same story. I'd strongly suggest reading all three versions of a story before using the slider to provide feedback on the various versions.

This is great work! I have wanted some version of this news site for years and have been making sketches for how it would work. My working title for the site is "Unspun" and very similar in spirit, except instead of the "Impartial" view, the "center" would just be a list of facts about things that happened that were referenced in both a Left and Right version of an article. And there would be lines connecting the center facts to where they show up (if they do at all) in the Left and Right articles. I'm pretty happy with this format, but I still kind of want to see all 3 versions at the same time so I can cross-reference. But I must like it because I just sent links to a whole bunch of friends and family. :-)

For those wondering like I was after reading this comment, both the anandtech and phoronix test setups are patched for Spectre.

For Machine Learning, I found "Learning From Data" http://amlbook.com/ to be a very strong, foundational book. It does the best job of any reference I've seen of answering the question: "Is learning possible and feasible?" Provides and excellent mathematical foundation for understanding when and why ML systems should succeed and fail. Associated MOOC from Caltech at https://work.caltech.edu/telecourse.html

A study group meetup (Every Tuesday evening in Austin, TX): https://www.meetup.com/cppmsg_ai/

Just Q&A - no presentations. Study from whatever books (http://amlbook.com/ and http://www.deeplearningbook.org/ are popular in our group) or courses (Andrew Ng's are also popular) you like throughout the week and then show up with any questions you have. We've been meeting for a couple of months now and new folks are always welcome no matter where you are in your studies!

I don't know...based on his writing I'm guessing that under special circumstances he might.

That's one big thing I was worried about. Also, "We included only races with an average number of finishers greater than 2,000 (for all four distances)" could be a major confounding factor. What if many "competitive" runners switched to smaller events as big events became more "general"? Until this major issue is addressed, the conclusions can only apply to the self-selected set of American runners in large races.

Yes to lossy dielectrics. The caveat being that right now we only have support for homogeneous materials - we have some thoughts on how to bring our methods to continuously varying materials, but that's still a research topic.

We started out focused on RCS problems for algorithm development and validation, but we're shifting to more antenna design and analysis (mounted antennas, installed performance, placement optimization). We have done near-field excitation of our own models on large structures, but usually our goal has been to maintain accuracy so our use case has us solve the driven antenna and the platform together in one go.

>Have you tried stripline as a benchmark?

No - just pulled the paper on it and put that on my to-do list. We've been focused on large problems recently and people seem happy to stick with scattering by spheres (PEC or dielectric) and comparison with the Mie solution. Way too much symmetry to serve as a comprehensive benchmark, but a decent way to compare computational efficiency. Our current benchmark run for a 100 wavelength diameter PEC sphere is 48 minutes on 256 CPU cores with 0.13% RMS error in the far field. We recently got 1.8% far field error for the 500 wavelength case on 300 cores in 17.8 hours. Our preliminary 1,000 wavelength numbers are very promising, too. No GPU/MIC or unusual hardware for those tests - all on a cluster of modern servers with Intel Xeon CPUs with 2-4 GB RAM per core.

Finding good benchmarks for sharp corners has been more challenging. The one we've been using for that is planewave scattering by a PEC cube and we test that the fields inside are 0 everywhere (including arbitrarily close to the surface at corners and edges).

Thanks for your other comments - geometry translation comes up often. Post-processing as you mentioned elsewhere is a common pain point, too, but solutions there seem to be pretty application/domain specific.

What's your take on model accuracy? I ask because my startup has developed a high-order accurate CEM solver (think MoM full-wave type solvers but every time you double the mesh density you get a 100x accuracy improvement thanks to a bunch of algorithmic breakthroughs). But we've also been putting in a lot of work to make CAD import fast and easy and it's not clear to me what's going to get people more interested - the technical performance of the solver or the usability of the GUI. I'd be interested in thoughts on that. Or anyone interested in getting hands-on, send me an email (see my profile).

We should be talking about "consistency" being the fundamental idea, not causality. I think "retrocausality" is an especially distracting choice of words.

I find the clearest way to think about this is in terms of Feynman Diagrams. Here's a quick intro to some of the rules: http://bolvan.ph.utexas.edu/~vadim/classes/2008f.homeworks/Q... If you're not familiar, ignore the math and just look at the pictures on the first couple of pages.

Take a simple diagram of a photon interacting with an electron (google QED Vertex). Here's a very tiny ASCII version: ~< The squiggly line represents a photon propagating and the straight line segments (which should have arrows pointing a direction) represent an electron propagating. We can rotate this thing around in a bunch of different ways in time so that we have 1 or 2 inputs ("before") an 2 or 1 outputs ("after") in time. For example, with time going left to right, ~< represents a photon decaying into an electron and positron. Flipped around, >~ represents a electron and a positron colliding/annihilating to create a photon. Turned another way you could have a photon and electron as input, and an electron with a changed momentum as the output. For the last case you could say the electron absorbed the photon. But really, these are all the exact same pattern just rotated around in spacetime. So, what is causation? If it's all the same pattern, it's clear that consistency with the pattern is more important than the direction of time's arrow.

