I once read a book by Linda Hill that I personally found amazingly valuable, but only because I was careful not to light any matches, because her presentation was dry, dry, dry.
Because of the Indian incompetence story here on Slashdot this morning, I went to paste a link into my files, and chanced upon a past entry concerning HCL Technologies, a topic that Linda Hill has addressed in video, and soon I found myself watching a clip of hers on YouTube I hadn't seen before.
The problem in India with the educational system is that the system dictates and student repeats.... We all had to unlearn how we were educated. And the leaders had to unlearn what they thought leadership was about. Because if you grow up in that kind of system, when you're a leader what you think your role is, is that you're supposed to set direction and make sure nobody deviates from it. That's fundamentally how they saw their role.
And here we have this Wikipedia article, where the unstated premise seems to be "Surprise! Derf-derf-derf, Wikipedia doesn't actually practice zero-deviation culture, despite their publicly assigned role as the plastic–pocket-protector paragon of geek dysfunction.
No, instead what we have is this: if a source is broadly flagged as tainted, it becomes open season to replace this source with a better citation wherever and whenever, without expecting significant blow back.
Isn't that leadership enough?
Is the underlying zero-deviation fixation that motivates this story just a tired strawman? Or is this derf-derf strawman meme playing to a real audience?
Well, I personally would run, run, run if I found myself in that audience, because anyone who doesn't is doomed to be soon be looking up at India as the management enlightenment movement that just passed you by with a big whoosh.
I don't see a problem with this. You want specific levels of error handling? Put it in the spec.
#include "no_abe_normal.h"
If you're not familiar with this convention (it appears you haven't been in this business long enough to hear the pathetic whimpers of Forma L. val d'Ation sequestered away from public shame in an attic antechamber), the "h" stands for "head".
Republicans only care about money. Can you do the job? Good. Get to work.
Cutting red tape to ribbons is an intrinsically easier job than building up effective layers of regulation that prevent the public interest being bent over a barrel, while the longest of all possible rubber gloves rummages around for the better part of a trillion dollars.
Evidently, no money was harmed in the operation.
The job, as I see it, is a little harder to accomplish, once you concede that there is such a thing as effective regulation, though it's yet far from a science; science also being a discipline where time after time ones best efforts fall short, and yet one perseveres.
In the best case scenario, even after regulation becomes more of science, it will still be double hard: hard to do and hard on the ego.
Kind of makes a guy want to double down on only caring about money, setting oneself up on a lavish private beach, and watching the glorious Egos soar.
Now obviously something was wrong with the polling data.
Because why? Because popular opinion has guaranteed monotonic convergence? At a guaranteed quadratic convergence rate?
Just what part of "moving target" is so hard for people to understand?
Candidate A shits her pants at the front of the boat. Everybody rushes to the back of the boat.
Candidate B begins barfing up a taco bowl. Everybody rushes back to the front of the boat.
Blather. Foam. Repeat.
The only reason Trump won is because on the day of the election, there were more disgusted voters headed to the back of the boat than the front of the boat.
Polls, especially rolling meta polls, have an intrinsic lag of two to four days. I clearly saw on the 538 graphics momentum building toward the rear bulkhead as we rounded into election day. After I extrapolated the trend a few days forward to compensate for polling lag, I was not surprised by the final outcome.
There was more than enough disgust in both directions to support either outcome. Another Billy Bush tape in the final week could easily have turned the tide. This was not a normal election where the polls were tracking a slow convergence of the undecideds. The polls were tracking a mad (and futile) scramble for the voters to distance themselves from whichever paragon of disgust was recently the most salient.
Moving target. The polls were no less instantaneously accurate than they've ever been. They just don't work very well when neither candidate has a redeeming feature, and the electorate goes into orbit around a positive pole in the complex plain.
How is this even remotely difficult to comprehend?
From where I sit, it's all bog-standard Electoral Engineering for Dummies, 101.
I'm a card-carrying Goldman Sachs conspiracy theorist. I really do think they systematically rig the market to their ultimate advantage, one ballsy five- or ten-year bamboozle after another.
Their main obstacle is that it's very hard to fool people twice with the same bullshit, so there always has to be a new fundamental disruption lurking around the next corner. I don't yet know what AI really is, but I sure know it's covered in fleece.
A sure tell is the narrative of nested boxes: each shiny thing within the dark thing is ever an order of magnitude more disruptive, more profound, and more lucrative.
The innermost thing is so brilliant, set against such a dark background (exponentially dark, by the miracle of multiple contrast-enhancing frames), it blinds you altogether.
On 25 August 2009, U.S. Patent 7,580,533 was issued for a device that uses a laser beam and smoke or vapor to detect sound vibrations in free air ("Particulate Flow Detection Microphone based on a laser-photocell pair with a moving stream of smoke or vapor in the laser beam's path"). Sound pressure waves cause disturbances in the smoke that in turn cause variations in the amount of laser light reaching the photo detector.
Or you could drop a few thumb-sized motes.
I like 2001, the Russian Solaris, and A Scanner Darkly.
Blade Runner and Alien were better than a jab in the eye.
If we further widen the net to include Space Fiction, The Wrath of Kahn rocks; while A New Hope and WALL-E both have their moments.
