We're both old timers, but apparently I've kept up better than you have.
First of all, cache (and the rest of the communications fabric) is more than the half the design of a high performance CPU. Long ago now are the days where the core itself was the anchor tenant, and the rest of chip amounted to window dressing. The primacy of the core to the chip (and the ISA to the core) was the central (and false) conceit of the original RISC paradigm. If the window dressing hadn't been more important than they wished to acknowledged, there's a good chance that one of the RISC designs would have succeed in unseating Intel, long ago.
Intel was almost forced into this by accident. Starved of registers in the ISA, but having a tight read/modify/write instruction format that efficiently allowed the local stack to function as an extended register set (more efficiently than for RISC), Intel was forced to accept that their competitive foundation was memory agility (without taking this view, their ISA was the crippling liability all their RISC competitors so loudly proclaimed).
When Intel's first OOO chip came out in the mid 1990s with the first Pentium Pro there was the great day of reckoning in the RISC camp. They had all naively assumed that x86 would never achieve those kinds of performance numbers on heavy, server workloads. RISC people read the numbers and muttered under their breath "oh, shit, we're doomed". And they were right.
RISC still easily won single threaded workloads, and floating point workloads by a factor of 2:1, but on a heavily loaded server, the P6 simply never caved. Small register sets make for faster task switching. Intel had provisioned several layers of cache, with lots of internal concurrency, and an external split-transaction data bus. Departmental file and mail servers all went straight to the P6, while dedicated COTS workstations, especially engineering workstations, went in for Alpha or MIPS (you could obtain Windows NT in a variety of flavours back then). Which market would you rather have? COTS Windows NT workstations were a niche market poaching from Sun's well-defended back yard.
The press roundly thrashed the P6 because it wasn't very good at running Window 95. Talk about short-term small-minded priorities. Meanwhile, it ran 32-bit protected OSes like a champ. Most important chip in Intel's history, in my opinion, and the one true reason why x86.die.die.die never came to pass as confidently foretold by every enlightened RISC chip-head to ever awaken under a juniper bush after eating way too much majestic, desert-sunset peyote.
Except for the Pentium IV debacle, every major chip Intel has released since is basically just a P6 fitted out with a king cab and jacked suspension. AMD kindly contributed an expanded ISA with more and wider registers. Intel gradually provided wider decoders, more dispatch paths, more execution units, more in-flight instructions, better branch prediction, larger TLBs, larger caches, better cache prediction, some fancy new SIMD instructions, etc. but it was all just more of the same.
As the multicore era progressed, an actual new technology was the invisible core added to manage the thermal envelope. This was not something the P6 needed to do. There was no instruction mix that would burn the chip out, if it didn't self limit (though some especially pernicious instruction mixes would separate the men from the boys in your CPU's cooling system.)
This ushered in a new design regime where peak performance (aka bragging rights) had to compromise with performance/watt. Just because a clever design would make some subsystem faster, didn't mean that design would win (you had to also look at the thermal cost). Gradually, the performance/watt criteria became the senior cook in the kitchen.
Performance/watt is joined at the hip with your fabrication node. Modern nodes don't offer just a single transistor dimension, but multiple choices of transistor dimension, depending on whether you wish to emphasize speed or thermal efficie
China doesn't need to import foreign brain power any more than any other country does. Sheesh. (Every economy benefits from mixing the pot to some degree, no matter how plush their own labour pool.)
Have you ever looked at the highschool math education data? (In related news, New York is going to found a new law school with not a single slack-ass Jewish law professor passing over the admittance bar of the inaugural faculty.)
Nor is there any great lack of commercial opportunity with Baidu and Alibaba rivaling any company in Silicon Valley (with even more lax data collection laws—lax to such a degree that collection verges on mandatory).
And, of course, the German auto sector is asleep at the wheel, as always. It's just not in the German national character to invest in or perfect shiny new technologies.
I suspect the main reason no German engineer is ever found asleep at the wheel is because, by the time the steering wheel is completely unnecessary, the steering wheel has already been removed altogether.
Germans are pretty big on shit or get off the pot. Does it work, or doesn't it? Very direct. If it works, amputate the vestigial wheel already. And if it doesn't work, don't show your flaccid, disgraceful face in the beer hall until you've got the situation under control, like a good German should—which doesn't even need to be stated, it runs so deep.
Winston cigarettes were sponsors of such television series as The Beverly Hillbillies and The Flintstones.
The former series would show stars Buddy Ebsen, Irene Ryan, and Nancy Kulp extolling the virtues of Winstons while smoking them and reciting the jingle.
The latter series would later come under fire for advertising cigarettes on an animated series watched by many children, but Winston pulled their involvement with the series after the Pebbles Flintstone character was born in 1963.
