logjammin

"You better agree with consensus" has always been the current culture in science. It's nothing to do with science itself -- just human emotionality. We don't change unless we have to, unless there's overwhelming evidence that we need to.

I think this is good advice in general -- you can't really ever go wrong with Noctua.

I built a Threadripper workstation last year but went with liquid cooling. However, I put 3 Noctua fans on the radiator and haven't looked back. Terrific company.

I think you mean Camden, NJ.

I came here to write exactly this about Louis. Hell, he was doing that long before he became a pariah.

I think that cries of "censorship!" from bien-pensant tech folks are largely overblown, and not only for the reasons cited in this thread. If some heterodox ideas of today truly will be part of future orthodoxies, they'll find a way - YouTube or not. A powerful concept that dies on the vine because it cannot get onto Facebook isn't a powerful concept.

I'd quibble slightly with the interpretation of the Hazard Ratio (or rate) here and say that it's more like "the instantaneous relative risk (or probability, if you like) of the event of interest [in this case mortality but not necessarily] between those with confirmed coronavirus and those without it, given that both groups have survived until time t and holding [set of covariates of interest, e.g. race, age, etc] constant."

Don't mean to sound like a pedant here even though it totally reads that way. I figured with enough epidemiology in the news these days, people may see a lot of "Hazard ratios" or "Cox regression" or "survival analysis" and be at risk of some confusion. I work with these concepts and I get tripped up myself sometimes.

Seconded (well, eighth'd, if the rest of the comments are any indication). I got a 6th gen X1C at the end of '18 and it's brilliant. It'd be neat to have a Ryzen option on them in the future, as I have a desktop I build with an AMD CPU and love it.

My only complaint about my X1? Windows.

You make good points, but you may be missing why OP's comment is getting the upvotes it's getting, or at least why I'd wager it is: he describes the kind of contentment that most of us can't even imagine possessing without winning the lottery, and he described having that contentment with surprisingly (to him/her) modest means. I can't speak for anyone else, but this is why I upvoted it. Hell, I can think of at least two times in my life where I've been scolded about being unproductive or ambition-less for even mentioning the possibility of living such a life (without being a millionaire).

Solely with respect to this person's comment, what we call such an arrangement matters a lot less than the high quality of the life he/she describes.

Came here to rave about McElreath. The book is great, as are the attached materials and YouTube lectures. Thumbs way up, especially when Gelman is giving you trouble.

Eh it's just as well, NZ could use an environmental breather from all the tourism. Silver linings.

It really, truly cannot.

This is really helpful, thank you. What you say puts some form and empirical evidence to my concerns about Hackintosh. I guess my gut had it right this time, which is unusual. So you can brick your fancy new homemade, warranty-less PC!

Anyway maybe one day I'll do it for fun on a crappy laptop I get off Craigslist. Sounds like this pays off most when it's a low-risk effort.

Great question. The answer may boil down to the way my memory works. I work in research, so I'm constantly reading and citing papers and studies for lit reviews, general understanding, making sure no one else has done the project I just thought up in the shower (usually they have), and mainly just staying at the crest of the wave in my field and subfields.

With the exception of the big famous names (famous for the 13 of us in our niche, anyway), I rarely remember those studies' authors, nor the titles of the papers most of the time - i.e. the data encoded in the filename. But I do remember certain phrases, numbers, and the like that they use in the body of the paper - in other words, the material that actually interests me. File content search for PDFs and other text allows me to enter one of these snippets and find the paper in question without having to spend minutes upon minutes scratching my head about "who was that lady at NYU ... or was it a guy at Berkeley"?

Oh yeah for sure, I found everything pretty quickly and it's definitely indispensable. I'd love if it did content search, but file search is great and fast and straightforward.

I didn't know about that upcoming feature. I'm really excited to hear that, so thanks. The PDF thing is a wild oversight, too, but I found SumatraPDF and am really impressed by its light weight and speed and bare-bones ethos. There's barely a UI but man can it handle 53 open PDFs at once.

I have both WSL and dual boot, but I dual boot way more. This is likely because I am a command line idiot, and never got fully fluent with navigating my computing life using one. It's high on my list of skills to master in life, because I know what a force multiplier it can be, but I haven't gotten around to it. I mostly use Linux to write, weirdly enough: from org mode to LaTeX to statistics coding, it's a smooth experience - again because it's pretty "just the basics" and does them well. I have no actual need for Linux - I'd be fine with just windows. I guess it's more aspirational on my part.

Apple's moves: yeah man, they're taking that walled garden shit to serious extremes. I know there's a logic to it that works for them, so I don't begrudge them their choices necessarily. But I did really like the company for a long time, so it's a bummer to see them the way they are now.