Speaking very loosely now (there are a bunch of constraints and caveats on what I'm about to say), you can plug these diagrams together to make arbitrarily complicated internal structures. But if they have the same inputs and outputs, they are in a sense consistent. And if you do it just right you can constrain which types of patterns can link up with the one you've set up. So, in the end, only patterns that are consistent with your setup can happen. Which is pretty much what this article is describing.

That's not really the right reasoning to use here. The number of cell divisions isn't 50 vs. 54, it's more like 2^50 vs. 2^54.

At some point the animal will exit the growth phase and reach a stable cell count and an elephant that reaches adulthood will just simply have more cells than a mouse. A 5,000kg elephant has a lot more cells that could develop cancer than a 0.02kg field mouse. And if that elephant lives 60 years instead of 1.5 for the mouse (let's just say for the sake of argument that cells divide once per year for replacement), that could be something like a 10,000,000 fold difference in the number of "cell-years" and cell divisions (at once per year) for something to go wrong and cause one of those cells to become cancerous.

"Peto noted that, in general, there is little relationship between cancer rates and the body size or age of animals. That is surprising: the cells of large-bodied or older animals should have divided many more times than those of smaller or younger ones, so should possess more random mutations predisposing them to cancer. Peto speculated that there might be an intrinsic biological mechanism that protects cells from cancer as they age and expand."

So, yeah, it seems like something important has to be going on. If a mouse can die of cancer at 1 year old, how can any elephants survive to 60?

I see a couple of other responses here, but let me address why fluoride belongs in _drinking water_

Fluoride belongs in drinking water in much the same way iodine belongs in salt---it's an extremely cheap, low-risk, and effective way to improve public health. Read some of the references for more details https://en.wikipedia.org/wiki/Water_fluoridation#Effectivene... but the short version is that fluoride is safe and effective and because it's so widespread in use, even if you only drink bottled water you're probably getting secondary exposure from other food/drink sources. And if you brush your teeth regularly with a fluoride toothpaste, then you've got that delivery mechanism covering you as well. In the end, the public health benefits are there because fluoridation is pervasive and hard to avoid. By similar argument, you probably also don't have an iodine deficiency because of all the iodized salt in use.

This is interesting because people relying on anecdotal evidence is pretty much the same failure mode for why so many people don't recognize the importance of vaccination. Because it's so pervasive in the US (and other places) it's easy to find stories of "I wasn't vaccinated and I didn't get sick" or similar "I didn't X and Y didn't happen" but it's not just the primary exposure, but all the secondary exposure and effects that also play an important role in public health efforts.

And last week I heard that story but the difference was that the images of the NATO tanks were in perfect focus while the Soviet tanks were out of focus in the training images. Clearly this story is all over the place. I wonder if even the likely source is correct. https://www.webofstories.com/play/marvin.minsky/122

"There's no mechanism that would cause someone's inner character to be reflected in their appearance in any consistent way." That's a pretty big leap. Down Syndrome has consistent physical characteristics that correlate with a particular set of cognitive/behavioral characteristics. While it is important to carefully critique scientific findings that may be motivated by political biases, it is also important to give science as a process the chance to find truths even if we might not like their political implications.

It's the lack of transparency that's the real problem. I like to know as much as I can when something I own is sharing personally identifiable data about me and my habits to companies (and governments). The fact that the whole effort on Vizio's part was under the radar means that consumers lacked important information about the functioning of their TV's. If they had known about the depth and breadth of the data collection, maybe some portion of purchasers would have made other decisions. Once that's on the table, then you're free to make the choice to let that data be shared if you're comfortable with it (as you indicate you'd be in your case).

https://projecteuler.net/ is an excellent tool for developing algorithmic thinking. From the project description:

The problems range in difficulty and for many the experience is inductive chain learning. That is, by solving one problem it will expose you to a new concept that allows you to undertake a previously inaccessible problem. So the determined participant will slowly but surely work his/her way through every problem.

>And of course, the bottom line is that these bombings worked. All else is speculation.

The imminent USSR invasion from the north played a significant part in the calculus of Japan's surrender. Too often the use of nuclear weapons alone gets credit, but there was a more complex political/diplomatic context surrounding _why_ they worked in the case.

Links for anyone interested in reading more on the topic: https://www.amazon.com/Racing-Enemy-Stalin-Truman-Surrender/... http://archive.boston.com/bostonglobe/ideas/articles/2011/08... http://foreignpolicy.com/2013/05/30/the-bomb-didnt-beat-japa...

Clearly the author didn't find the results from the paper you pointed to particularly compelling. It's listed as reference 136 and mentioned several times. The last of which, "In the absence of more precise knowledge of the extents of analgesic consumption and the durations of exposure, the reliability of these reports [45, 136, 137] is open to question." Is there a reason you find the Stewart paper to be more definitive than this author suggests?