Interesting advertising tactic: "I'll have what she's having!" [points at Daisy May "Granny" Moses]
You know, to hell with celebrity endorsements from the stacked Donna Douglas, or the studly Max Baer Jr. We've got our eyes on the prize: the old, the ugly, and the infirm.
Kulp was once described as television's most homely girl or, as one reviewer put it, possessing the "face of a shrivelled balloon, the figure of a string of spaghetti, and the voice of a bullfrog in mating season."
Others described her as tall and prim and praised her comedic skills.
So which is it? Are the Germans good engineers who generally stay on top of modern technology, or asleep at the wheel, which isn't even there any more, because they got rid of that useless appendage long ago?
In any version of the story I've ever heard, the only correct answer is both of the above.
It wasn't a big jump from Quake to Counter-Strike, and Counter-Strike (or near replicas) ended up training real troops. For real combat. With real weapons. Real wounds. Real blood. Real gravestones.
What hasn't been pointed out here is that Quake wasn't just innovative for true 3D, but had pretty good 3D physics as well. In particular, you could bounce the grenade off walls and ceilings at different angles and velocities in all kinds of unpredictable ways. I wasn't that big into rocket jumping, but I think that was a thing, too.
The physics alone basically changed the game from a finite universe (the usual stream of self-contained rabbit warrens with endless texture make-overs) into an effectively open-ended combat dynamic, even if it was open-ended in a fairly small way.
The biggest limitation with single-player Quake was the polygon budget. The game had to ration your opponents in every region. You could have maybe a dozen standard monsters, or up to three boss monsters in the current zone (if they were carefully limited in graphic complexity). The whole place ended up feeling more like a crypt frozen in time, or a haunted temple only activated by Indiana Jones daring to enter the portal. There was no sense that the place had its own daily rhythm, or economic rhythm, or even a religious or cult rhythm.
Making a big noise here rarely had any sustained effect on what woke up there.
The game wasn't horrible for alternative routes, but the alternate routes always felt explicitly designed (or accessible only to controller monkeys). You always felt the walls funnelling you back into the standard sequence.
This was not far from the era of the infamous FreeBSD GIANT lock. I'm pretty sure it was Daniel Kahneman's book where I learned that the number one predictor of success in the Israeli Air Force was the ability to quickly dial into a cognitive task and then filter out everything not relevant to that task, and then to shift focus in an instant, choose a completely different cognitive task, and again filter out everything not relevant to the new task (activating a completely different set of filters).
That's really hard to do, so most people who aren't experience in that kind of dynamic environment learn to coordinate themselves through a cognitive GIANT lock. At certain key points of the process, your behaviours are rigidly serialized. The brain is inherently SMP, but learning how to coordinate several concurrent processes all at once without them stepping on each other is extremely difficult.
It was through videogames that I sandpapered away the GIANT lock out of my own cognitive codebase. I basically played from the dual-joystick Robotron era through to Half-Life (with a sprinkle of Diablo and Age of Empires on the side). Then the games started to become more vested in faux realism (aka fantasy gratification) and more like skateboards and less like harried brain-teasers, and I quickly lost interest in the entire scene. (Rail guns and mental GIANT locks are like chocolate and peanut butter, so even Half-Life was already on the downslope; lumbering around the entire maze fumigating snipers in single-file was less enjoyable to me than the "real" work I was probably avoiding, so I just ignored the snipers and went full on into the nearest exposed melee and rarely lasted more than 40 seconds before my head popped off, yet again, no matter how well I meleed—presumably by some mono-tasking 12-year-old rail gunner who probably chuckled about how clueless and stupid the other players were to get shot over and over and over again in the same way.)
Eradicating your personal GIANT lock is exactly what you need to do if you want to become Robin Williams or Tina Fey. At the height of improv, you need your wits online across the board all at once. Railgunnery hails from the tired, gaseous province of scripted of fart jokes.
Improv, itself, is an essential skill if you want to contemplate how language really works, at its deepest level. Your highschool grammar teac
Ordered from most likely to improve your career, to least likely: * conscientiousness * openness to experience * agreeableness * extraversion * neuroticism
(In some career tracks, you'd reverse openness and agreeableness, but not computer science, in my opinion.)
I've had two experiences with pair programming, and in both cases my partner hated it, because they were extreme introverts, and they liked to burrow into the details. The first guy liked to achieve working code at a breakneck speed, and would usually cut many corners along the way. We had a problem where we had two linguistic input spaces, and we formed a conceptual superset, into which both of these could be mapped one-to-one.
