patience_and_focus

From the article - "AI should not waste time trying to understand the viewpoints of people who distrust artificial intelligence for a living." Who has the day time job as "distruster" of AI?? GPT-3, are you looking at me. :o

AstraZeneca Covid-19 vaccine study put on hold due to suspected adverse reaction in participant in the U.K. https://www.statnews.com/2020/09/08/astrazeneca-covid-19-vaccine-study-put-on-hold-due-to-suspected-adverse-reaction-in-participant-in-the-u-k/

I'm not gonna defend the results. However, what in your (and author's) opinion should have been done? The simple answer is not having exams at all. Would that have worked better for poor kids? Colleges would have been forced to use the same (or similar) info for their admissions that the computer used for exam: "an array of student information, including teacher-estimated grades and past performance by students in each school". There is also a conclusion by author: "Algorithms should not be used to assign student grades." This is bullcrap. First of all, they already are with pretty much zero opposition: https://www.ets.org/gre/revised_general/scores/how/ (yeah, human is in the loop, but still) Second, the answer is to improve algorithms rather than discard them. Author also is wrong on a number of other counts: they don't agree with "Computers make neutral decisions" - yeah, computers can have bias, but human graders definitely have bias - and are susceptible to fatigue, misunderstandings, and even fraud. I'd guess that's one of the reasons why ETS uses algorithmic scorer in addition to human one. Author tries to score a lot of points with claims: "Algorithms can’t monitor or detect hate speech, ... they can’t predict crime, they can’t determine which job applicants are more suited than others, they can’t do effective facial recognition" - except that algorithms can do all of these and they do all of these and they are getting better in doing all of these. Yeah, you can prohibit using AI for facial recognition by law, but it does not mean that algorithms are or won't be better in recognizing people than people are. Anyway, it sucks to be caught in this, but the way to go is to improve algorithms rather than giving up and going back to warm and fuzzy human-graded default. Yeah, I didn't claim that use of algorithms is bad in all cases (I'm not a luddite). Also the author is certainly biased herself in many ways. But I do think the academic process such as admission is opaque to start with, has become highly political and contentious with emotions running high. We don't need another dose of opaque criterion in this mix right now. If we can lay out the rules of the game beforehand, that will help (this applies to the current process as well). The way I understood, there was no attempt to explain how the algorithm reached its decision (for example provide weight on each of the factors chosen by the algorithm). Without transparency this is a recipe for disaster.

Reviving this thread. This is unfortunate use of algorithms/machine learning/AI. When Algorithms Give Real Students Imaginary Grades https://www.nytimes.com/2020/09/08/opinion/international-baccalaureate-algorithm-grades.html

This is absolutely fascinating. Scientists, independently of this finding above, are also trying to now create a "lattice-work of models" from so many different streams of data and findings and putting them all together. Impressive that it has happened in less than 1 year of first reporting this virus. A model like this, even with some errors here and there, is going to be invaluable to further improve discovery of new treatments and medical practice. Another example of such a model is in cancer immunotherapy with the understanding of cancer immunity cycle (see https://www.roche.com/research_and_development/what_we_are_working_on/oncology/cancer-immunotherapy/cancer-immunotherapy-cycle.htm). It took at least 30 years to get here for this one.

Fact check: sales of one company's divorce software are up 34%. Exactly. If only someone would do a simple click to the original source article from the ny post page, it also has following - "During the COVID-19 crisis, we experienced the highest purchase of divorce agreements per capita in the South followed by the Midwest, the West, and the North East. The rate of divorce in the south was two to three times higher than the rest of the US regions. The Southern states with the highest divorce rates were: Mississippi Oklahoma Arkansas Alabama Louisiana" Hmmmm, so states mostly resisting lockdown had higher divorce. May be we should lock down more to save marriage instead of banning gay marriage. :o

The average person should just try to find a low cost broad based index etf or similarly low cost target date fund and call it a day. Too bad many fall into traps of "advisors." :-\ Unfortunately the reality is not that simple. Many private employers do not offer good low cost index funds but only ETF's with high fee structure (my former employer did that unknowingly - thank you outsourced HR). Employees have no control in many cases and too much hassle to go outside to set up Roth IRA or something like that due matching contributions and other benefits from the employer.

