What you need are citations to trustworthy sources and to be reviewed by trustworthy peers.
You've already lost the fight: no human system outperforms its incentive structure.
Peer review is hopelessly ensnared by academic advancement culture. Entire disciplines can end up publishing bunk, if that becomes the tenure-track fashion of the decade. Tulip bubbles are not restricted to the business cycle. Even hard sciences have been hit pretty bad. Et tu, string theory?
The fundamental theorem of peer review is due to Max Planck:
Science advances one funeral at a time.
The zone of convergence of peer review involves the passing of interested parties. In most of the hard sciences, fifty years pretty much weeds out the crap.
However, if you take a field such as nutrition science, I dare say it's still inadvisable to take fresh "peer review" at face value. John Yudkin was on the right track in 1958. Fifty years downstream, the truth is out there, but it's still far from evenly distributed in the public imagination.
Nutrition science was subverted by a white coat army of industry apparatchiks. These studies are expensive and, oh yeah, replication crisis.
Most human systems can be trusted some of the time. The real art of bullshit detection is figuring which times are those times. Even the best human systems are bullshit on the margin.
What you need to understand here is that the journalist impulse to publish is directly proportional to the tenuousness of the result in question.
Well, if the speed of light falls derp derp wormholes derp derp Stargate derp derp dusty von Daniken booster spice derp derp human immorality derp derp Omni Magazine alternate-reality cum shot. Well, you got your $4 worth, didn't you?
There's an enormous term in the human condition centered around escape from reality. This makes sense to some degree, because human reality usually contains a giant heal spur of oppression of the downtrodden masses (success has a habit of being highly asymmetric).
Trump's monosyllabic barrage becomes tremendously more convincing if you want to believe the underlying message.
Somehow, one supposes, being suckers for false hope must be evolutionarily adaptive (who, after all, is qualified to challenge the modern evolutionary synthesis?)
And then you get right down to it, the anchor tenants of modern bullshit culture are the major religions (being largely incompatible, at most 1 of N could anywhere close to broadly correct). Because, you know, life without bullshit would be empty and meaningless.
Deep down, most of us don't really want to drain the bullshit pond. And it's not just one pond. It's pond after pond. Never get comfortable.
The fundamental theorem of bullshit busting is due to Richard Feynman:
The first principle is that you must not fool yourself and you are the easiest person to fool.
Evolution took a long look at Hamlet, and came up with satisficing.
Make happy assumptions that are compatible with medium-term survival (generally best obtained from proven survivors—aka your parents and select community), then behave with the efficiency of assuming their truth, until the shit really hits the fan; then sit back, renounce, regroup, and repeat.
Dawkins pretty much feels about religion the way Einstein felt about cosmic inflation and quantum indeterminacy. Right model, wrong hope, long painful row to hoe. Even when our best minds get something right, they're often left wishing they hadn't.
So there's this unhappy observation about the human
The article has a poor to false understanding of how blue light interacts with DLMO (dim light melatonin onset).
I'm pretty sure the entrainment effect of blue light is via direct neuronal connection to the SCN, and I doubt it involves melatonin, except indirectly.
The homeostatic sleep pressure signal builds up (more or less linearly) for as long as you're awake. On its own, this would mean that you taper into drowsiness all day long. So the sleep system has another mechanism that suppresses response to the sleep pressure signal. I vaguely recall that what happens with DLMO is that melatonin onset signals the body to turn off the suppression switch, so that the body begins to notice the homeostatic sleep pressure signal.
DLMO, however, is easily inhibited by exposure to blue light at a point in time approximately an hour before bedtime. If you're outdoors hunting moose in the bright light of late-evening arctic summer, this is a useful adaptation.
You'll get to bed later, which means you'll sleep a bit later (but not much) and then you will get less blue light early the next morning, which will affect your entrainment, gradually, on the slow-drip program.
—
As a rough, empirical ratio, for every extra hour you stay up, you'll sleep about twenty minutes later the next morning. It's not uncommon to stay up for an extra two hours, then barely sleep in for an extra half hour. (We need to ignore here that modern society tends to run a massive, permanent sleep deficit, which can suddenly turn into sleeping four to six hours late at the first opportunity that allows this to happen. That's a different beast entirely.)
I have a circadian rhythm disorder, and I know from decades of sleep tracking that morning wake-up time is about three times more reliable in estimating my sleep phase than time of retirement.
—
This is a worthwhile paper from the top of my notes, but it's hard to wade through:
I just love the page break at the end of page 6. But then I'm really into microscopic moments of small page-formatting humour. (It's probably not unrelated to all those long, lonely nights, before I found a viable treatment.)
—
Here's a good summary, I just found for the first time.
The reason I only vaguely remember this mechanism is that all the phase response curves in the literature are dose dependent.
There is no PRC I've ever seen that computes the phase response differential to endogenous melatonin levels. No, what you do is administer some dose/formulation (which can include sustained-release components) at staggered times over several weeks, and then you plot the graph averaged over your test population (which thus includes all the metabolic uptake and clearance variability).
There was a time I desperately wanted to consult one of these curves and then to declare "I am here", but it never happened. These are, in effect, better regarded as qualitative curves than quantitative curves.
The model was never predictive enough to be worth memorizing exactly. And thus I remain slightly dim on DLMO when I really shouldn't be after all these years.
At the age of 22, I fell in love with my boss, and at the age of 24, I learned the devastating consequences.
The Republican weaponization of Clinton's misdeed was to claim that this behaviour made Bill unfit to govern. (If powerful men having extramarital affairs with young women was incompatible with leadership, well, the vast sweep of history does not so record.)
