braindongle

"...They are an acceptable way to represent a data definition, but a poor way to write queries and a bad way to store object state." Rails perspective: Pump your brakes there, friend. This claim depends completely on your application's needs. I write web apps for research that don't need to scale and don't need lighting performance. I have built apps that ended up needing some optimized-SQL-versus-ActiveRecord experimentation, and sometimes, writing raw queries is certainly worth it. AR has been evolving for 15 years, is extremely useful, and all that is needed for many use cases. We're not all trying to do the same thing.

Thank you for the zero-indexed reference list. Brought a smile.

If you're unhappy with NYT, consider their mainstream-educated-conservative counterpart, WSJ. NYT had a porous paywall for a long time, intentionally. WSJ has always been on lock-down. In both cases, it's a strategy, not do-goodery, but regardless NYT is no bad actor in this area.

Nicely done. I wonder: what is the use case that prompted this investigation? 40-50 milliseconds is correct in terms of latency that musicians notice. If you're aiming at any realtime musical collaboration over the internet, the Big Problem is network latency. Unpredictable and just too long. And, if you're not working over the internet, why use a browser?

Edit: https://hello-magenta.glitch.me seems to be the use case. Cool and inevitable that this would be worked-on, but after spending years with both algorithmically-supported composition and all-human composition, I'm skeptical. There is a special sauce that machines will never grok. Or, I'm wrong.

I'm coming from a web development perspective. When the browser is your gui, or your app doesn't need one, Docker (running on Linux, Mac, or, cough, Windows) just takes away your concerns about environment consistency. Like a VM, but not really. A "container", a minimal Linux image, running a "service" for your app, is a beautiful and portable thing! (Looks like running guis on Mac from Docker is messy: https://cntnr.io/running-guis-with-docker-on-mac-os-x-a14df6...)

Docker is the way. Darwin is a mess for development, but it doesn't matter. Start with a 137 MB bare-bones Debian image, add just what you need, Docker is the way.

Wet blanket researcher: No randomization, purposive sampling, high attrition, this is not great science. Doing randomized trials for this sort of thing is difficult but necessary. A correlational pre-post study is a good start!

This is the most entertaining (false) stereotype that I've seen in some time. Those artists with their big egos and weak constitutions! Makes me nostalgic.

“Artists have very fine and delicate natures, they have a very delicate nervous system, and, unfortunately, all of them have a strongly inflated self-image,”

Woo-woo! Engineering word police, open up! "We dampen sponges. We damp vibrations."

Riders can certainly be over-concerned about weight, but uh, it is pretty important. How much to these expensive beasts weigh?

I'm interested to see what the future of the Research Scientist role is. It's basically a place for people who didn't do a post-doc and aren't bad-ass enough to not need one. And, one could argue, people who just don't want to play the faculty game. I hope we see that space grow in academia. I think we will.

...also, in a ponzi scheme there is no product. I aim to produce useful, generalizable knowledge. That's what the NIH is paying for. A valuable product.

The academic ponzi scheme thing is overplayed. I'm a Phd Research Scientist at a good school. I'm not faculty, but I have a lot of freedom and I get to do research that is fulfilling. Yes, it's a pyramid (aren't all org charts?) and I'm not at the top, but I'm happy.

Those of us in the field of psychiatric research encounter innumerable illness narratives of this sort. Everyone has their own story to tell about what went wrong and what helps.

While these anecdotes may be of value to the storyteller, they are at best worthless if you are serious about understanding your options in mental health. Evidence-based psychiatry is messy, but good people are working hard to build credible evidence, and not just for pills. "Everything is biased by pharma money" is often said by people on the sidelines. Psychiatric services researchers know better.

Interested in meditation vs medication for depression? Start with a systematic review in a reputable journal, like this: https://goo.gl/yN1asm

There's the convenience of these tools, then there's convenience of having a small number of non-standard tools to carry along with you, however easy that may be.

For example, for any diff'ing or cat-like need, I find it SUPER handy to just pipe stdout to vim like so:

$> thing-with-output-that-needs-navigating | vim -

I'm not an Emacs user, but my impression is that over in that world, the shell integrations are crazy good.

Wait. Incentive structures in academia are screwed up? No. It can't be.

I see similarities here with pharma and medical research. Conflicts of interest are endemic there, and they will be here, too, but this sort of sponsorship is inevitable. What can be done? In the case of pharma, I'm OK with them sponsoring peer reviewed, academic research, as long as policies are in place to make it immediately obvious where the money is coming from when I read the literature.