If we further wide the net to include any form of thematic overlap, I'd include The Right Stuff, Apollo 13, the first Back to the Future, the first Iron Man, select chunks of The Terminator franchise, RoboCop, Young Frankenstein, Dr Strangelove, and certain aspects of The Fifth Element. One might even include the sensibility of Tree of Life or Hugo.
I'll also give an honourable to The City of Lost Children, because I would actually rewatch that movie. Can't recall much of anything about the plot (not usually a good sign), but there's plenty of there there in other regards. In a pinch, I could rewatch Dune as an entertaining car wreck.
Unfortunately, much of the rest of the canon only serves to rouse my appetite without entirely beddin' her back down.
Note that I did not exclude any Spielberg movies by accident. If I had to rewatch one, it would be THX 1138. Spielberg is so sentimental, I'm soon humming Indian Love Call and wishing it would work.
On my list as the least science fiction film ever made would be the original Matrix. Perhaps the humans harvested for their cerebellar electricity was a satirical neoliberal talking point adapted from Ayn Rand.
Obviously there are disadvantages to having only one graphics language, but the benefits outweigh the disadvantages.... Android made the same decision to go that way. Even Wayland to a certain degree has been doing that. They have to support EGL and GL, simply because it's very convenient for app developers and toolkit developers — an open graphics language. That was the part that inspired us, and we wanted to have this one graphics language and support it well. And that takes a lot of craft.
So, once you can say: no more weird 2D API, no more weird phong API, and everything is mapped out to GL, you're way better off. And you can distil down the scope of the overall project to something more manageable. So it went from being impossible to possible. And then there was me, being very opinionated. I don't believe in extensibility from the beginning — traditionally in Linux everything is super extensible, which has got benefits for a certain audience.
If you think about the audience of the display server, it's one of the few places in the system where you've got three audiences. So you've got the users, who don't care, or shouldn't care, about the display server.
How is it that I never fall into the category of people described as "users"?
Does what I do for ten hours a day, every day, not fall into the semantic category of "using"? Me, and everyone like me? How do we always find ourselves filed under "a certain audience"? Well, this "certain audience" is today crying no giant room-temperature crocodile tears—neither any small, steamy gnat tears.
Here's the underlying problem: "user", as fantasized by far too many software developers, is the centerfold normalization of real womanhood.
I thought Poker was a game of understanding your opponents not only based on past actions with cards but also by looking at facial expressions, body language and determining whether or not they have a good hand. Along with that, a big part is developing subtle gestures to throw your opponents off.
Hollywood much?
Also, cops fight crime mainly by experience tragic science field trips in early childhood.
Here's a pro tip. If the actors are loving it, it has probably had the living Snopes kicked out of it, supposing there was any at all to begin with.
Actually, you only came second out of the small group of readers who didn't also have a second response to the day's story that was more worth the bother of typing out.
It's one of those tasks in life that naturally goes to the few.
Perhaps if I had twenty hands, and could race to be first on all my instantaneous takes at the same time I would finally embrace the joy—the joy I have witnesses so many times in others—of rushing to post the obvious.
I'm at such a huge disadvantage here. By the time I winnow my winning candidate(s) down to hands available—just that small extra delay is enough—I'm probably last off the starting blocks. Hence my decades of sour grapes at those who excel me.
The IBM PC was an overpriced, slipshod piece of hardware even by the standards of the time.
The keyboard was indestructible. The case was a tank. The monochrome monitor displayed 25 rows of 80 columns, including upper and lower case letters.
Before the IBM PC was introduced, the personal computer market was dominated by systems using the 6502 and Z80 8-bit microprocessors, such as the TRS 80, Commodore PET and Apple II series, which used proprietary operating systems, and by computers running CP/M.
More than 50 new business-oriented personal computer systems came on the market in the year before IBM released the IBM PC.
Very few of them used a 16- or 32-bit microprocessor, as 8-bit systems were generally believed by the vendors to be perfectly adequate, and the Intel 8086 was too expensive to use.
Unfortunately, the IBM PC also included a few deliberate sandbags.
IBM decided to use the Intel 8088 after first considering the Motorola 68000 and the Intel i8086, because the other two were considered to be too powerful for their needs.
IBM never figured this was their last entry into the PC business. It was supposed to be a trial balloon. They had a zillion reasons to cripple the first edition, both in terms of processing power and in terms of memory expansion capacity.
By the way, did you actually use a TRS 80, Commodore PET, or Apple II? I used all three. Realistically, these all sucked for any serious purpose—except for learning the difficult art of programming the hard way.
It was just the other night I realized how starting my programming career on a TRS 80 with the notoriously unreliable tape drive influence my programming style for years to come.
No, BASIC did not ruin me. (I also picked up APL, several dialects of assembler within a year, rudimentary Pascal, some LISP, some FORTH, and most of C just as soon as I could get my hands on it.)
What did ruin me was the inability to curate a subroutine library of my favourite helper code. It just too took long to merge one chunk of code off cassette into another. (I believe the merge mode was that whatever new BASIC program you loaded just wrote right over top of any existing line numbers.) The TRS 80 was the computer I could use at school for free, which I did after school every day. Never had one at home until much later.
There was a certain kind of robustness you just didn't worry about, because every single program was pretty much home-cooked from scratch. At most, one might load something vaguely similar and then cannibalize some of the common bits.