Family values aside, the power imbalance creates the risk that Bill would abuse his immense power to cover up the vastly exaggerated blot on his record. The Republicans actually knew that anyone with an accurate base rate of human history / human culture would not regard his behaviour as incompatible with leadership—though a common and damning blot nevertheless, so the tactic was to escalate the stakes until Bill felt compelled to lie about it—which, unfortunately, was extremely easy to anticipate.
Lying to formal body of review is considered incompatible with leadership, sort of, incrementally, since not all that long ago. For example, it barely extends as far back as the Reagan's Iran–Contra affair. (Some people roll with family values and view Clinton's offense as the worse offense. I happen to roll with geopolitical transparency, and so I view Reagan's offense as the worse offense—he appointed those clucks, and it was his ultimate responsibility to know all the big shit).
Bill was plenty smart enough to figure out that the public perception battle would play out exactly as it did, leaving him boxed into a corner where he could—according to his established character—only choose to lie (perhaps he overestimated his power to blow off the investigation, but even there, had he succeeded, he would have mortgaged a sizeable fraction of his presidential energy in ruthlessly defending his momentary gratification).
Clearly, his judgment in this matter fell short of the mark by any standard.
However, I rate it not quite as bald as boasting about sexual harassment with a camera rolling. Whatever Bill purportedly said to Donald on the golf course (that was "far worse" in Donald's personal judgement), there was no film at eleven after the fact.
The modern world contains a lot of cameras and microphones. Trump's world has contained many cameras and microphones since way back. A prudent man in his position wouldn't be openly bragging about his magical power to get away with sexual harassment just to impress Billy Bush. And it's not like Donald didn't have a front row vantage point on Bill sinking his own boat through which to consider and amend his own standard of personal conduct. Donald had every opportunity to know better, and the penny never dropped.
So in summary, a whole lot of things are "not okay" but still the world largely spins as it has always done for thousands of years.
I know that's no guarantee but you do the best you can.
Considering how much you originally paid for your deoxygenated speaker wire, I would say $40 is the least you can do.
Were you to model the price signal with quaternion rotation instead, you would discover that the price signal really can spin around to a perfect 180-degree inversion of "the best you can do", given but a sparse free-energy input of mindless optimism, and a scant few months to capture abandoned area under the curve (and that's not even including the machine learning revolution).
Perhaps capitalism eventually does the right thing, but not until after imbibing all the loose sugar.
—
First Law of Mice and Men: If something can go wrong, it will go wrong.
Corollary: If something can lead to an easy buck, it will lead to an easy buck.
Unfortunately, all the quaternions in this picture belong to the increasingly neutered EPA.
So you've replaced a single CEO with the results of a vote between a few "senior staff". That leaves the rest of the company "not in charge".
Why stop your argument there? With another sentence or two, we can rewind western civilization all the way back to the Taliban's conception of marriage.
Because in any human collective, only one party can ultimately wear the pants.
How the truth stings. Resistance is futile. Sauron does not share power. Those poor, deluded Swedes. Yada yada yada.
Ultimately, this will affect almost no one. Planning for this change has been happening for a long time now. Your favorite add-ons will continue to work.
Trust, but keep one foot wedged in the emergency exit, and one ear cocked for a fell voice on the air:
It took me a year and a half of extensive rewriting to make my add-ons e10s/multiprocess compatible, something that is being rolled out only now, all with the prospect of a long-lasting life for them. And the WebExtensions announcement was made not two months after. "Demotivating" doesn't quite cover it...
Much to the point of this thread, the day is not long in coming where a series of small, hotly debated innovations in machine learning will culminate in a robust classifier able to ferret out a dozen egregious offenses against working brain cells in that toxic screed of hot-button click-bait you just posted.
The wise among us will use this forthcoming capability to accept all well-formed signals, the fools will filter bubble to black. The later group being larger than the former group, while the former group holds all the marbles, much derived from clue will change, while much derived from populism changes little, as our social algorithms contrive to sneer, in an arms race of eloquent dissection.
If Linus had ever worked on a truly hard problem, he might think different. Not every impasse in life can be resolved by industrious sleeve-rolling.
Operating systems don't hold opinions on human social dynamics. Earth-shattering innovation need not apply.
But props to MIT/Zhang for having a better understanding of patent law. That counts for a lot these days.
Your implication being that UC Berkeley doesn't see fit to make this caliber of legal advice available to faculty self-evidently working in fields with billions of dollars at stake?
If Berkeley fell short on sound legal acumen with the Holy Grail of the Biological Revolution inches away from the tips of their greedy little fingers, god help man with garage.
After pressing "submit", in a split-second second evaluation, I noticed that that sentence I wrote does not quite work.
Problematic:
Because some human process defined the solution gradient that the "genetic" software optimized over—ad infinite turtle—in an act of algorithmic emancipation now glibly lumped under the verb "write" by the baby-impervious membrane all-too-tragically-often comporting itself as "logic".
Less problematic:
Because some human process defined the solution gradient—ad infinite turtle—that the "genetic" software optimized over, in an act of algorithmic emancipation now glibly lumped under the verb "write" by the baby-impervious membrane all-too-tragically-often comporting itself as "logic".
Cognitively, this is a turtle too soon for full effect.
Even less problematic, but horrifying:
Because some human process defined the solution gradient that the "genetic" software optimized over—ad infinite turtle—optimized over in an act of algorithmic emancipation now glibly lumped under the verb "write" by the baby-impervious membrane all-too-tragically-often comporting itself as "logic".
Horrifying because I've always regarded this kind of sub-phrase repetition as the hallmark of hack speechwriters. One can partially excuse this by placing a semicolon in front of the repeated phrase, but here the semicolon is incompatible with the closing mdash.
