kem

I'm speechless, but also habituating to that feeling.

It's as if we reincarnated oil barons from the early 1900s and appointed them all to lead the government without telling them it was 2017.

Pay-to-publish has its own problems.

I think the current model has issues, and there should be a movement in general toward open dissemination of research, but I'm not sure the PLoS model is much better in a lot of ways. Maybe not worse, but maybe not better.

Where this is all headed is anyone's guess. If I had to, eventually most scientific research will just appear in blogs, being treated like personal journals or something, or in research society journals, kind of like what is referenced in the article, basically run by academic research organizations.

My experience with the second model is that there are organizations that would like to run things that way, but run into problems with copyediting, layout and design, and reviewing infrastructure. They kind of think it's something they can just do on the fly but as they do it they realize it's more work. Maybe a company that just provides the tools is the way to go, but at some point that would probably basically become a publisher.

Not bad to think about this, but I think much smaller changes would have huge effects.

Thinks like approval or rank voting, for example.

I don't understand why these minor changes don't get implemented.

I think there's a bigger problem the article is getting at, that people are kind of missing or downplaying, which is the culture around math skill acquisition and ability.

It's obvious this guy had a love of math and ability in it, from his majors and subsequent events.

But it's also actually not that unreasonable for him to start second-guessing himself when he struggled with math earlier in life.

I do think there's this idea that if you're good at math and have something to offer in it, it will show early on regardless of life circumstances or mentors or role models or whatever, that if it's not immediately obvious that you're a mathematical genius you should forget about it.

"Realizing a test you took in elementary school needn't define you" is actually a nontrivial thing to overcome in today's society, maybe even especially in STEM circles.

Sometimes I wish STEM culture was more focused on sharing the joys of STEM and trying to be as open-minded and inclusive as possible, instead of brandishing it as a competitive tool.

I appreciate this article being posted, and have the utmost respect for NumPy developers. The urgency and discrepancy between use of certain important open-source libraries, and their support, is bewildering sometimes.

As I was thinking about it, though, I'm not surprised NumPy hasn't been funded before. The reasons why say a lot about biases in memory.

It wasn't that long ago that the sorts of things NumPy does were seen as fairly niche, and in the domain of statistics or engineering. It's only with relatively recent interest in AI and DL that this has been seen as within the purview of Silicon Valley-comp sci-type business, as opposed to EE or something different. I still am kind of a little disoriented--the other day, looking through our university's course catalog, I realized that certain topics that would have been taught in the stats or psychology departments are now being seen as the territory of comp sci. Statisticians have written excoriations about being treated as if they don't exist, as comp sci blithely barrels forward, reinventing the wheel.

I'm not meaning to take sides with these issues, only pointing out that I think the world we live in was very different not so long ago. It might seem puzzling that NumPy hasn't had more funding, but I think that's in part because what it's most profitably used for now wasn't really seen as much more than academic science fiction not too long ago.

The other part of it too, is that until relatively recently, if you were to do numerical heavy lifting, you'd almost certainly be expected to do that in C/C++ or maybe Fortran. There's a tension in numerical computing, between the performance and expressiveness that's needed, and Python is on one end of that continuum, far from the end that is traditionally associated with complex numerical computing. Sure, you had things like MATLAB with Python in the same functional role, but those were largely seen as teaching tools, or something that engineers did for one-off projects, having learned to do that in school (I still think the use of python in ML derives from the use of Python as a teaching tool in uni).

I'm not trying to knock Python or NumPy or anything, just kind of trying to convey a different perspective, which is that I can remember a time not too long ago when the use of Python in numerics was seen as primarily didactic in nature, or for limited circumscribed applications.

FWIW, it seems to me Python is kind of on a path similar to what happened with javascript, which was treated as kind of an ancillary helper language on the web, until Google started pushing its limits. Then there was browser wars 2.0, and huge efforts put into javascript, and it became a main player in network computing. To me, there's a similar trend with Python: it really kind of existed as a language for prototyping and scripting tasks, and now finds itself in a different role than it has been used for traditionally, and projects in that area are getting an influx of money accordingly. What I see happening is (1) a blossoming diversity of numerical computing communities (Haskell, Python, Julia, Kotlin, Scala, Rust, Go, etc.), due to competition and variation in application scenarios and preferences, (2) a huge influx of resources being put into Python to make it more performant, or (3) people jumping ship from Python into one of those other platforms to get more bang-for-the buck [or (4) some combination of all of these.]