Covid-19 Is Creating a Wave of Heart Disease Emerging data show that some of the coronavirus’s most potent damage is inflicted on the heart. By Haider Warraich Dr. Warraich is a cardiologist. https://www.nytimes.com/2020/08/17/opinion/covid-19-heart-disease.html "Eduardo Rodriguez was poised to start as the No. 1 pitcher for the Boston Red Sox this season. But in July the 27-year-old tested positive for Covid-19. Feeling “100 years old,” he told reporters: “I’ve never been that sick in my life, and I don’t want to get that sick again.” His symptoms abated, but a few weeks later he felt so tired after throwing about 20 pitches during practice that his team told him to stop and rest. Further investigation revealed that he had a condition many are still struggling to understand: Covid-19-associated myocarditis. Mr. Rodriguez won’t be playing baseball this season."

Nail in the coffin Malaria Drug Chloroquine Does Not Inhibit COVID-19 Infection in Human Lung Cells https://www.nature.com/articles/s41586-020-2575-3 "These results indicate that chloroquine targets a pathway for viral activation that is not operative in lung cells and is unlikely to protect against SARS-CoV-2 spread in and between patients."

If this study is independently "replicated" by other researchers, this is very valuable information not just for clinicians on the front lines but also for patients themselves. For example, imagine a person in his/her 40's or 50's has Covid-19 like symptoms, doesn't have too many co-morbidity issues like heart disease. Given the testing mess, they also don't have the results (its taking average of 5 days or more again in many hot-spots). If the telemedicine clinic they call for advice is not useful they can at least have some idea of whether they should be driving to the nearest emergency room by matching this list and overlap with their symptoms. It may save lives.

On the other hand a cheap generic drug has actually been shown to be effective in double blind randomized clinical trials for saving lives of the very sick Covid-19 patients (hint: its not HCQ) https://fivethirtyeight.com/videos/how-a-60-year-old-drug-became-our-best-hope-for-saving-people-with-covid-19/

https://www.medrxiv.org/content/10.1101/2020.07.15.20151852v1 The mortality rate is high of about 25%. The number of days since symptom onset 9 days before medicine administration On oxygen or ventilation at baseline: 67% Many doctors who use Hydroxychloroquine say 1) It works better with Zinc, and optionally Azithromycin. 2) Works when given early but not when given later when symptoms of pneumonia have started. For example Dr. Harvey Risch from Yale says: I think that there has been confusion about treating the cold versus treating the pneumonia. These medications don’t seem to work so well for treating the pneumonia. As early as possible is crucial, within the first five to six days of symptoms. https://medicine.yale.edu/news-article/25085/ I think you are on the wrong forum trying to push for HCQ+/- other agents. You are wasting everyone's and your time here. I would suggest you either talk to like minded physicians and start a trial with your design in mind and/or fund them personally with your money for this. Another option is to do a go-fund-me site to raise money for such a trial or create a social network group of like-minded people with money to pool resources for this. It is not unheard of because there are many patient and specific drug advocacy groups around that do this kind of thing in other disease areas. Also, there will be enough physicians in this country (and participants as well) that believe in this HCQ+other agent treatment and will be willing to participate/help. In-fact you can start by contacting the docs you keep citing (given the papers you keep posting). Just to be clear, I am not being sarcastic. If you so strongly believe in this, have some skin in the game by putting your own money into it. Just like we all do when investing.

It is bullshit. Don't know if the tweet is deliberate misrepresentation or incompetence, but given its fox news reporter, it probably both. I would not recommend following twitter accounts of fox new reporters for accurate reporting. See the source here (linked in the fox new article) and specifically the table from page 25 onwards for testing by laboratory - http://ww11.doh.state.fl.us/comm/_partners/covid19_report_archive/state_reports_latest.pdf 1. The testing results by lab is sorted high to low by # of tests administered and columns show # inconclusive, # of negatives and # of positives as well. It is clear that > 90% of the data is reporting correctly all numbers - inconclusive, positive and negative numbers. 2. Not only that, this is a long tail where the first 50 or so labs (page 25) account for ~ 80% of the total test results and are correctly reported. Infact, the top lab 10 labs itself report ~ 2 million tests. 3. There are labs reporting 100% positivity (not reporting negatives - they start to show up on page 29) but they are conducting very few tests comparatively. 4. Lastly, a lot of labs have reported conducting a total of 3 tests or less (page 40 on-wards). That is a very low number of tests and it is actually quite likely that all of them turn out positive or all of them turn out negative. This is because sample size is small. In fact, many labs do report 0% positivity when # of tests are very few. Vast majority of data is solid and of-course there are labs that do not report accurately but their influence on the data is rounding error. Thankfully the FL state health dept is still competent and not yet overtly political.