Language is a complex solution gradient, one that humans have yet to successfully express. Sad. All those unemployed genetic algorithms, awaiting human clue.
You're not considering that machines can write algorithms. And they certainly can.
I've spent my entire life trying not to be this dim. Yes, very clever work there treading on the narrow definition—while engaged in 100% baby flush.
The ridiculousness of this is apparent to any thinking person in less time than it takes to type "Wittgenstein".
Because some human process defined the solution gradient that the "genetic" software optimized over—ad infinite turtle—in an act of algorithmic emancipation now glibly lumped under the verb "write" by the baby-impervious membrane all-too-tragically-often comporting itself as "logic".
For myself, negative ideation (which for me is not suicidal, just an endless litany of everything I've ever done wrong in front of another person) strongly correlates with sleep quality.
As soon as I reach the threshold where the negative thoughts are disruptive to daily life, I take of several effective sleep medications, and—if I manage to get the hard and deep sleep I need—negative ideation is gone again the next day.
Hours in the sack don't count, either. I've had times where I've been getting a solid eight hours of consolidated sleep for an entire week, but still the negative ideation has made an appearance. Sometimes my sleep seems to hollow out so that it's non-restorative on some hidden, inner dimension.
At present, my best sleep-quality aid is 3–4 mg of nortriptyline, a dosage I have custom-compounded at a local compounding pharmacy.
Its off-label uses include treatment of chronic fatigue syndrome, chronic pain and migraine, and labile affect in some neurological disorders.
Fewer and milder side effects occur with nortriptyline than tertiary tricyclic antidepressants such as imipramine and amitriptyline.
I was originally prescribed amitriptyline, which also worked great to restore my sleep quality, but left me fuzzy-headed the next day.
The later substitution with nortriptyline was based on my own research effort, as was carefully titrating my minimal therapeutic dose (which ends up costing me six times than as much as the minimum standard dose of 10 mg, available in capsule form only—what's less fuzz in the brain the morning after worth to you?)
If I really need to drop the sleep hammer, I take two, then write off part of the next morning to light housework.
Also, 84 percent of C-suite, and 81 percent of IT teams believe they have the right protection set up.
In related news, 85% of both groups combined think they are good at their jobs.
Interviewer: You get paid the big bucks. Are you doing it wrong?
Interviewee #1: Well, gosh, I don't know.
Interviewee #2: Every damn time, and twice for breakfast.
Interviewer: Uh, #2, how long have you held your current rank.
Interviewee #2: The previous numbnut is still fumbling for his keys in the parking lot, with all his executive possessions packed in an open box, tucked under his left arm.
Interviewer: How about you, #1?
Interviewee #1: Twenty-two years.
Interviewer: Really? You've been running the IT department for twenty years?
Interviewee #1: Actually, no. I'm the janitor. The chief custodian wears a shirt and tie, so I do, too. Always dress like the boss, you know. Good career advice passed down from my grandfather. You can tell a lot from the texture and density of crumpled, yellow Post-It notes at the bottom of an executive can. I'm not sure about our current IT head. There are days where I think he's in the danger zone.
As this goes, that's probably more useful than the intended interview.
The company received 40 million dollars in its Series B round of funding. The round was led by such notables as Mark Zuckerberg, Elon Musk, Peter Thiel, Vinod Khosla, and Ashton Kutcher.
The patent and copyright systems are also wealth redistribution programs.
If we get right down to it—say half as reductive as Dawkin's The Selfish Gene—any societal agreement is a wealth redistribution program. The politics of human reproduction are definitely wealth redistribution programs—probably the very pinnacle of the wealth redistribution pyramid. But that's okay, because morality.
The main difference between my wider version and the ordinary use of the term is that the wealth under redistribution is abstract enough that the bloodymindedness of loss-aversion politics reaches a lower grade of fever. The highest grade of fever is reserved for any dollar bill that has actually been in your wallet at any point in time while the full accounts were settled.
Go into any casino and ask random people how much they have won or lost, on that evening, or in any aggregate period. Be prepared to enjoy some very interesting narrative math. The narrative math won't, in most cases, have the slightest resemblance to bottom line math, from a more objective vantage point, e.g. a person engaged in a Martingale betting algorithm faces but a single outcome, almost surely, whatever character the ride appears to have for the uncertain duration. Perhaps there was a high point where the bedazzled schlep felt able to afford some drinks and hookers.
The technology behind self-driving cars has come up in a number of episodes of the O'Reilly Data Show hosted by Ben Lorica. Ben knows his stuff well enough to perform this role, but to my taste, he's pretty softball most of the time; his show is more of a polite survey than a contest of minds.
The guest is Shaoshan Liu, "co-founder of PerceptIn and previously the senior architect (autonomous driving) at Baidu USA".
As I recall it, Liu says that the instrument package for a fully autonomous self-driving car—in the not too-distant past—costs around $100,000 and requires 3000 W to post-process (heterogeneous algorithms, including some neural networks).
The "cheaper" self-driving cars require that the road network is completely pre-programmed, such that the car is almost more of a trail running on data rails than a fully autonomous vehicle.
It didn't sound like even the most expensive car can handle all conditions of day/night and heavy weather.
Some vain, vainglorious, or just plain glorious startup was mentioned with the goal of driving the cost of the LIDAR unit down to $200, probably with some fancy chip (the audio is not always well-balanced on this show; even with buds, I miss a lot unless I'm in a quiet room).
I personally wouldn't invest this kind of money at this time scale unless I thought "just plain glorious" was rounding the bend with mini-Craig Venter hard on the lash.
Do they know something we don't know? Or is this just another fifth-generation-AI industrial drum bang?