Agile, look out—you ain't gonna need it. Every line of code ended up written in the least general way possible, so long as it sped up the code entry process.
Fortunately, I never had to use a cassette drive on an IBM PC.
On the IBM PC, I still had to multi-pass the compiler by switching floppy disks during the compile and link cycle, but that's a whole other story.
You know, the "standards of the time" included Heathkit, and Hewlett Packard (back when that still meant doing the right thing), and Tektronix, and later Compaq. The hobby computers were junk in part because everyone knew it was going to be a brisk ride. SOMEDAY SOON THERE MIGHT EVEN BE LOWER CASE. Basic economics.
Except for the IBM sandbag trick. That was old school economics, a first sour taste of something us hobbyists had not yet had to worry about.
That was the true legacy of the original IBM PC. It was the first coldly calculated, deliberate consumer diss. We all hates it forever for exactly that one thing.
because Canada uses the metric system but probably still buys its printing presses from the U.S.
Or maybe, just maybe it's the entire infrastructure of automated bill-handling equipment, including vending machines?
But I'm sure it will grief some neatfreak crack dealer who feels compelled to store giant mountains of bills in his metric utility room—no wait, Canada still uses two-by-fours (for which neither dimension is a round number in any system of measurement) and 16" stud centers.
You will be hard-pressed to make a case that human intelligence is anything but a catastrophic failure and/or malfunctioning system by any rational standard.
Yes, but the key step in your argument is not what you think. The exact moment you admit "rational standard" as a viable yardstick, your human intelligence steps off a plank into catastrophic failure.
Your unreasoning, intensely emotional response (larded at random with non-sequitur hot buttons) is why we all tend to be better off, long term, as competent, cool, and collected automatic control systems are further instigated.
There are plenty of competent control systems out there. By their very nature, they're not in the news on a daily basis.
This is actually calibrated behaviour. I click "submit" when the odds of noticing such a mistake fall below the bar of justifying the extra time investment.
And then sometimes I spot it right as the afterimage fades.
We have a name for people who are best served by quick decision making: we call them first responders (a category which sometimes includes pilots).
Much of the rest of the time, quick decision making is used to whip up drama rather than obtain a broad, positive outcome.
Poor slow decision: going out drinking without first planning transportation home.
Seemingly great quick decision: serving to avoid hitting the lamppost.
Poor quick decision: serving to avoid hitting the lamppost into oncoming traffic.
Happens in politics all the damn time. Yet we still keep electing these charismatic blowhards. Many of Trump's voters explicitly praised his really short approval pipeline between brainstem and mouth. Who knew healthcare could be so complicated?
Mike is a smart guy. I like him. But watch him blow his own foot off and gnaw upon the bloodied remains starting at about 7m50 into the following presentation:
He wants to tackle this "next" project by some loose, unspecified analogy with recent surprising progress in computer vision and computer robotics.
Of course, how could the frontal lobe be different than dedicated perceptual and motor lobes?
If this is the standard—and this is by no means unusual among even the smartest people in the field—accurate prediction of forward progress will continue to be a cheque in the mail until the end of freaking time.
How can any thinking person in this day and age not realize that AI predictions are HARDLY EVER worth the paper they're no longer printed upon?
Successful AI prediction is five years away—always has been, and always will be.
And no, the worthless prediction of consensus merit was not "decades" but "about a decade". I've been consuming machine learning podcasts day in and day out since the end of January. AlphaGo is not infrequently mentioned.
My two favourite guests so far (not specific to Go) have been Ilya Sutskever (deep intuition; now of OpenAI, which bodes well) and David Blei (broad, nuanced perspective, with particular expertise in LDA and topic modelling).
Probably those interviews were from Talking Machines (sadly, no longer active) or The O'Reilly Data Show (the intense industry focus and reverence for Apache Spark gets on my nerves after a while, but it's reasonably meaty for the most part).
A technology so rooted in public acceptance that it's no longer necessary to communicate up front with the users who will most suffer from the upcoming change cycle.
I didn't leave Ubuntu because of Unity.
I left Ubuntu because no transition plan was put forward to aid me in riding out the early adoption cycle from a safe remove whereby I retained the full use of my extra monitors and the meticulous workflow depending upon these that I had painstakingly adopted over many years.
There's nothing intrinsic to mainstream that I reject, other than how becoming meanstream seems to immediately entitle the proprietor to carpet yank—without even the courtesy of a gruff dentist, who at least mutters reassuringly "this won't hurt a bit".
If you remove the contributions of rms, the open source movement never would have started. If you remove the contributions of Linus, the open source movement would have never reached the heights it reached today. If you remove the contributions of esr, the open source movement would have fractured and broken into nothingness.
This is the kind of post-hoc hagiography that fuels Ayn Rand's objectivism crap mountain.
Out of these three, it's only hard to imagine a different, yet equally grand path if RMS had never occurred. His foundational dog work on binutils invited many others to host parties they could not have otherwise managed to throw.
Due to RMS as a unique personality, we got a highly political license sooner than we would have by another probable path. This was both a strength and a liability, whose relative magnitudes are almost impossible to judge in retrospect.