NIH halts clinical trial of hydroxychloroquine https://www.nih.gov/news-events/news-releases/nih-halts-clinical-trial-hydroxychloroquine

I think we all should do some deep DD (as we do when investing) before posting on technical things like effectiveness of a treatment. Moreover, I really think these BS studies should not be published and will not be published in any reputed journal. There are plenty of substandard journals that exist where these show up and someone incompetent in the media picks it up. Here is why this is a total BS study - (a) This is an retrospective observational study, not a randomized double blind clinical trial which is the gold standard and has shown no effectiveness. (b) Look at Table 1 of this study (https://www.ijidonline.com/article/S1201-9712(20)30534-8/fulltext) - the age itself can explain the difference in mortality. The patients receiving hydroxychloroquine alone (median age = 53) were younger by 18 years on average than who received no treatment (median age = 71). We already know younger patients survive better than older ones. So the patients treated with hydroxychloroquine simply lived because they were younger. One couldn't have picked a worst retrospective dataset than this for the analysis.

It depends. Although its still unknown scientifically, one emerging trend from the data so far says that if a person gets "mild" or "moderate" disease then the antibodies fade in a few months. But in the "sufficiently" sick, they have high levels of antibodies until and after convalescent phase of the infection (recovery and after). So far, the followup time has been limited. Remember, we only have had ~5 months of real tracking of this virus in humans and the infection itself lingers for at-least month in many sick patients. So this data is very preliminary. It is well known scientifically that how long immunity will last for an infection (bacterial or viral or fungal) is very variable and depends on a lot of factors. For example polio and measles immunity (gained via an infection or a vaccine) lasts lifetime. Whereas flu lasts a year, getting tetanus infection and Hep C infection provides barely any long lasting immunity. So in short, this is so new that we don't know yet.

It is true that 30,000 cases today is not the same as 30,000 cases in April. But, the "just doing more testing" is wrong and is essentially a cover-up. Ontario is testing at roughly the same per-capita rate as Florida. And as I posted up-thread, there is actually an inverse correlation between testing in Ontario and cases. I agree that we aren't at a "gloom and doom" phase. But Texas needed to cancel elective hospital procedures because they reopened bars. This seems like a poor use of your "Rt budget". Don't think that you can attribute solely reopened bars for the increased spread. I think the widespread protests that has been occurring for a month now likely created plenty of vectors for spread. That some places are now seeing outbreaks shouldn't come as a surprise. What is a bit surprising is that NYC, by far the hardest hit city in the US, keeps seeing declines. Both are culprits but to different degrees. Bars are inside an enclosed place exposure, protests (and beaches and parks) are outside in open space exposure. This has dramatic impact on viral load and consequent spread and severity.

This is a good short read on this question - https://www.nature.com/articles/s41564-020-0690-4.pdf?proof=true1

Programming is a very good case study. Webex has been around for 20 years. Skype has been around for 15 years. Yet the image of a lone programmer coding away on a Hawaii beach is (mostly) a myth. Yes there are those, but very small percentage of overall number. If anything despite the new age tools, co-location was driving innovation in the software world. Now there is always an argument that webex and skype were clunky. But they were better than nothing during early-mid 90's, and still the co-lation of software engineers into city centers accelerated after that.