The benefit to the forty-niners was that the gold was simply "free for the taking" at first. In the goldfields at the beginning, there was no private property, no licensing fees, and no taxes.
The rules attempted to balance the rights of early arrivers at a site with later arrivers; a "claim" could be "staked" by a prospector, but that claim was valid only as long as it was being actively worked.
Miners worked at a claim only long enough to determine its potential.
If a claim was deemed as low-value—as most were—miners would abandon the site in search for a better one. In the case where a claim was abandoned or not worked upon, other miners would "claim-jump" the land. "Claim-jumping" meant that a miner began work on a previously claimed site.
Disputes were often handled personally and violently, and were sometimes addressed by groups of prospectors acting as arbitrators.
And it all goes down hill from there.
Pretty soon the disputants get dragged into an actual confangled courtroom where you're not allowed to address the judge as "your esteemed ball sack".
Move over, Mount Rushmore, Google now has an algorithm that can wallpaper the Ceres asteroid with the face of every American who has ever been photographed—all the way back to a pinhole camera exposing an onion skin soaked in lemon juice and potato starch.
It's been resoundingly debunked that present society can foot the bill on such an extreme fart-tolerant posture on free-speech purity. There just aren't enough clothespins in the known universe to make this ideal workable.
After a while, clipping the seven thousandth clothespin onto your nose begins to feel like something that imperialist Europeans would have quickly attributed to the Song Dynasty.
Call me fascist if you will, but I'm going to pass on self-inflicted free-speech warrior lotus nose.
I have a circadian rhythm disorder. Not that long ago I free ran on my innate 25.5 hour circadian day for three years, before I discovered that only sustained release melatonin is able to fix my problem (the shortest circadian day I achieved on any dose/formulation of non-SR mel was 24.15 hours, which required me to discontinue for two out of every five-six weeks for vertical retrace).
2016 was the first year of my adult life where I didn't lose entrainment with the calendar day.
But it's still not a bed of roses. I still feel a trace of my 16.5 day free-running circadian period, and it primarily affects sleep quality, more so than duration. I'll go through bad weeks where my afternoon melatonin induces extreme fatigue in the late afternoon / early evening.
This is not normal fatigue. I was once told by a German linguist that some parts of Germany have a way of saying "go downstairs" as "go down from up". In English, we normally say "fall asleep", in which consciousness unwinds top down. When my melatonin overpowers me, my consciousness unwinds from the bottom up. I'm not falling into sleep. Sleep is welling up like a glacier trapped in a tanning salon. I don't know if other people have experienced brief waking moments at night, where you are kind of consciously awake (you look around the dark room, check the time, groan a bit to yourself, etc.) but all the while whatever dream you were having continues apace. It's more like that than normal sleep, only the underlying alternative brain state becomes progressively more intense, while conscious tasks become progressively more difficult as the underlying circuits abandon consciousness to crash a different party.
During the years I spent free running, my sleep architecture was highly structured, but not in a good way. I divided my 16.5 day cycle into five lobes of three days each, because in each lobe my sleep architecture was noticeably different. By convention, my cycles began anew the first day I found myself waking up shortly after midnight (sleeping across midnight was a welcome condition). My five lobes progressed from fabulous, to good, to mediocre, to piss poor, to downright horrible. It had something to do with increasingly non-restorative sleep. Getting eight hours of solid sleep during each sleep period didn't seem to make much difference. (My sleep intervals were more consistent free-running than they are now, fully entrained.)
There seems to be two competitive cycles in my SCN: when my body wants to sleep, and when my body fully responds to sleep as a restorative function.
While free running, the cycle determining when I wanted to sleep had the upper hand at all times. Whether the sleep was fully restorative depended entirely upon which lobe I was in. For one glorious lobe, both cycles were in complete alignment. Then they began to drift. By the fifth lobe, it was like they were 180 degrees out of alignment. Being 180 degrees out of alignment seemed to send my SCN straight into the Bermuda triangle, with all dials spinning madly in the magnetic anomaly. On the last day of my last lobe, I usually experienced a very long waking period I called "stretch day". I wouldn't be very productive, but I'd be strangely alert—almost manic—for a 24-hour waking period. Bed time would jump from 07:00 directly to 14:00 (rather than the customary 1-1.5 hour daily jump during the rest of my cycle).
What I tended to do in my long nights of piss poor to appalling level of function was read a lot. I couldn't generate any kind of outward task signal. I needed to do something that provided a strong external signal, and I needed to latch onto it like the last tit in paradise, and never let go. My comprehension and retention were as high as ever, with one gotcha: not yet. While in this state, I could understand the book I was reading, but not fully relate it to any other knowledge I might have. First day of lobe one, I would often experience manic recall of every connection I had failed to
Slashdot Mirror
We used to call them "hot grits". In was an everyday topic of discussion.
You've already lost the fight: no human system outperforms its incentive structure.
Peer review is hopelessly ensnared by academic advancement culture. Entire disciplines can end up publishing bunk, if that becomes the tenure-track fashion of the decade. Tulip bubbles are not restricted to the business cycle. Even hard sciences have been hit pretty bad. Et tu, string theory?
The fundamental theorem of peer review is due to Max Planck:
The zone of convergence of peer review involves the passing of interested parties. In most of the hard sciences, fifty years pretty much weeds out the crap.
However, if you take a field such as nutrition science, I dare say it's still inadvisable to take fresh "peer review" at face value. John Yudkin was on the right track in 1958. Fifty years downstream, the truth is out there, but it's still far from evenly distributed in the public imagination.
Nutrition science was subverted by a white coat army of industry apparatchiks. These studies are expensive and, oh yeah, replication crisis.