Without Linus, FreeBSD either would have become far more participatory, or some variant with a far greater embrace and tolerance of messiness would have forked within two years. And since this wouldn't have embraced GPL at the system level, Gnu HURD might even have been finished, with perhaps a necessary course correction or two under mounting pressure from a large install base.
Saying that open source wouldn't exist as it now does without ESR is pretty close to saying that the internet boom of the late nineties would not have happened without George Gilder. ("George who?" all the children ask. Exactly my point.)
Certainly charismatic figures come along when the moment is ripe to crystallize the zeitgeist, but history does not record that these people have ever been in short supply (something that would become immediately obvious if that stupid scheme from Atlas Shrugged had ever been tried for real in the real world.) Generally, you can never have more than a few of these types at any given time, because the human psychology of prophets and prophecy accrues special powers to the lone voice. A solitary howling wolf is divine revelation, a million howling wolves is just a statistical noise (see again the howlingly ludicrous legacy of one Ayn Rand).
Is it just me or does this whole diatribe just ooze "pathetic marketing maneuver"?
For every action there's an equal and opposite reaction. Now that half the population has levelled-up on aggressive ignorance, the standard for those who still pretend to know better only becomes that much higher.
Intel is very much one of those pretenders.
Paging Gordon Moore.
Gordon—dressed exactly like Bob Pinciotti—steps into the Tardis that just materialized outside his office, and grabs the ringing phone—midway through the fiftieth ring.
Gordon: Hello. Get this damn thing off my lawn. Can't you pre-foresee I'm trying to calculate here? Hello?
Weedy, tired voice of greybeard future: Uh, Gordon, terribly sorry for the interruption, we'll keep this short.
Gordon: Wait a minute, you do sound stereotypically old and exhausted to the bone by an aggressively stupid society gone to the maddest of dogs, but how do I really know you're from the future? I wasn't born yesterday, you know, and I'm sure as hell not falling for any old, mind-blowing future-ish looking technology, just because.
Weedy greyish greenbeard: Uh... good instinct there. Let's see, how shall we do this? Okay, I know. Get this. IEEE Standard 754 for floating point in silicon. It defines +infinity, -infinity, as well as signed zero.
Gordon: Oh yeah? Zero in both directions? I'm still not convinced.
Weedy: But wait, there's more. We've also got two types of NaN.
Gordon: You've got naan? Why didn't you send me some, in this crappy contraption? Sheesh, what's it even good for? Greedy bastards. This is still the damn seventies back here, you know. Still can't get a good naan on the left coast for love or money.
Weedy: That will change, trust me. But it's not what I meant. I mean N A N for "not a number".
Gordon: You've got two types of Not a Number? Oh, Christ, I'm not even sure I want to know.
Weedy: You probably don't.
Gordon: Well, you've already ruined my happy train, so hit me.
Weedy: Quiet NaN and signalling NaN.
Gordon: Ugh. But I've seen worse—and we're barely even begun this revolution thing...
Weedy:... think ahead, way ahead, to thousand of loosely-synchronized vector stream processors.
Gordon: Seriously? Thousands of parallel FP stream processors? What kind of atomic acreage you got there, when you come from?
Weedy: Roughly an untrimmed pinky fingernail from a large, Caucasian male, with normal appendages.
Gordon: No shit?
Weedy: Hey, it's your law—
Gordon: —extrapolation, just between you and me. Marketing people. Real asses. Yeah. But who really believes it?
Weedy: Right you are. It was tough all along. Anyway, so I've got this small question for you. Technically, your law keeps going and going—you'd hardly believe it yourself—but we're run into this small problem with thermal extraction. So the way this thing went is, we've now got tens of millions of transistors dedicated to supervising a rolling blackout—it would just be too hot to run every transistor, all the time. But we still make more and more with every shrink.
Gordon: And you came all the way back in time to ask me whether we should count transistors that are USUALLY TURNED OFF due to engineering constraints as part of my extrapolated transistor bounty? Hey, buddy, I was doing real work here! Now I'm angry again. Real angry. What lunk-headed agency or corporation would even begin to routinely count transistors that are almost always depowered?
Huh? I think if you read that program and don't know that `C++ > C` == false, you don't know how the post-increment operator works and you're not much of a C programmer.
This is the reason this entire thread is pointless.
Going all the way back to the seventies, it was a core design decision in the C language not to over-specify the semantics of expression evaluation to the degree that tricky, efficient implementations became impossible on realistic architectures.
So the decision was: we're not going to nail your nuts to the floor on side-effects until you reach a sequence point in the expression.
The C language has a fairly low cognitive burden, and likewise, the world back in the 1970s had a fairly low cognitive burden. One could afford an entire day to actually read K&R.
Then along comes C++, which follows mostly the same instincts and intuitions, with a desire to maintain a fairly hard standard of compatibility, and now we have a language with a high cognitive burden (in a world with a high cognitive burden) painstakingly built on a foundation that was already trickier than your average ninnie-poo can be bothered to properly master.
C++ is a fine programming language IF AND ONLY IF you're willing to take on the cognitive burden in the broad spirit of no compromise.
Before the first Zortech C++, I had already taken this step with the supposedly simple and trivial C language. And I just kept on going in the same spirit when C++ came along. Most of its quirks are there for a reason. It's a big commitment to know all the reasons.