Agree in the short term. Disagree in the medium to long term (> 5 yrs). The powerful economic forces that have been making large cities grow even larger, especially over the last 30-40 years (shall I hazard to say at least a 100 years) are not going anywhere. This is a bump in that process. Large cities like London survived bouts of plague in the distant past. Modern cities like SF and NYC survived spanish flu, HIV etc. Asian cities like Hong Kong, Shanghai, and Singapore survived SARS (far more deadly than this current SARS-CoV2) and grew nevertheless in the last 20-25 years. That doesn't mean there won't be damage and bankruptcy in real-estate tied to offices in the next 1-5 years. But I don't see long term trends changing, especially when this virus does not hit the young and urban crowd hard. The old folks living in cities or nearby crowded suburbs may move out at faster pace. Frankly that is desirable because at least in Bay area housing and construction is depressed partly due to NIMBY supporting older folks and outdated zoning laws. Young people are far more receptive of changing zoning and allowing construction of both residential and office buildings.

Clearly many in the white house have not gotten this memo and are setting a different example for their supporters to emulate.

Of course masks are good. I was also highly in favor of masks early in this thread. You just can't keep saying that masks aren't being pushed because of WHO. It's also because of Trump, who won't even wear one, who won't use his influence and reach to convince people to wear them (one of the main powers of the presidency is explaining things to citizens and persuading them of something), and who didn't order emergency production of PPE months ago despite talking about it repeatedly, and has greatly contributed by his words and actions to politicizing them (should seatbelts and condoms and vaccines also be politicized? so stupid). It's not like other countries can't and haven't pushed masked whatever the WHO said early on, or if they listened at first, haven't later course-corrected rather than kept going in the wrong direction. Hallmarks of a good investor (and should I dare say someone who is good at anything else) is to realize their mistakes if they have been made (they don't even have to publicize them) and course correct when facts change on the ground. Can't say that with the current white house - https://www.cnbc.com/2020/05/21/trump-doesnt-wear-coronavirus-mask-to-ford-plant.html

This is what I call crappy "data analysis". This person who wrote this bloomberg article will fail any data analysis 101 course. Here are just a couple of the many issues - (a) The article says - "But, as our next chart shows, there’s little correlation between the severity of a nation’s restrictions and whether it managed to curb excess fatalities — a measure that looks at the overall number of deaths compared with normal trends." Where the hell is correlation coefficient value? And the p-value? And the chart is NOT an X versus Y plot of stringency score versus deaths. The stringency score used for correlation was at what time (since stringency concept changes over time)? Is the correlation a time series auto-correlation or point correlation? A claim is made in the above sentence about "little correlation", but no numbers to back it up. (b) Here the article conveniently provides an X vs Y plot for severity of restrictions score vs economic activity and shows a trend line. Where the hell is correlation coefficient value? And the p-value? Just looking at that graph I can tell you that correlation for this will be crappy as well (both coefficient and p-value). The trend line is just smoke and mirrors. This is obfuscation 101 when someone posts a trend line but no numbers associated. Then of-course, there are numerous question like why certain countries in Europe were left out in the charts for this article etc. It requires a good journal paper level rigor, not this mess of an article. Seems to me that author had a theory in mind and simply massaged/ scrubbed the original source data to prove a point. The original data is here - https://www.bsg.ox.ac.uk/research/research-projects/coronavirus-government-response-tracker