Most human systems can be trusted some of the time. The real art of bullshit detection is figuring which times are those times. Even the best human systems are bullshit on the margin.
Faster-than-light neutrino anomaly
What you need to understand here is that the journalist impulse to publish is directly proportional to the tenuousness of the result in question.
Well, if the speed of light falls derp derp wormholes derp derp Stargate derp derp dusty von Daniken booster spice derp derp human immorality derp derp Omni Magazine alternate-reality cum shot. Well, you got your $4 worth, didn't you?
There's an enormous term in the human condition centered around escape from reality. This makes sense to some degree, because human reality usually contains a giant heal spur of oppression of the downtrodden masses (success has a habit of being highly asymmetric).
Trump's monosyllabic barrage becomes tremendously more convincing if you want to believe the underlying message.
Somehow, one supposes, being suckers for false hope must be evolutionarily adaptive (who, after all, is qualified to challenge the modern evolutionary synthesis?)
And then you get right down to it, the anchor tenants of modern bullshit culture are the major religions (being largely incompatible, at most 1 of N could anywhere close to broadly correct). Because, you know, life without bullshit would be empty and meaningless.
Deep down, most of us don't really want to drain the bullshit pond. And it's not just one pond. It's pond after pond. Never get comfortable.
The fundamental theorem of bullshit busting is due to Richard Feynman:
Evolution took a long look at Hamlet, and came up with satisficing.
Make happy assumptions that are compatible with medium-term survival (generally best obtained from proven survivors—aka your parents and select community), then behave with the efficiency of assuming their truth, until the shit really hits the fan; then sit back, renounce, regroup, and repeat.
Dawkins pretty much feels about religion the way Einstein felt about cosmic inflation and quantum indeterminacy. Right model, wrong hope, long painful row to hoe. Even when our best minds get something right, they're often left wishing they hadn't.
So there's this unhappy observation about the human
The article has a poor to false understanding of how blue light interacts with DLMO (dim light melatonin onset).
I'm pretty sure the entrainment effect of blue light is via direct neuronal connection to the SCN, and I doubt it involves melatonin, except indirectly.
The homeostatic sleep pressure signal builds up (more or less linearly) for as long as you're awake. On its own, this would mean that you taper into drowsiness all day long. So the sleep system has another mechanism that suppresses response to the sleep pressure signal. I vaguely recall that what happens with DLMO is that melatonin onset signals the body to turn off the suppression switch, so that the body begins to notice the homeostatic sleep pressure signal.
DLMO, however, is easily inhibited by exposure to blue light at a point in time approximately an hour before bedtime. If you're outdoors hunting moose in the bright light of late-evening arctic summer, this is a useful adaptation.
You'll get to bed later, which means you'll sleep a bit later (but not much) and then you will get less blue light early the next morning, which will affect your entrainment, gradually, on the slow-drip program.
—
As a rough, empirical ratio, for every extra hour you stay up, you'll sleep about twenty minutes later the next morning. It's not uncommon to stay up for an extra two hours, then barely sleep in for an extra half hour. (We need to ignore here that modern society tends to run a massive, permanent sleep deficit, which can suddenly turn into sleeping four to six hours late at the first opportunity that allows this to happen. That's a different beast entirely.)
I have a circadian rhythm disorder, and I know from decades of sleep tracking that morning wake-up time is about three times more reliable in estimating my sleep phase than time of retirement.
—
This is a worthwhile paper from the top of my notes, but it's hard to wade through:
Estimating Dim Light Melatonin Onset (DLMO) Phase in Adolescents Using Summer or School-Year Sleep/Wake Schedules — 2006
I like this paper because it shows how social convention (adolescent schooling) also influences DLMO phase.
—
The sleep pressure signal eventually overwhelms the suppression of this signal, regardless of the DLMO mechanism.
James Maas is a good representative of the modern sleep science orthodoxy:
Surefire Strategies to Sleep for Success!
I just love the page break at the end of page 6. But then I'm really into microscopic moments of small page-formatting humour. (It's probably not unrelated to all those long, lonely nights, before I found a viable treatment.)
—
Here's a good summary, I just found for the first time.
Phase Response Curve
The reason I only vaguely remember this mechanism is that all the phase response curves in the literature are dose dependent.
There is no PRC I've ever seen that computes the phase response differential to endogenous melatonin levels. No, what you do is administer some dose/formulation (which can include sustained-release components) at staggered times over several weeks, and then you plot the graph averaged over your test population (which thus includes all the metabolic uptake and clearance variability).
There was a time I desperately wanted to consult one of these curves and then to declare "I am here", but it never happened. These are, in effect, better regarded as qualitative curves than quantitative curves.
The model was never predictive enough to be worth memorizing exactly. And thus I remain slightly dim on DLMO when I really shouldn't be after all these years.
The price of shame — March 2015
The Republican weaponization of Clinton's misdeed was to claim that this behaviour made Bill unfit to govern. (If powerful men having extramarital affairs with young women was incompatible with leadership, well, the vast sweep of history does not so record.)
Family values aside, the power imbalance creates the risk that Bill would abuse his immense power to cover up the vastly exaggerated blot on his record. The Republicans actually knew that anyone with an accurate base rate of human history / human culture would not regard his behaviour as incompatible with leadership—though a common and damning blot nevertheless, so the tactic was to escalate the stakes until Bill felt compelled to lie about it—which, unfortunately, was extremely easy to anticipate.