And here we are in a world where the burden of mastering a tiny handful of C-language quirks lies beyond 90% of the programming population.
For you people, there's Java. Java even tried to make floating point deterministic across all architectures, present and future, so that ninny-poo legion would never have to learn the first thing about the artful compromise of sequence points and guard digits.
Java is such a nice cradle to grave language—coffin lid included.
Personally, I always wanted to be buried in a big boy box, so I put in the effort to know better. The world is so complex now, I would almost classify knowing better as a form of leprosy.
I'll grant you restaurant variety will suffer the smaller the town. That said I live in a place with a population of less than half a million and we've got several Thai places, numerous Korean joints, at least one Japanese eatery I'm aware of, and a couple Indian places.
That describes one side of a single city block in downtown Vancouver, if you add a Vietnamese Pho, a Starbucks, two decent coffee shops, and a superior French bakery.
In any scam, there's always an enabling mark with the deep pockets.
In this case, the mark is the tired sectors of button-down industry (presently known as The Swamp & its canyon suburbs) that only hire the Big Bucks Diploma from Big Bucks School, while the bargain bin is stuffed to the brim with brilliant scions of cheap alternative education.
Slashdot Mirror
I once read a book by Linda Hill that I personally found amazingly valuable, but only because I was careful not to light any matches, because her presentation was dry, dry, dry.
Because of the Indian incompetence story here on Slashdot this morning, I went to paste a link into my files, and chanced upon a past entry concerning HCL Technologies, a topic that Linda Hill has addressed in video, and soon I found myself watching a clip of hers on YouTube I hadn't seen before.
Linda Hill on empowering young sparks at HCL — July 2016
And here we have this Wikipedia article, where the unstated premise seems to be "Surprise! Derf-derf-derf, Wikipedia doesn't actually practice zero-deviation culture, despite their publicly assigned role as the plastic–pocket-protector paragon of geek dysfunction.
No, instead what we have is this: if a source is broadly flagged as tainted, it becomes open season to replace this source with a better citation wherever and whenever, without expecting significant blow back.
Isn't that leadership enough?
Is the underlying zero-deviation fixation that motivates this story just a tired strawman? Or is this derf-derf strawman meme playing to a real audience?
Well, I personally would run, run, run if I found myself in that audience, because anyone who doesn't is doomed to be soon be looking up at India as the management enlightenment movement that just passed you by with a big whoosh.
#include "no_abe_normal.h"
If you're not familiar with this convention (it appears you haven't been in this business long enough to hear the pathetic whimpers of Forma L. val d'Ation sequestered away from public shame in an attic antechamber), the "h" stands for "head".
Cutting red tape to ribbons is an intrinsically easier job than building up effective layers of regulation that prevent the public interest being bent over a barrel, while the longest of all possible rubber gloves rummages around for the better part of a trillion dollars.
Evidently, no money was harmed in the operation.
The job, as I see it, is a little harder to accomplish, once you concede that there is such a thing as effective regulation, though it's yet far from a science; science also being a discipline where time after time ones best efforts fall short, and yet one perseveres.
In the best case scenario, even after regulation becomes more of science, it will still be double hard: hard to do and hard on the ego.
Kind of makes a guy want to double down on only caring about money, setting oneself up on a lavish private beach, and watching the glorious Egos soar.
Because why? Because popular opinion has guaranteed monotonic convergence? At a guaranteed quadratic convergence rate?
Just what part of "moving target" is so hard for people to understand?
Candidate A shits her pants at the front of the boat. Everybody rushes to the back of the boat.
Candidate B begins barfing up a taco bowl. Everybody rushes back to the front of the boat.
Blather. Foam. Repeat.
The only reason Trump won is because on the day of the election, there were more disgusted voters headed to the back of the boat than the front of the boat.
Polls, especially rolling meta polls, have an intrinsic lag of two to four days. I clearly saw on the 538 graphics momentum building toward the rear bulkhead as we rounded into election day. After I extrapolated the trend a few days forward to compensate for polling lag, I was not surprised by the final outcome.
There was more than enough disgust in both directions to support either outcome. Another Billy Bush tape in the final week could easily have turned the tide. This was not a normal election where the polls were tracking a slow convergence of the undecideds. The polls were tracking a mad (and futile) scramble for the voters to distance themselves from whichever paragon of disgust was recently the most salient.
Moving target. The polls were no less instantaneously accurate than they've ever been. They just don't work very well when neither candidate has a redeeming feature, and the electorate goes into orbit around a positive pole in the complex plain.
How is this even remotely difficult to comprehend?
From where I sit, it's all bog-standard Electoral Engineering for Dummies, 101.
I'm a card-carrying Goldman Sachs conspiracy theorist. I really do think they systematically rig the market to their ultimate advantage, one ballsy five- or ten-year bamboozle after another.
Their main obstacle is that it's very hard to fool people twice with the same bullshit, so there always has to be a new fundamental disruption lurking around the next corner. I don't yet know what AI really is, but I sure know it's covered in fleece.
A sure tell is the narrative of nested boxes: each shiny thing within the dark thing is ever an order of magnitude more disruptive, more profound, and more lucrative.