Yeah that's a great point. I thought about it initially, but I got a little nervous (perhaps unjustifiably) about the Hackintosh universe being a little slapdash and then about Apple maybe issuing some under the radar OS update that bricks my machine. I'm catastrophizing, maybe, but I never pulled the trigger. That said, I never actually put in the 3-4hrs of reading I'd need to do, so I definitely wouldn't rule it out.

Do you have experience with doing the Hackintosh thing at all?

"I swear, I wish I didn't love macOS so much (or wasn't so heavily invested in it), or I'd happily ditch it for a really powerful thermally cooled desktop and use that as my machine."

I was you at this time about two years ago. Exact same thing. Fed up with Apple hardware bullshit, and with their pricing. With Apple in general. I had a sick iMac (as sick as an iMac can be, I mean). Loved and was invested in OSX. But something tipped the scales. Can't remember what specifically, but I said "fuck it" and put the machine on Craigslist. Got a buyer immediately, and lost very little money on a three year old machine. Took the cash, plus a little more, and built a PC. A liquid cooled PC. An egregious, so-ugly-it's-kinda-neat monster with tubes and a radiator and fans and the whole deal. I run Windows 10 Pro and Ubuntu. Neither is perfect for me. Windows especially can be maddening. My personal pet peeve is the lack of powerful device search -- in OSX, I could use Spotlight or better yet Alfred to look inside PDFs, for example. SOL on that in W10.

But it was worth it. My AMD-powered PC smokes anything in my Zip code, I'm pretty sure, and it's more than enough for my work needs (and my work does actually put the thing through its paces.)

I miss OSX, but not enough to go back.

This is a problem of insufficient data. We don't have a denominator, so there's no way to tell if it's more widespread than we thought. In fact there's no good way to tell how widespread it is at all. We have numerators: number of people sick enough to show to up to an ER, or number dead, or whatever, take your pick. But in every fraction we might want, we lack a denominator: total number of people infected.

There's a stuningly simple way to get one: sample randomly from a population of interest in large enough numbers (say e.g. Cook County, home of Chicago, or hell, the entire US) and test everyone for the virus. Voila! You have a representative chunk of your population, an estimate of virus prevalence, and a quantifiable degree of precision about that estimate (SE's for a confidence interval, say). With enough people in your sample, you can have extremely accurate estimates of the spread of Covid in your population of interest. You repeat this process over and over, through time, to track the numbers.

I'm not trivializing this kind of effort: it takes rigorous sampling designs and dozens or hundreds of field workers, among other things. It's intense but very straightforward. Political pollsters (and thousands of researchers in different fields) do it every day.

No one in the US has done this, and no one with any visibility from Fauci on down has even suggested it (that I know of; please correct me). I work in public health, and this is first year, first semester of grad school stuff.

This matters because the public health responses to coronavirus when there's 1% or 10% or 25% or 75% of a population infected all look very different. In short, we may be over- or underreacting to the situation with the measures we currently have in place.

I've been going half crazy wondering what's happening about this for a while now. After the initial fuckup, we now have tests. We have money. We can do this. Why don't we?

(Iceland has done this!)

I've been using Wire [1] for years on desktop in large part because of their E2E claims; I wanted something simple and secure that worked relatively fluidly for video calls. I've had mostly pleasant experiences with it, and with many calls I'm surprised by the video quality. I've never tried their mobile app, but they've got one and it looks nice, aesthetically. Main drawback has really been that few people I know use it and I've had to cajole people into doing so a little bit, which stinks.

But I'm not a cryptographer and am unable to verify the company's security claims. For all I know it's go zero encryption whatsoever. Are there people on HN who feel qualified to comment? Has anyone used Wire before?

[1] https://wire.com/en/

Hey, truly thanks for this. It's not Dark Sky, but it's not bad, and I never would've heard about it were it not for folks on HN. I'm getting good data for where I live, the UI is nice and clean. It's a start!

Just came here to say "fuck". I'm so let down. Can't have nice things, I guess.

Right, yes, exactly: my point is that that state of affairs you describe is fucking bananas. That's all.

Put aside for a second the fact that being rich enough from word processing software to surround yourself with experts and repeat what they tell you does not (necessarily) make you an expert. Apart from that, I agree with the thrust of what you're saying: he's not the Waltons, and he's funded meaningful public health work.

The point I'm trying to make in my original comment isn't about BG himself, it's exactly about "the people in charge of our institutions". Rereading it, I didn't make this as clear as possible, so let me take another stab:

We think of ourselves as an advanced democracy (or democratic republic) and economy. We certainly pay for that in taxes and in other ways (obviously we're not Nordic in terms of income tax, but still). But when it's go-time for that advanced democracy, for the institutions and experts we pay to train through years of schooling in public universities and then pay to command highly specialized bodies at local, state, and federal levels, their performance is, excuse my language, so fucking dismal that the vaunted CDC cannot make a working test, that there are not enough latex gloves in that advanced democracy to keep doctors and nurses from getting sick ... their performance is so dismal and we're so untrusting of the basic decency and competence of our experts and elected officials that we crowd into a Reddit AMA to hear a software magnate maybe have a better solution ... Is that not just bizarre to you? Infuriating? Sad?