There are pretty low-hanging fruit out there, though.

Approval voting or ranking are almost universally agreed to be better than what we have (in most places--some places do have that). That alone would make a huge change.

Independent districting commissions would be a big step forward.

Election day holidays would help.

Actually, a lot of the stuff in this proposed bill would help.

You're right that there's open questions out there, but it's not like there aren't solutions that are obvious first steps.

Re: healthcare, I think that's probably even more complicated than electoral reform. For example, missing from your example are cost-cutting measures, like requiring more transparency in pricing and charging, and deregulating certain things to increase competition.

fwiw, it's important to remember that there's a lot of diversity involved.

In my university, for example, the library pushes hard for movement toward open public standards. I've also seen junior faculty make efforts to publish in open journals. But then those junior faculty get slammed by [some] senior faculty for not publishing in higher profile journals and not having enough of an impact factor.

There's also a lot of legitimate grievances against open journals, due to the financial incentives that are created and the types of review processes that are encouraged by their policies.

So there's a chicken-and-egg factor and a lot of disagreement about what should be happening.

Personally, I think there should be a huge shift in publishing practices in academics, but it's going to take some time.

Not really an expert in this particular area, but the research I'm aware of kind of suggests if anything the opposite is true, depending on what you mean by "the ability to learn new things."

There's some cognitive decline, but it tends to start after 25 or so, and it's mostly associated with slowed speed per se rather than learning ability. There's also some controversy in that the declines might be associated with serious health conditions, that are associated with age, rather than age per se.

The other things you mention either stay the same or increase with age, which strikes people as kind of counter-intuitive, which speaks to stereotypes people have.

Your point is important, although I think farmers' political impact exceeds their voting power, especially in GOP-voting communities.

I know from personal experience that this issue and related issues have become very salient to farmers and those around them. It's not a small issue. When your livelihood depends on a very expensive piece of machinery, being able to repair it yourself (or have it repaired by whomever you want) is important, and farmers are talking to their friends and family about it. They're being screwed, and all they want to do is farm.

Also, this issue is bigger than what it nominally seems, because it affects equipment dealers and mechanics as well. I know dealers who are pissed as hell at Deere--they feel stabbed in the back by them because of this stuff.

This repair stuff is also the tip of the iceberg in terms of Deere's behavior lately too. Many people I know would list the right-to-repair issue as just one in a list of grievances they have against Deere. Over the last decade or so, there's been a huge shift in Deere's behavior, and they've really become kind of monopolistic in their general orientation toward who they deal with. This has angered a lot of people. I suspect the right-to-repair in rural Republican areas might be a kind of symbolic issue, representative of much broader anger at Deere.

Regulatory capture, advertently or inadvertently, is one of the elephants in the healthcare room. I'm tired of all the discussion (on both sides) totally ignoring this issue.

I agree with the "not necessarily" part, but one thing that concerns me is in seeing large, critical institutions, like large hospitals and universities, outsourcing critical infrastructure off-premises.

If you look at operational costs per se, and licensing fees in a scenario like that, you're missing a huge proportion of the most threatening costs. Consider hospitals, for example, and recent problems with ransoming electronic health system data.

What concerns me with SaaS is that the biggest costs are very catastrophic or serious risks (like lock-in dependencies, or hacking due to monocultures--by the way, is that a formal, modeled concept in computer security, like it is in agriculture? The vulnerability of a market to "infection" due to monocultures?) that are maybe high probability over a very long-term period, but very low over a short-term period, and hidden.

"as long as there are good competitors around"--that's the key, and not something I necessarily trust to be the case. Net neutrality, for example, would be a non-issue if there were good competition, but there's not.

FWIW, I think now technically indexing can either be 1- or 0-based.

As someone who used Python and R, and also Julia, my take on Julia has changed over time.

My overall assessment is that yes, it's definitely worth getting your feet wet with. It has all the advantages of R or Python (for numerics, a point I will return to), but with much, much better performance. I feel like the syntax is also cleaner than either, although it has more of an advantage over R than Python in that area (I like Python's syntax more than R's).

I had an experience of some prototype R code running for about a day without finishing. The same code in Julia finished in about 5 minutes. It was kind of the final straw that convinced me to gradually move.