The innermost thing is so brilliant, set against such a dark background (exponentially dark, by the miracle of multiple contrast-enhancing frames), it blinds you altogether.
This is one of the most ridiculous memes ever. Sound is a mechanical vibration, and Jupiter probably vibrates like hell after it ignites.
What people mean is that there is no direct transmission of physical sound waves through the vacuum of space.
Snooping Through Walls with Microwaves
Laser microphone
Or you could drop a few thumb-sized motes.
I like 2001, the Russian Solaris, and A Scanner Darkly.
Blade Runner and Alien were better than a jab in the eye.
If we further widen the net to include Space Fiction, The Wrath of Kahn rocks; while A New Hope and WALL-E both have their moments.
If we further wide the net to include any form of thematic overlap, I'd include The Right Stuff, Apollo 13, the first Back to the Future, the first Iron Man, select chunks of The Terminator franchise, RoboCop, Young Frankenstein, Dr Strangelove, and certain aspects of The Fifth Element. One might even include the sensibility of Tree of Life or Hugo.
I'll also give an honourable to The City of Lost Children, because I would actually rewatch that movie. Can't recall much of anything about the plot (not usually a good sign), but there's plenty of there there in other regards. In a pinch, I could rewatch Dune as an entertaining car wreck.
Unfortunately, much of the rest of the canon only serves to rouse my appetite without entirely beddin' her back down.
Note that I did not exclude any Spielberg movies by accident. If I had to rewatch one, it would be THX 1138. Spielberg is so sentimental, I'm soon humming Indian Love Call and wishing it would work.
On my list as the least science fiction film ever made would be the original Matrix. Perhaps the humans harvested for their cerebellar electricity was a satirical neoliberal talking point adapted from Ayn Rand.
Interview: Thomas Voss of Mir — October 2014
How is it that I never fall into the category of people described as "users"?
Does what I do for ten hours a day, every day, not fall into the semantic category of "using"? Me, and everyone like me? How do we always find ourselves filed under "a certain audience"? Well, this "certain audience" is today crying no giant room-temperature crocodile tears—neither any small, steamy gnat tears.
Here's the underlying problem: "user", as fantasized by far too many software developers, is the centerfold normalization of real womanhood.
Hollywood much?
Also, cops fight crime mainly by experience tragic science field trips in early childhood.
Here's a pro tip. If the actors are loving it, it has probably had the living Snopes kicked out of it, supposing there was any at all to begin with.
Actually, you only came second out of the small group of readers who didn't also have a second response to the day's story that was more worth the bother of typing out.
It's one of those tasks in life that naturally goes to the few.
Perhaps if I had twenty hands, and could race to be first on all my instantaneous takes at the same time I would finally embrace the joy—the joy I have witnesses so many times in others—of rushing to post the obvious.
I'm at such a huge disadvantage here. By the time I winnow my winning candidate(s) down to hands available—just that small extra delay is enough—I'm probably last off the starting blocks. Hence my decades of sour grapes at those who excel me.
The keyboard was indestructible. The case was a tank. The monochrome monitor displayed 25 rows of 80 columns, including upper and lower case letters.
Unfortunately, the IBM PC also included a few deliberate sandbags.
IBM never figured this was their last entry into the PC business. It was supposed to be a trial balloon. They had a zillion reasons to cripple the first edition, both in terms of processing power and in terms of memory expansion capacity.
By the way, did you actually use a TRS 80, Commodore PET, or Apple II? I used all three. Realistically, these all sucked for any serious purpose—except for learning the difficult art of programming the hard way.
It was just the other night I realized how starting my programming career on a TRS 80 with the notoriously unreliable tape drive influence my programming style for years to come.
No, BASIC did not ruin me. (I also picked up APL, several dialects of assembler within a year, rudimentary Pascal, some LISP, some FORTH, and most of C just as soon as I could get my hands on it.)
What did ruin me was the inability to curate a subroutine library of my favourite helper code. It just too took long to merge one chunk of code off cassette into another. (I believe the merge mode was that whatever new BASIC program you loaded just wrote right over top of any existing line numbers.) The TRS 80 was the computer I could use at school for free, which I did after school every day. Never had one at home until much later.
There was a certain kind of robustness you just didn't worry about, because every single program was pretty much home-cooked from scratch. At most, one might load something vaguely similar and then cannibalize some of the common bits.
Agile, look out—you ain't gonna need it. Every line of code ended up written in the least general way possible, so long as it sped up the code entry process.
Fortunately, I never had to use a cassette drive on an IBM PC.
On the IBM PC, I still had to multi-pass the compiler by switching floppy disks during the compile and link cycle, but that's a whole other story.
You know, the "standards of the time" included Heathkit, and Hewlett Packard (back when that still meant doing the right thing), and Tektronix, and later Compaq. The hobby computers were junk in part because everyone knew it was going to be a brisk ride. SOMEDAY SOON THERE MIGHT EVEN BE LOWER CASE. Basic economics.
Except for the IBM sandbag trick. That was old school economics, a first sour taste of something us hobbyists had not yet had to worry about.
That was the true legacy of the original IBM PC. It was the first coldly calculated, deliberate consumer diss. We all hates it forever for exactly that one thing.
Oh Lord it's hard to surprise a cynic, ...
when you're a cynic in every way
All together now. With feeling.