Lying to formal body of review is considered incompatible with leadership, sort of, incrementally, since not all that long ago. For example, it barely extends as far back as the Reagan's Iran–Contra affair. (Some people roll with family values and view Clinton's offense as the worse offense. I happen to roll with geopolitical transparency, and so I view Reagan's offense as the worse offense—he appointed those clucks, and it was his ultimate responsibility to know all the big shit).
Bill was plenty smart enough to figure out that the public perception battle would play out exactly as it did, leaving him boxed into a corner where he could—according to his established character—only choose to lie (perhaps he overestimated his power to blow off the investigation, but even there, had he succeeded, he would have mortgaged a sizeable fraction of his presidential energy in ruthlessly defending his momentary gratification).
Clearly, his judgment in this matter fell short of the mark by any standard.
However, I rate it not quite as bald as boasting about sexual harassment with a camera rolling. Whatever Bill purportedly said to Donald on the golf course (that was "far worse" in Donald's personal judgement), there was no film at eleven after the fact.
The modern world contains a lot of cameras and microphones. Trump's world has contained many cameras and microphones since way back. A prudent man in his position wouldn't be openly bragging about his magical power to get away with sexual harassment just to impress Billy Bush. And it's not like Donald didn't have a front row vantage point on Bill sinking his own boat through which to consider and amend his own standard of personal conduct. Donald had every opportunity to know better, and the penny never dropped.
So in summary, a whole lot of things are "not okay" but still the world largely spins as it has always done for thousands of years.
Considering how much you originally paid for your deoxygenated speaker wire, I would say $40 is the least you can do.
Were you to model the price signal with quaternion rotation instead, you would discover that the price signal really can spin around to a perfect 180-degree inversion of "the best you can do", given but a sparse free-energy input of mindless optimism, and a scant few months to capture abandoned area under the curve (and that's not even including the machine learning revolution).
Perhaps capitalism eventually does the right thing, but not until after imbibing all the loose sugar.
—
First Law of Mice and Men: If something can go wrong, it will go wrong.
Corollary: If something can lead to an easy buck, it will lead to an easy buck.
Unfortunately, all the quaternions in this picture belong to the increasingly neutered EPA.
In a dozen lifetimes I could hardly scratch all the great content on the internet that's available for free, or with barely any strings attached.
The best defense against the dark arts of advertising is a curiosity streak that's a mile wide.
Two words: substitute good.
At the first sign of trouble, I open alternate tabs like a cowboy after a bar brawl.
Why stop your argument there? With another sentence or two, we can rewind western civilization all the way back to the Taliban's conception of marriage.
Because in any human collective, only one party can ultimately wear the pants.
How the truth stings. Resistance is futile. Sauron does not share power. Those poor, deluded Swedes. Yada yada yada.
Trust, but keep one foot wedged in the emergency exit, and one ear cocked for a fell voice on the air:
I cannot continue working on my add-ons anymore. I'm sorry, but it's time.
Also known as perpetual explanation machine.
Much to the point of this thread, the day is not long in coming where a series of small, hotly debated innovations in machine learning will culminate in a robust classifier able to ferret out a dozen egregious offenses against working brain cells in that toxic screed of hot-button click-bait you just posted.
The wise among us will use this forthcoming capability to accept all well-formed signals, the fools will filter bubble to black. The later group being larger than the former group, while the former group holds all the marbles, much derived from clue will change, while much derived from populism changes little, as our social algorithms contrive to sneer, in an arms race of eloquent dissection.
If Linus had ever worked on a truly hard problem, he might think different. Not every impasse in life can be resolved by industrious sleeve-rolling.
Operating systems don't hold opinions on human social dynamics. Earth-shattering innovation need not apply.
Your implication being that UC Berkeley doesn't see fit to make this caliber of legal advice available to faculty self-evidently working in fields with billions of dollars at stake?
If Berkeley fell short on sound legal acumen with the Holy Grail of the Biological Revolution inches away from the tips of their greedy little fingers, god help man with garage.
I have trouble comprehending the small mental world you live in where all of your knowledge is equally available at all times.
There's a reason why it's polite to gloss your acronyms on first use, even in the narrowest academic publications.
Just yesterday I was reviewing the literature on machine learning. The Juergen Schmidhuber review alone begins with the following glossary:
AE: Autoencoder
BFGS: Broyden—Fletcher—Goldfarb—Shanno
BNN: Biological Neural Network
BM: Boltzmann Machine
BP: Backpropagation
BRNN: Bi-directional Recurrent Neural Network
CAP: Credit Assignment Path
CEC: Constant Error Carousel
CFL: Context Free Language
CMA-ES: Covariance Matrix Estimation ES
CNN: Convolutional Neural Network
CoSyNE: Co-Synaptic Neuro-Evolution
CSL: Context Sensitive Language
CTC: Connectionist Temporal Classification
DBN: Deep Belief Network
DCT: Discrete Cosine Transform
DL: Deep Learning
DP: Dynamic Programming
DS: Direct Policy Search
EA: Evolutionary Algorithm
EM: Expectation Maximization
ES: Evolution Strategy
FMS: Flat Minimum Search
FNN: Feedforward Neural Network
FSA: Finite State Automaton
GMDH: Group Method of Data Handling
GOFAI: Good Old-Fashioned AI
GP: Genetic Programming
GPU: Graphics Processing Unit
GPU-MPCNN: GPU-Based MPCNN
HMM: Hidden Markov Model
HRL: Hierarchical Reinforcement Learning
HTM: Hierarchical Temporal Memory
HMAX: Hierarchical Model "and X"
LSTM: Long Short-Term Memory (RNN)
MDL: Minimum Description Length
MDP: Markov Decision Process
MNIST: Mixed National Institute of Standards and Technology Database
MP: Max-Pooling
MPCNN: Max-Pooling CNN
NE: NeuroEvolution
NEAT: NE of Augmenting Topologies
NES: Natural Evolution Strategies
NFQ: Neural Fitted Q-Learning
NN: Neural Network
OCR: Optical Character Recognition
PCC: Potential Causal Connection
PDCC: Potential Direct Causal Connection
PM: Predictability Minimization
POMDP: Partially Observable MDP
RAAM: Recursive Auto-Associative Memory
RBM: Restricted Boltzmann Machine
ReLU: Rectified Linear Unit
RL: Reinforcement Learning
RNN: Recurrent Neural Network
R-prop: Resilient Backpropagation
SL: Supervised Learning
SLIM NN: Self-Delimiting Neural Network
SOTA: Self-Organizing Tree Algorithm
SVM: Support Vector Machine
TDNN: Time-Delay Neural Network
TIMIT: TI/SRI/MIT Acoustic-Phonetic Continuous Speech Corpus
UL: Unsupervised Learning
WTA: Winner-Take-All
And it's but one of dozens of fields where I stick my finger into the alphabet pie.