Since that time, though, there's one issue that's kind of nagged me, and has only grown over time, which is that Julia is kind of a niche language, like R. It's a big niche, and it might not matter, but over time I've come to appreciate the fact that Python is more general purpose. I am also watching as things like Kotlin, Scala, and Nim gain in popularity and in resources. I suspect that Julia will expand over time, but those others have a head start in some ways (even if they are behind in other ways).

Like some others, I also had some experience of head-scratching changes that occurred with new API-breaking releases. They were subtle changes that were difficult to catch because they weren't deprecations or things that caused errors, but changes in how valid syntax is interpreted. I don't see that as a long-term problem, but it gave me pause.

I guess the TLDR is: if you're interested, I recommend you dip your toes in it, if that works for you, but with some caution. I see it more as a replacement for R long-term than Python, and I see serious competitors rising in popularity.

FWIW, I'm not sure why infidelity is so stigmatized. With all the things going on in the world that seems so minor in the grand scheme of things. Also, presumably the secretary is an adult, so they're two consenting adults (yes, I understand the power differential issue but there's also the risk of being accused of sexual harassment, which I'd argue is more damaging nowadays). Relationships are so complicated I've learned not to pass judgment on anything.

I'm definitely not saying infidelity is good, only that I think it's treated way out of proportion.

To be honest, I think it sounds more shitty to unilaterally declare respect or lack thereof for a colleague based on their private life rather than their professional performance.

Take credit for work that's not yours, and leave with millions in severance after they've fucked up the company? Totally fine. Sleep with your secretary? Unforgivable (and the CEO's fault).

I've published in mathematical/statistical areas as well as applied ones, and my sense is that interpersonal difficulties are more acute, but also more transparent, in math and statistics, for lack of a better way of putting it. My sense is that people tend to hold views more absolutely, because they see their positions as more logic-based, and so have a more difficult time recognizing unrealistic assumptions they're making, or outright errors in their derivations or thinking. An applied researcher can shrug and dismiss something as sampling error, or ambiguity in a study design, but in math, it's harder to invoke such excuses or pressure relief valves. A threat, of whatever real or misperceived magnitude, seems to take on greater significance in math and statistics.

There's kind of "converse" problems too, that I think is the focus of the linked article, in that there's an epidemic of hype in academics, math, and the sciences. So it's not just that people of high ability are being missed, but some individuals of high ability are being over-hyped to an extreme. The signal to noise ratio is very poor due to multiple types of errors.

People have been writing about this for a long time; it's nothing new but little seems to change (https://www.theatlantic.com/magazine/archive/2002/12/our-gen...).

... this is like walking through a minefield, but here it goes.

A lot of the illnesses they're talking about in the article, like IBS, lower back pain, and CFS have a strong psychosomatic component to it for a lot of people. They essentially say this in the article--something like "this won't work for malaria, but it will work for pain."

This can be a really controversial position, and it gets really twisted by a lot of people. No I'm not advocating mind-body dualism. What I'm saying is that there's top-down influences on physical symptom perception and that probably influences bottom-up processes, leading to some sort of vicious circle.

I think a lot of these psychosomatic conditions are really explained by the same processes involved in the nocebo effect. There have been studies to support this, showing that psychosomatic patients report more nocebo effects in control conditions than other patients and controls.

The idea you're discussing--reassurance effects--is interesting and actually really understudied I think. However, I think there's an equally plausible hypothesis, that you're kind of fighting negative psychosomatic effects with positive psychosomatic effects in certain situations.

I don't think that's all of the placebo effect, but I do think there's some broader causal system involved that encompasses placebo effects, nocebo effects, and psychosomatic illnesses and effects, including more psychosomatic psychiatric conditions (e.g., conversion disorder, psychogenic seizures, etc.), as well as psychological effects on disease process (e.g., stress effects on inflammation).

I kind of wish this area of research didn't get so bogged down with political infighting, which invariably happens. Patients start insisting that we're being dismissive by discussing psychosomatic processes, some researchers will spuriously start criticizing other researchers for advocating Cartesian dualism, as political smear tactic, when the real issues pertain to emergent processes and top-down mutual influences. It's a very interesting area of research with lots of potential, but tends to get oversimplified really quickly to score political points.