Or maybe, just maybe it's the entire infrastructure of automated bill-handling equipment, including vending machines?
But I'm sure it will grief some neatfreak crack dealer who feels compelled to store giant mountains of bills in his metric utility room—no wait, Canada still uses two-by-fours (for which neither dimension is a round number in any system of measurement) and 16" stud centers.
Yes, but the key step in your argument is not what you think. The exact moment you admit "rational standard" as a viable yardstick, your human intelligence steps off a plank into catastrophic failure.
QED.
Your unreasoning, intensely emotional response (larded at random with non-sequitur hot buttons) is why we all tend to be better off, long term, as competent, cool, and collected automatic control systems are further instigated.
There are plenty of competent control systems out there. By their very nature, they're not in the news on a daily basis.
This is actually calibrated behaviour. I click "submit" when the odds of noticing such a mistake fall below the bar of justifying the extra time investment.
And then sometimes I spot it right as the afterimage fades.
We have a name for people who are best served by quick decision making: we call them first responders (a category which sometimes includes pilots).
Much of the rest of the time, quick decision making is used to whip up drama rather than obtain a broad, positive outcome.
Poor slow decision: going out drinking without first planning transportation home.
Seemingly great quick decision: serving to avoid hitting the lamppost.
Poor quick decision: serving to avoid hitting the lamppost into oncoming traffic.
Happens in politics all the damn time. Yet we still keep electing these charismatic blowhards. Many of Trump's voters explicitly praised his really short approval pipeline between brainstem and mouth. Who knew healthcare could be so complicated?
Okay, I can't shut up. He's the other half.
Mike is a smart guy. I like him. But watch him blow his own foot off and gnaw upon the bloodied remains starting at about 7m50 into the following presentation:
Text By the Bay 2015: Mike Tung, Turning the Web into a Structured Database
"What about this frontal lobe?"
He wants to tackle this "next" project by some loose, unspecified analogy with recent surprising progress in computer vision and computer robotics.
Of course, how could the frontal lobe be different than dedicated perceptual and motor lobes?
If this is the standard—and this is by no means unusual among even the smartest people in the field—accurate prediction of forward progress will continue to be a cheque in the mail until the end of freaking time.
How can any thinking person in this day and age not realize that AI predictions are HARDLY EVER worth the paper they're no longer printed upon?
Successful AI prediction is five years away—always has been, and always will be.
And no, the worthless prediction of consensus merit was not "decades" but "about a decade". I've been consuming machine learning podcasts day in and day out since the end of January. AlphaGo is not infrequently mentioned.
My two favourite guests so far (not specific to Go) have been Ilya Sutskever (deep intuition; now of OpenAI, which bodes well) and David Blei (broad, nuanced perspective, with particular expertise in LDA and topic modelling).
Probably those interviews were from Talking Machines (sadly, no longer active) or The O'Reilly Data Show (the intense industry focus and reverence for Apache Spark gets on my nerves after a while, but it's reasonably meaty for the most part).
As Shuttleworth would have it:
* mainstream, n.
A technology so rooted in public acceptance that it's no longer necessary to communicate up front with the users who will most suffer from the upcoming change cycle.
I didn't leave Ubuntu because of Unity.
I left Ubuntu because no transition plan was put forward to aid me in riding out the early adoption cycle from a safe remove whereby I retained the full use of my extra monitors and the meticulous workflow depending upon these that I had painstakingly adopted over many years.
There's nothing intrinsic to mainstream that I reject, other than how becoming meanstream seems to immediately entitle the proprietor to carpet yank—without even the courtesy of a gruff dentist, who at least mutters reassuringly "this won't hurt a bit".
This has my vote as the stupidest rhetorical meme since Robinson Crusoe converts mute, g-string Friday to Christian buttlerhood.
The answer, my friend, is not enough Fridays.
Or Saturdays. Or Sundays & holidays.
This is the kind of post-hoc hagiography that fuels Ayn Rand's objectivism crap mountain.
Out of these three, it's only hard to imagine a different, yet equally grand path if RMS had never occurred. His foundational dog work on binutils invited many others to host parties they could not have otherwise managed to throw.
Due to RMS as a unique personality, we got a highly political license sooner than we would have by another probable path. This was both a strength and a liability, whose relative magnitudes are almost impossible to judge in retrospect.
Without Linus, FreeBSD either would have become far more participatory, or some variant with a far greater embrace and tolerance of messiness would have forked within two years. And since this wouldn't have embraced GPL at the system level, Gnu HURD might even have been finished, with perhaps a necessary course correction or two under mounting pressure from a large install base.
Saying that open source wouldn't exist as it now does without ESR is pretty close to saying that the internet boom of the late nineties would not have happened without George Gilder. ("George who?" all the children ask. Exactly my point.)
Certainly charismatic figures come along when the moment is ripe to crystallize the zeitgeist, but history does not record that these people have ever been in short supply (something that would become immediately obvious if that stupid scheme from Atlas Shrugged had ever been tried for real in the real world.) Generally, you can never have more than a few of these types at any given time, because the human psychology of prophets and prophecy accrues special powers to the lone voice. A solitary howling wolf is divine revelation, a million howling wolves is just a statistical noise (see again the howlingly ludicrous legacy of one Ayn Rand).