We have crappy diets, resulting in first world problems like high blood pressure, diabetes etc. In rural China and rural Africa, people have 110/70 blood pressure well into their old age. In the western world, with processed foods and high salt and high sugar, we have extremely high blood pressure and diabetes rates very early in life. Our bodies don't fight off viruses as well as healthy people. My view is it's largely diet based. All, the above and second and third world countries also have reporting issues, plus the epidemic has not run its. course yet. There are reports of bad situations in Ecuador (Quito) and Brazil but numbers are hard to come by. While these and other points mentioned above are all valid, I highly doubt they can fully explain the huge discrepancy between the developed and the developing countries. And especially considering the big factor that should make situations worse in developing countries -- their lack of good healthcare systems. I do wonder whether how some countries do not care much about this virus and this is being reflected in recognizing/reporting the COVID death numbers. My wife's coworker (they work in healthcare), who has families in Bangladesh, told her yesterday that while COVID is spreading there, people are more worried about going hungry than the virus. Suppose your people, media, and government do not really recognize this virus as anything novel or serious... In such countries, even if people die due to COVID or related illness, they might not warrant much attention and won't be tracked like some doomsday counter. In that sense, is COVID another "first world problem"? Yup. I think this is on the money. Media coverage and fear mongering have made this what it is. Its now widely recognized that this was here much earlier than some people thought, and guess what? Life was totally normal and folks got on with their normal business and the economy was humming along just fine. So yes, its a shame we manufactured a horror story and certainly did impair parts of the economy, probably unnecessarily. But, in other news. Shanghai Disney tickets sold out. RCL is reporting normal booking volume for 2021, guess not everyone is living under a table in their basement. Huh? Maybe I am missing something but the denominator for per capita death rate is simply total population of a country (https://www.statista.com/statistics/1104709/coronavirus-deaths-worldwide-per-million-inhabitants/). So China and India will show a small number for a very long time unless this is allowed to spread unhinged. I know there is general distrust of death numbers (numerator) attributable to covid but even if/when they start reporting "correct" ballpark numbers in the numerator, the denominator is what will skew and influence interpretation. This is really a very bad metric to compare countries on some kind of effectiveness or other theories on why deaths are higher because of the effect of denominator that has no direct relevance to deaths. It is simply an accident of population growth trajectory. Here it is without the denominator. I see a similar pattern, still. Now we just need to divide this by the approx # of days since first 1000 detected infections to get account for different starting points of large scale community spread. And why 1000? Because we need a reasonably large reported number to make sure community spread is well underway. We can pick another number here but has to be consistent across the board. For example Brazil will be early April, India will be starting in mid April whereas US will be in early to mid-Feb if I remember correctly. Slowly we are converging towards the correct way to look at things across countries using this metric. Even with this, it won't answer all questions. It will answer how quickly deaths rose (and plateaued) in each country post start of community spread. P.S: Just to be clear, this will be a lagging indicator in economic metric paralance.

We have crappy diets, resulting in first world problems like high blood pressure, diabetes etc. In rural China and rural Africa, people have 110/70 blood pressure well into their old age. In the western world, with processed foods and high salt and high sugar, we have extremely high blood pressure and diabetes rates very early in life. Our bodies don't fight off viruses as well as healthy people. My view is it's largely diet based. All, the above and second and third world countries also have reporting issues, plus the epidemic has not run its. course yet. There are reports of bad situations in Ecuador (Quito) and Brazil but numbers are hard to come by. While these and other points mentioned above are all valid, I highly doubt they can fully explain the huge discrepancy between the developed and the developing countries. And especially considering the big factor that should make situations worse in developing countries -- their lack of good healthcare systems. I do wonder whether how some countries do not care much about this virus and this is being reflected in recognizing/reporting the COVID death numbers. My wife's coworker (they work in healthcare), who has families in Bangladesh, told her yesterday that while COVID is spreading there, people are more worried about going hungry than the virus. Suppose your people, media, and government do not really recognize this virus as anything novel or serious... In such countries, even if people die due to COVID or related illness, they might not warrant much attention and won't be tracked like some doomsday counter. In that sense, is COVID another "first world problem"? Yup. I think this is on the money. Media coverage and fear mongering have made this what it is. Its now widely recognized that this was here much earlier than some people thought, and guess what? Life was totally normal and folks got on with their normal business and the economy was humming along just fine. So yes, its a shame we manufactured a horror story and certainly did impair parts of the economy, probably unnecessarily. But, in other news. Shanghai Disney tickets sold out. RCL is reporting normal booking volume for 2021, guess not everyone is living under a table in their basement. Huh? Maybe I am missing something but the denominator for per capita death rate is simply total population of a country (https://www.statista.com/statistics/1104709/coronavirus-deaths-worldwide-per-million-inhabitants/). So China and India will show a small number for a very long time unless this is allowed to spread unhinged. I know there is general distrust of death numbers (numerator) attributable to covid but even if/when they start reporting "correct" ballpark numbers in the numerator, the denominator is what will skew and influence interpretation. This is really a very bad metric to compare countries on some kind of effectiveness or other theories on why deaths are higher because of the effect of denominator that has no direct relevance to deaths. It is simply an accident of population growth trajectory.