After pressing "submit", in a split-second second evaluation, I noticed that that sentence I wrote does not quite work.
Problematic:
Less problematic:
Cognitively, this is a turtle too soon for full effect.
Even less problematic, but horrifying:
Horrifying because I've always regarded this kind of sub-phrase repetition as the hallmark of hack speechwriters. One can partially excuse this by placing a semicolon in front of the repeated phrase, but here the semicolon is incompatible with the closing mdash.
Language is a complex solution gradient, one that humans have yet to successfully express. Sad. All those unemployed genetic algorithms, awaiting human clue.
I've spent my entire life trying not to be this dim. Yes, very clever work there treading on the narrow definition—while engaged in 100% baby flush.
The ridiculousness of this is apparent to any thinking person in less time than it takes to type "Wittgenstein".
Because some human process defined the solution gradient that the "genetic" software optimized over—ad infinite turtle—in an act of algorithmic emancipation now glibly lumped under the verb "write" by the baby-impervious membrane all-too-tragically-often comporting itself as "logic".
For myself, negative ideation (which for me is not suicidal, just an endless litany of everything I've ever done wrong in front of another person) strongly correlates with sleep quality.
As soon as I reach the threshold where the negative thoughts are disruptive to daily life, I take of several effective sleep medications, and—if I manage to get the hard and deep sleep I need—negative ideation is gone again the next day.
Hours in the sack don't count, either. I've had times where I've been getting a solid eight hours of consolidated sleep for an entire week, but still the negative ideation has made an appearance. Sometimes my sleep seems to hollow out so that it's non-restorative on some hidden, inner dimension.
At present, my best sleep-quality aid is 3–4 mg of nortriptyline, a dosage I have custom-compounded at a local compounding pharmacy.
I was originally prescribed amitriptyline, which also worked great to restore my sleep quality, but left me fuzzy-headed the next day.
The later substitution with nortriptyline was based on my own research effort, as was carefully titrating my minimal therapeutic dose (which ends up costing me six times than as much as the minimum standard dose of 10 mg, available in capsule form only—what's less fuzz in the brain the morning after worth to you?)
If I really need to drop the sleep hammer, I take two, then write off part of the next morning to light housework.
In related news, 85% of both groups combined think they are good at their jobs.
Interviewer: You get paid the big bucks. Are you doing it wrong?
Interviewee #1: Well, gosh, I don't know.
Interviewee #2: Every damn time, and twice for breakfast.
Interviewer: Uh, #2, how long have you held your current rank.
Interviewee #2: The previous numbnut is still fumbling for his keys in the parking lot, with all his executive possessions packed in an open box, tucked under his left arm.
Interviewer: How about you, #1?
Interviewee #1: Twenty-two years.
Interviewer: Really? You've been running the IT department for twenty years?
Interviewee #1: Actually, no. I'm the janitor. The chief custodian wears a shirt and tie, so I do, too. Always dress like the boss, you know. Good career advice passed down from my grandfather. You can tell a lot from the texture and density of crumpled, yellow Post-It notes at the bottom of an executive can. I'm not sure about our current IT head. There are days where I think he's in the danger zone.
As this goes, that's probably more useful than the intended interview.
I was looking into the deep learning celery diet earlier today.
Uber Buys a Mysterious Startup to Make Itself an AI Company
Many smart people in deep stealth.
Vicarious (company)
When has Peter Thiel ever been wrong?
The patent and copyright systems are also wealth redistribution programs.
If we get right down to it—say half as reductive as Dawkin's The Selfish Gene—any societal agreement is a wealth redistribution program. The politics of human reproduction are definitely wealth redistribution programs—probably the very pinnacle of the wealth redistribution pyramid. But that's okay, because morality.
The main difference between my wider version and the ordinary use of the term is that the wealth under redistribution is abstract enough that the bloodymindedness of loss-aversion politics reaches a lower grade of fever. The highest grade of fever is reserved for any dollar bill that has actually been in your wallet at any point in time while the full accounts were settled.
Go into any casino and ask random people how much they have won or lost, on that evening, or in any aggregate period. Be prepared to enjoy some very interesting narrative math. The narrative math won't, in most cases, have the slightest resemblance to bottom line math, from a more objective vantage point, e.g. a person engaged in a Martingale betting algorithm faces but a single outcome, almost surely, whatever character the ride appears to have for the uncertain duration. Perhaps there was a high point where the bedazzled schlep felt able to afford some drinks and hookers.
"Redistribution" is a term of narrative art.
I've worked hard to get through life, never adding this particular club to my bag.
There are days, though, when the grass does seem greener on the other side of the fence.