For every action there's an equal and opposite reaction. Now that half the population has levelled-up on aggressive ignorance, the standard for those who still pretend to know better only becomes that much higher.
Intel is very much one of those pretenders.
Paging Gordon Moore.
Gordon—dressed exactly like Bob Pinciotti—steps into the Tardis that just materialized outside his office, and grabs the ringing phone—midway through the fiftieth ring.
Gordon: Hello. Get this damn thing off my lawn. Can't you pre-foresee I'm trying to calculate here? Hello?
Weedy, tired voice of greybeard future: Uh, Gordon, terribly sorry for the interruption, we'll keep this short.
Gordon: Wait a minute, you do sound stereotypically old and exhausted to the bone by an aggressively stupid society gone to the maddest of dogs, but how do I really know you're from the future? I wasn't born yesterday, you know, and I'm sure as hell not falling for any old, mind-blowing future-ish looking technology, just because.
Weedy greyish greenbeard: Uh ... good instinct there. Let's see, how shall we do this? Okay, I know. Get this. IEEE Standard 754 for floating point in silicon. It defines +infinity, -infinity, as well as signed zero.
Gordon: Oh yeah? Zero in both directions? I'm still not convinced.
Weedy: But wait, there's more. We've also got two types of NaN.
Gordon: You've got naan? Why didn't you send me some, in this crappy contraption? Sheesh, what's it even good for? Greedy bastards. This is still the damn seventies back here, you know. Still can't get a good naan on the left coast for love or money.
Weedy: That will change, trust me. But it's not what I meant. I mean N A N for "not a number".
Gordon: You've got two types of Not a Number? Oh, Christ, I'm not even sure I want to know.
Weedy: You probably don't.
Gordon: Well, you've already ruined my happy train, so hit me.
Weedy: Quiet NaN and signalling NaN.
Gordon: Ugh. But I've seen worse—and we're barely even begun this revolution thing ...
Weedy: ... think ahead, way ahead, to thousand of loosely-synchronized vector stream processors.
Gordon: Seriously? Thousands of parallel FP stream processors? What kind of atomic acreage you got there, when you come from?
Weedy: Roughly an untrimmed pinky fingernail from a large, Caucasian male, with normal appendages.
Gordon: No shit?
Weedy: Hey, it's your law—
Gordon: —extrapolation, just between you and me. Marketing people. Real asses. Yeah. But who really believes it?
Weedy: Right you are. It was tough all along. Anyway, so I've got this small question for you. Technically, your law keeps going and going—you'd hardly believe it yourself—but we're run into this small problem with thermal extraction. So the way this thing went is, we've now got tens of millions of transistors dedicated to supervising a rolling blackout—it would just be too hot to run every transistor, all the time. But we still make more and more with every shrink.
Gordon: And you came all the way back in time to ask me whether we should count transistors that are USUALLY TURNED OFF due to engineering constraints as part of my extrapolated transistor bounty? Hey, buddy, I was doing real work here! Now I'm angry again. Real angry. What lunk-headed agency or corporation would even begin to routinely count transistors that are almost always depowered?
Weedy: Are you sure you want to know?
Gordon: I've got a bad feeling about this
This is the reason this entire thread is pointless.
Going all the way back to the seventies, it was a core design decision in the C language not to over-specify the semantics of expression evaluation to the degree that tricky, efficient implementations became impossible on realistic architectures.
So the decision was: we're not going to nail your nuts to the floor on side-effects until you reach a sequence point in the expression.
The C language has a fairly low cognitive burden, and likewise, the world back in the 1970s had a fairly low cognitive burden. One could afford an entire day to actually read K&R.
Then along comes C++, which follows mostly the same instincts and intuitions, with a desire to maintain a fairly hard standard of compatibility, and now we have a language with a high cognitive burden (in a world with a high cognitive burden) painstakingly built on a foundation that was already trickier than your average ninnie-poo can be bothered to properly master.
C++ is a fine programming language IF AND ONLY IF you're willing to take on the cognitive burden in the broad spirit of no compromise.
Before the first Zortech C++, I had already taken this step with the supposedly simple and trivial C language. And I just kept on going in the same spirit when C++ came along. Most of its quirks are there for a reason. It's a big commitment to know all the reasons.
And here we are in a world where the burden of mastering a tiny handful of C-language quirks lies beyond 90% of the programming population.
For you people, there's Java. Java even tried to make floating point deterministic across all architectures, present and future, so that ninny-poo legion would never have to learn the first thing about the artful compromise of sequence points and guard digits.
Java is such a nice cradle to grave language—coffin lid included.
Personally, I always wanted to be buried in a big boy box, so I put in the effort to know better. The world is so complex now, I would almost classify knowing better as a form of leprosy.
Emphasis on almost.
That describes one side of a single city block in downtown Vancouver, if you add a Vietnamese Pho, a Starbucks, two decent coffee shops, and a superior French bakery.
In any scam, there's always an enabling mark with the deep pockets.
In this case, the mark is the tired sectors of button-down industry (presently known as The Swamp & its canyon suburbs) that only hire the Big Bucks Diploma from Big Bucks School, while the bargain bin is stuffed to the brim with brilliant scions of cheap alternative education.
Milk + MOOC = mook