I fully grant that I may be alone in feeling this way. On some level I'm reacting strongly to the fact that the country I immigrated to as a teenager is revealing itself to be profoundly third-world in certain ways - like, this isn't what I thought I signed up for. Like I said earlier, there's something truly funny about it at the same time that it's terribly sad. (One funny thing is that the solution the guy comes up with is "let's put a microchip in everyone!".) Don't mean to impose my views on anyone, and again, I'm not saying Gates has enabled, via Windows wealth, zero good in the world. It's not about Gates; it's about us.

If he's so prescient about pandemics, why didn't he spend the last 20 years, I don't know, making sure healthcare workers in his own darned city had enough masks and gloves to deal with one? Surely he could've saved a few lives in his own backyard while we was busy in Africa, right?

Excuse a bit of hyperbole, but what sort of medieval nightmare-world are we living in when the citizens of the world's most revered democracy are waiting with bated breath for the pronouncements of the wealthy on fighting a viral pandemic?

Gates Foundation, research, etc etc. I get it. He's been involved with health stuff for a while. In theory, too, I'm open to a good idea regardless of its source. But this remains a guy who dollars to doughnuts I'd bet does not know how much milk costs at his supermarket. And he's telling us about chipping and digital tattoos?

It'd be legit pretty hilarious if it weren't so insane.

Who doesn't have skin in the game in that sentence?

Calling for repeated random draws of a country's population and testing everyone in that draw is stirring shit up?

While I'm not defending the entirety of Ioannidis's argument (he makes a couple of mistakes; e.g. we know social distancing has some effect), it's important to point out what his central argument actually is, and how it's different from how the CBC article that purports to describe it. He doesn't assert that the policy response has gone too far; instead, he asserts that in the absence of good data on the pandemic, it's difficult to know whether the policy response is too much or not enough. In my view, on this point he's exactly right.

From the actual Stat article:

"At a time when everyone needs better information, from disease modelers and governments to people quarantined or just social distancing, we lack reliable evidence on how many people have been infected with SARS-CoV-2 or who continue to become infected. Better information is needed to guide decisions and actions of monumental significance and to monitor their impact."

On "we lack reliable evidence": his point is that we don't have what we need, namely an accurate denominator in calculations of morbidity and mortality. "We don’t know if we are failing to capture infections by a factor of three or 300." He's not wrong! In epidemiology, accurately estimating, say, case fatality rates is notoriously difficult - especially in the middle (or the beginning!) of an epidemic. Given the issues we're all familiar with surrounding the lack of testing, this should come as no surprise.

On "better information is needed": I don't think it's that far off-base to say that in the absence of (what asserts is) good data, it's impossible to calibrate policy responses to 2019-nCov so that you're saving as many lives as possible while doing as little damage to the rest of society (e.g. economy) as possible at the same time. This suggestion is eminently reasonable: "The most valuable piece of information for answering [questions about how many people will die of the coronavirus] would be to know the current prevalence of the infection in a random sample of a population and to repeat this exercise at regular time intervals to estimate the incidence of new infections.".

Ioannidis (a physician and an epidemiologist) has made a career out of shedding light on science's mistakes and blind spots, and rightly so. I think the controversy surrounding this piece is overblown: the "fiasco" he's talking about is not a policy response fiasco, but an "evidence fiasco" - without knowing (or, in the US, even attempting to know) the true prevalence and incidence of coronavirus, it's impossible to say how well we're responding.

* Edit / addendum:

Counterpoints to Ioannidis to the effect of "well this is all moot anyway, when faced with high uncertainty about a really scary thing, you act as cautiously as possible regardless of the quality of your data" have a lot of merit. If you're a policymaker right now and you see what's happening in Italian ERs, hell yes you implement social distancing on a mass scale, maybe full lockdown / shelter in place, etc.

But these two sets of tools aren't mutually exclusive; we have a lot of smart and capable people, plus a lot of money, to sic on each. We can allow policymakers to react strongly and cautiously right now to maybe save lives, while also setting up the sorts of basic epidemiological studies Ioannidis calls for: repeated random samples of a representative chunk of a population (e.g. the US population) with the ability to test everyone for corona (preferably using multiple different tests) in order to accurately estimate the denominator.

Source?

Not sure why this is greyed out. I might not agree with all of it, but you're not out of line.

That's a good question, no doubt. A related one I had was (and maybe I missed this), but what's the "dose" of virus one needs to get infected? In the discussion of decay / half life of the virus on surfaces and in the air, I wondered about this. It's one thing if the virus can hang around for three hours and infect you at any time, and another thing if it's in the air for three hours but decays at ____ rate and can only infect for 1 hour.