The technology behind self-driving cars has come up in a number of episodes of the O'Reilly Data Show hosted by Ben Lorica. Ben knows his stuff well enough to perform this role, but to my taste, he's pretty softball most of the time; his show is more of a polite survey than a contest of minds.
Here is one link I could quickly find:
* The technology behind self-driving vehicles
The guest is Shaoshan Liu, "co-founder of PerceptIn and previously the senior architect (autonomous driving) at Baidu USA".
As I recall it, Liu says that the instrument package for a fully autonomous self-driving car—in the not too-distant past—costs around $100,000 and requires 3000 W to post-process (heterogeneous algorithms, including some neural networks).
The "cheaper" self-driving cars require that the road network is completely pre-programmed, such that the car is almost more of a trail running on data rails than a fully autonomous vehicle.
It didn't sound like even the most expensive car can handle all conditions of day/night and heavy weather.
Some vain, vainglorious, or just plain glorious startup was mentioned with the goal of driving the cost of the LIDAR unit down to $200, probably with some fancy chip (the audio is not always well-balanced on this show; even with buds, I miss a lot unless I'm in a quiet room).
I personally wouldn't invest this kind of money at this time scale unless I thought "just plain glorious" was rounding the bend with mini-Craig Venter hard on the lash.
Do they know something we don't know? Or is this just another fifth-generation-AI industrial drum bang?
And it all goes down hill from there.
Pretty soon the disputants get dragged into an actual confangled courtroom where you're not allowed to address the judge as "your esteemed ball sack".
Fucking language police.
Move over, Mount Rushmore, Google now has an algorithm that can wallpaper the Ceres asteroid with the face of every American who has ever been photographed—all the way back to a pinhole camera exposing an onion skin soaked in lemon juice and potato starch.
It's been resoundingly debunked that present society can foot the bill on such an extreme fart-tolerant posture on free-speech purity. There just aren't enough clothespins in the known universe to make this ideal workable.
After a while, clipping the seven thousandth clothespin onto your nose begins to feel like something that imperialist Europeans would have quickly attributed to the Song Dynasty.
Call me fascist if you will, but I'm going to pass on self-inflicted free-speech warrior lotus nose.
I have a circadian rhythm disorder. Not that long ago I free ran on my innate 25.5 hour circadian day for three years, before I discovered that only sustained release melatonin is able to fix my problem (the shortest circadian day I achieved on any dose/formulation of non-SR mel was 24.15 hours, which required me to discontinue for two out of every five-six weeks for vertical retrace).
2016 was the first year of my adult life where I didn't lose entrainment with the calendar day.
But it's still not a bed of roses. I still feel a trace of my 16.5 day free-running circadian period, and it primarily affects sleep quality, more so than duration. I'll go through bad weeks where my afternoon melatonin induces extreme fatigue in the late afternoon / early evening.
This is not normal fatigue. I was once told by a German linguist that some parts of Germany have a way of saying "go downstairs" as "go down from up". In English, we normally say "fall asleep", in which consciousness unwinds top down. When my melatonin overpowers me, my consciousness unwinds from the bottom up. I'm not falling into sleep. Sleep is welling up like a glacier trapped in a tanning salon. I don't know if other people have experienced brief waking moments at night, where you are kind of consciously awake (you look around the dark room, check the time, groan a bit to yourself, etc.) but all the while whatever dream you were having continues apace. It's more like that than normal sleep, only the underlying alternative brain state becomes progressively more intense, while conscious tasks become progressively more difficult as the underlying circuits abandon consciousness to crash a different party.
During the years I spent free running, my sleep architecture was highly structured, but not in a good way. I divided my 16.5 day cycle into five lobes of three days each, because in each lobe my sleep architecture was noticeably different. By convention, my cycles began anew the first day I found myself waking up shortly after midnight (sleeping across midnight was a welcome condition). My five lobes progressed from fabulous, to good, to mediocre, to piss poor, to downright horrible. It had something to do with increasingly non-restorative sleep. Getting eight hours of solid sleep during each sleep period didn't seem to make much difference. (My sleep intervals were more consistent free-running than they are now, fully entrained.)
There seems to be two competitive cycles in my SCN: when my body wants to sleep, and when my body fully responds to sleep as a restorative function.
While free running, the cycle determining when I wanted to sleep had the upper hand at all times. Whether the sleep was fully restorative depended entirely upon which lobe I was in. For one glorious lobe, both cycles were in complete alignment. Then they began to drift. By the fifth lobe, it was like they were 180 degrees out of alignment. Being 180 degrees out of alignment seemed to send my SCN straight into the Bermuda triangle, with all dials spinning madly in the magnetic anomaly. On the last day of my last lobe, I usually experienced a very long waking period I called "stretch day". I wouldn't be very productive, but I'd be strangely alert—almost manic—for a 24-hour waking period. Bed time would jump from 07:00 directly to 14:00 (rather than the customary 1-1.5 hour daily jump during the rest of my cycle).
What I tended to do in my long nights of piss poor to appalling level of function was read a lot. I couldn't generate any kind of outward task signal. I needed to do something that provided a strong external signal, and I needed to latch onto it like the last tit in paradise, and never let go. My comprehension and retention were as high as ever, with one gotcha: not yet. While in this state, I could understand the book I was reading, but not fully relate it to any other knowledge I might have. First day of lobe one, I would often experience manic recall of every connection I had failed to
Why do dogs lick their genitals? Because they can. Such is the human condition that self-love is insufficient unless you're extraordinarily limber.
And yes, it's a metaphor.
At the end of the day, you can't eat self-opinion. I suspect Marcus Aurelius was deliberately playing dumb for comic effect.