What to buy instead? My 2015 MBP has been a wonderful machine to work on, and there's no way I'm getting one of these little gray turds. So, I'll be using Ubuntu for my daily driver, which is fine. But that last generation MBP is so nice, what other hardware compares?

Neighbor to neighbor, not medical advice:

Navigating the mental health care system in this country is difficult. From what you've described, here's what I would do:

- Start with an in-network therapist. My personal bias is to look for someone with a PhD in clinical psychology, but the important thing is to get started. Talk with your therapist about medication. They are trained to do this and to get you do an MD (psychiatrist) if you agree it makes sense. They will take it from there. Managing meds is basically what they do. (There are psychiatrists who also do talk therapy. If they are good at both, it won't be cheap).

- Do not feel at all obligated to stick with your first provider. This part is hard, due to the awkwardness of firing your therapist. Therapy takes time, but it may be clear after a couple of sessions that you are, e.g., with someone who won't talk about solving practical problems and only wants to talk about your upbringing, or the other way around. If you think it might be a good fit, give it more time, but it's true: you do have to click with your therapist for it to really work, and there is no formula for that.

- If you are not a danger to yourself or anyone else, and you are able to take care of yourself (basic social functioning and providing for yourself), then consider thinking about your care in terms of cost and benefit. If you are having to make excuses to duck out of work, that's stressing you out, and the therapy isn't helping, that's a bad c/b. If it's costing you an arm and a leg, your mind is more important, but that's still a bad c/b.

If you are isolated and in constant pain, strongly consider staying in some sort of care as a safety net, even if you can't find someone with whom you really click.

As a person who has experience with severe depression, studying evidence-based psychiatry, and conducting psychiatric services research, I wholeheartedly endorse the advice in this piece.

I disagree. If you Google for primary literature related to Tuskegee, you will find, front-and-center, publications describing the ethical disaster that it was. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2609060/ https://dash.harvard.edu/bitstream/handle/1/3372911/brandt_r... One does indeed grasp the horror.

And, where did the Nazis publish, again?

We certainly need people to responsibly interpret peer-reviewed literature for lay audiences. Responsibly.

> Facebook has acted on this and intentionally made hundreds of thousands of people fall into a depression, just to see if they could, and then they bragged about it.

No, they did not. They adjusted the proportion of positive emotional expressions and negative emotional expressions in users' news feeds in order to test for an association between that proportion and emotional expressions in users' posts. There are serious problems with what they did, but "intentionally made hundreds of thousands of people fall into a depression" is late-Slashdot style trolling.

http://www.pnas.org/content/pnas/111/24/8788.full.pdf

What specific user-affecting changes would the new overlords make that suck for developers? I don't doubt that there are good answers to this. They could treat it like a cash cow and just start charging more for less, but I mean functional changes.

All good points. The most interesting thing about Reddit to me (again, the tiny slice of it that I see, I always come in through the front door) is civility in the presence of anonymity, in 2018. "Come for the cats, stay for the empathy" is pretty right-on.

IMO, groupthink is worse here than on Reddit, or at least the subs I visit. Divergent thinking is often celebrated. Here it's more about convergence to the most defensibly informed perspective. I may be wrong, and these differences are subtle in any event. Regardless, any discussion of groupthink on HN should include some critical self-reflection. Please downvote me!

Is there a prevailing approach to thinking about (accounting for?) false negatives in ground truth data? I'm new to this area, and the question is relevant for my current work. By definition, you simply don't know anything about false negatives unless you have some estimate of specificity in addition to your labeled data, but can anything be done?

It's a bad sign when a self-proclaimed authority lifts the definition of the topic from Wikipedia. Unless said authority wrote the Wikipedia entry, in which case I humbly apologize.

I for one am fine with the idea that 'phone' means something different now. That's how important smart phones are. They changed the meaning of 'phone' in under a decade. I disagree that this skews the conversation about them. People know how often they use the "phone" app. Nobody thinks "phones taking over our lives" means people are using them for voice communication too much.

Absolutely correct. And so-called general AI may never happen. Regardless, this is shocking. It immediately needs to be factored into any speculation about what the world will look like in 20 years. Innumerable questions.

This is an excellent list. I actually fired off a snarky reply to a snarky work email just before reading this. Thanks for reminding me to check myself.