Legislators who stand up to nonsense like this need to be supported for protecting children against an overreaching state.

The law is absurd in what it targets, but even if children were actually guilty of a crime (as opposed to just engaging in sexual behavior), they should be allowed to mature.

This is just another example of the GOP outlawing sex. Society won't accept their archaic social morals, so they shove it down our throats under other guises.

I had a similar reaction. The headlines might as well been "extremely skilled climber makes wildly risky bet and lives to tell the tale."

Sure he's skilled; sure he is strategic. What I want to know is could he do this over and over again without dying? My guess is not. His odds would probably be higher than most others' but still risky.

But in the end, if it's what he wants to do, so be it. To me it's interesting but not something I'd like to do, regardless of the risk. Spending my time climbing rock walls is just not appealing to me, even if it's out in the wilderness, which I love.

Survivorship bias needs to be taught in schools at a young age, like in kindergarten. It's so insidious.

On the other hand, I didn't think of it until I read this. It made me a little satisfied, and then it made me quite angry.

One of the things I wondered when I read this story is if it would be possible to develop software that would somehow circumvent this type of situation. For example, using autoencoding or something to lose watermarking details intentionally or something. But what you're discussing seems more advanced than the yellow dots idea.

I was surprised to see printers being involved--I thought something like this leak would all be digital. I was also surprised that the Intercept would not be more savvy about printer identification because it's been publicized so much over the years.

But this isn't about the NSA, it's about state-level secretaries of state offices.

That's similar to my take here. Whether or not Trump actively colluded with Russia to affect election results, he's still behaving abnormally about the issue. I mean, if you step back and look at this, and just give Trump the benefit of the doubt and allow that maybe he was ignorant about Russia's attacks on the US elections, is his response reasonable for a president?

He should be vigorously defending the US, and aggressively trying to get to the bottom of it. Instead, he's actively obstructing the investigation and thwarting US defenses. In short, he's doing the opposite of what he should be doing as president. He's acting like a damn mole using the office to further his own interests and those of his business associates, including Russia.

At worst Trump is a traitor, and at best he's grossly derelict of duty. How this is not front and center of US political discussion on all sides, including the GOP, befuddles me.

This also seems critical to interpreting Trump's voter fraud commission, as the Russia efforts were focused on voter registration. I predict Trump will somehow spin what was being reported on in this article into something having to do with voter fraud. That is, "let's not talk about Russian hacking of voter registration records, as that's just speculation. We agree, though, that there are anomalies in the voter registration records, and given that we all agree on that, it's time for voter ID laws."

A monolithic private institution doesn't inspire confidence in me any more than a monolithic public one. I don't see the purpose in switching one for the other, and would probably rather see it run by the government.

If it was replacing the current system with one where you had 12 different companies competing with one another, maybe so but I don't really see the benefit in what I'm reading. But I don't feel strongly about it.

What worries me the most about this in some ways is how it seems to be focused on the large airlines and not on passengers and private pilots. This is the current trend that disturbs me the most about Trump and the GOP: they seem pretty blatantly to be representing large corporations and their own party rather than the public at large. The assumption seems to be that what benefits the fat cats is what benefits everyone.

I'm not anti-large-corporation per se, but a big part of the reason for democracy is to protect the average person.

I don't really see any arguments here other than "it makes the airlines happy." As if they needed more latitude.

It's hard to tell from either article what's going on. The Science piece you linked provides a bit more detail, but I wouldn't say those details are particularly damning (about half of the audited healthy participants didn't have an approved physical as part of the study, but they did have physicals elsewhere at the institution because they were participating in other studies. The most concerning thing is that 2 women didn't have pregnancy test results confirming they weren't pregnant.

Sure, you have to be careful about scientists maintaining ethical practice, but the other side of this is a trend toward research IRBs being focused on legal technicalities in reviewing research. For the most part, whether you're talking about NIH or universities, IRBs tend to focus on legal liability rather than ethics. This sounds like it might be reasonable, but in practice what it means is that IRBs aren't asking "is something really problematic happening here?" but instead "is there anything, even the smallest irrelevant detail, that could be used against us in a legal case?" Usually this means that IRBs are representing the institutions' best interests, rather than those of either the researchers or participants.

A related thing came up with social scientists recently, where a bunch of them apparently said "screw this, we've had enough"(https://www.nytimes.com/2017/05/22/science/social-science-re...). The issue wasn't that they don't care about ethics, it's that they are fed up with IRBs treating the risks of asking people about their political attitudes as if they are equal to having a surgical procedure done.

It sounds like there was some drift from protocol, which should be scrutinized. But nothing I've read so far suggests that the harm incurred justifies a ban on the use of the research data already collected. Maybe have some external review committee review the situation and make recommendations for changes; maybe indefinitely postpone future data collection by certain personnel, until things can be ironed out. But to ban publication of anything based on existing data?

To me this story illustrates two problems with academic research today: the unreasonableness of IRBs, and the corruption of grant politics.

Your comment about business administration threw me for a loop--I couldn't tell if you were being sarcastic or not anymore.

These sorts of articles always strike me as naive, and the discussions sort of disingenuous.

I'll go along with this thought experiment when I see places stop requiring MBAs of administrators just 'cause (one of the most useless degrees in my opinion), when medical professionals don't need a license to practice, and when companies actually stop looking at degrees when hiring individuals, instead of the exceptions to the rule you always see on the internet ("hey, I hire for my company, and I don't pay attention to degrees. Great. You're like 1/100).

Put aside all the unrealistic assumptions being made, that others are posting, such as that money will just fall from the sky to invest, or that 7% annualized return is actually a value that people can count on.

It's still ignoring questions like: will I be able to do what I want just teaching myself? Could I get hired as a civil engineer? For every one of these articles, there's another talking about how comp sci graduates are just better prepared for software development, etc. on average than people from other majors who are self-taught.

It's funny to me that these statistics and discussions ignore grad school, when a prereq for grad school is going to undergrad.

Believe me, I wish everyone would go along with what someone is capable of rather than what their degree rubber stamps (that is, after all, for all the jokes made, the premise of a liberal arts degree--that you don't need to specialize in school to have any given skill in real life). It would be better in so many ways. But society is more superficial than we let on.

Blockchain-based contracts seem like a useful thing in the abstract, but I've always had two questions that no one has really been able to satisfactorily answer. These might have been answered, though, because it's been awhile since I looked into it:

1. What about blockchain length? The article kind of alludes to this, but there seems to be this "we'll deal with that problem later" idea, even though it seems critical. The answer I always got that chains would fork or be stored distributively but then that suggested the primary use would be in small networks, or that there would be critical problems to solve sooner rather than later.

2. Isn't a guaranteed decrease in monetary supply a problem? I was kind of under the impression that ideally a currency experiences a small amount of increase monetary supply, to avoid things becoming prohibitively expensive. The process of generating coin seems kind of backward to me in many ways, although I'm not an expert in the area.

Not to absolve Purdue's responsibility in this, but I do think this exposes the fraud that is the US healthcare system.

First and foremost, it illustrates that drug regulation is basically a failure, and that the FDA and physicians cannot be trusted as sole authorities on drug safety. There needs to be more diversity of authority.

You say "unless doctors ran their own studies" as if it is not the responsibility of physicians to be thinking critically about scientific information about drugs, in terms of who is providing the information and how to respond as scientists. Your statement basically assumes that doctors are not competent to evaluate and conduct scientific research. Why shouldn't doctors run their own studies? Maybe not each physician, by why don't they as a group? Why doesn't a large hospital run a trial? A conglomerate? It's not that difficult given the magnitude of the consequences involved and the frequency of prescription.

It is the job of the medical community to be proactive participants in scientific process, especially when clinical consequences are on the line. That we've reached a point where we lay the responsibility of critical evaluation at the lap of the corporation says as much about the fault of the physicians and FDA as it does about the corporations.

It's kind of like saying that an individual who knows a child is being abused but goes out of their way to transport the kid day after day to the abuser without reporting it bears no responsibility because they weren't abusing the child themselves.

Anyone involved in the healthcare system knew that oxy was an addictive substance and there were rampant problems. The idea that someone with an MD shouldn't have a clue that increasing the dose would lead to greater problems is absurd--it's like claiming that a chemical engineer shouldn't understand that adding jet fuel to a fire wouldn't increase the blaze. Acting like physicians were innocent victims is a joke.

The corporations and physicians were both complicit. Either that, or they were incompetent. It can't be neither. The FDA and physician groups need to be dethroned from their perches.