And I would watch him sail into an edge case, and not really deal with the problem. And I would interject, "uh, if you do that, our conceptual mathematical function isn't one-to-one anymore". But then we would discover that he was trying to work around previous work that hadn't gotten this entirely straight, and I would want to refactor everything in sight simultaneously, and then he would slap my fingers, and perform all the refactoring in prudent single file. Except that he wanted to constantly quit, before the job was done, so I kept slapping his fingers. Probably the longest he ever endured between successful compiles in his entire professional life.
We moved more dirty rocks that day working together than we had in the previous week. The problem was that my partner couldn't see the big picture. He felt like it had been the slowest most aggravating day of his life. What I saw was a system which finally had all its formal ducks in a row, one that never gave us any trouble again, ever after. But his dopamine loop didn't work on that scale. I seriously cramped his cruddy kLOC style, and he avoided sitting beside me ever again, even though he was far better at properly pacing complex code refactoring than I'll ever be (because I tend to think about the big picture all the time) and those complementary skills made our partnership highly effective.
The second person hated it because I kept reminding him of how much work remained to be done. Every time he made a change, I would go "oh, this is going to impact that" and I would go running off to check out that on another screen. It's funny, because this time I was the one catching loose stitches, but my my way of doing this was mostly leaving source code comments in related modules about how "on such and such a day, we changed X and Y, and even though we tested U and V, we couldn't figure out how to test Z quickly". It's not that I'm low on conscientiousness, it's that my conscientiousness is always global in nature.
It's actually pretty hard, as a solitary person, to be working that hard on the central task, and to simultaneously leave so many useful breadcrumbs scattered around.
The funny thing is, you bump into those comments a few weeks or months later, and you're surprised, because in your present state of mind, it's completely slipped your consciousness that Z is even something that needs to be tested as a result of your local changes. The comment mutates from "sorry, we skipped Z under pressure in the trenches" to "OMG, I had better pay attention to Z, which had almost completely slipped my mind".
It's always a counterfactual proposition, but I'm sure that a prescient comment of this sort has saved me hours, or even half a day compared to how I might otherwise have proceeded (until the light dawned).
My experience is that people have different emotional management loops. People who cope with stress by trying not to consider the wh
It's a big jump—fraught with insecurity—to go from living at home to living on your own. Who's going to tell you you're still awesome when the best laid plans of callow fledglings screw the pooch?
Apple Inc.: I will! Me, me, me, meeeeeee!
Oh, what a complex world you weave when Fiction as a Service becomes the killer application of sugar-laced independence.
Any explanation which drops the "simple" bomb at word four sets itself up for an ankle-high bar, but that was awesome.
Left unexamined: why does Microsoft care enough about the desktop experience to bother doing this? Governments have nowhere else to go, anyway (not since Linux came down with Munich syndrome). Or is Chromebook ultimately a viable threat in the hidebound arena of the public purse? If so, that would be the real story here.
All that is true, but normally you aren't hedging into a tulip euphoria where you pony up seven figures for a strike price that winds up 14x out of the strike zone, though results may vary in Japan, where Godzilla is DH eligible.
With a proper security model, suspect Android downloads could be sandboxed with permissions to do SFA, and all the IP endpoints it initiates could be thoroughly logged.
Who else out there is breathlessly awaiting Comcast to bring us the next AlphaZero?
They say that a watched pot never boils, without mentioning that an unwatched pot can't boil, for the type of pot where the boiling point goes up as fast as the market will bear.
... network neutrality regulations would discourage construction of high-speed internet lines that telephone and cable giants are spending tens of billions of dollars to deploy...
Perhaps you could argue that. But the other side is that without network neutrality, no-one will want to use all the shitty, overpriced applications, even if all the bits arrive like a [highly metered] bat out of hell.
A tremendous amount of innovation derives directly from the old adage "too cheap to meter".
There are things you can build in the absence of friction, that simply won't fly with friction added (nor does it help that this friction is almost always capricious and unpredictable—if we can agree to term a consistent program of sudden, unannounced price jacks "unpredictable").
Innovation is fundamentally an ecological good. "Ecology" is a term of art meaning "you can try, but you'll never manage to attach oversimplified, hard numbers that don't make you look stupid the first time a non-idiot raises a hand in challenge."
The Tyranny of Metrics (2018) by Jerry Z. Muller is a good antidote, if you're stuck in a work culture that worships only that which emerges from a trance of numbers.
So along comes a guy like William P. Barr, prince of the silver lining: that the measurement of ecological goods (such as innovation) ranges from difficult to impossible is easily reconstrued by the glib as a feature, not a bug, this line of argument always working out to "physical-asset sunk-cost is king" (kings being known for their palatial spreads paved with acres of green).
Hard innovation is bone, soft innovation is flesh (so easily melted away with a bathtub full of neocon acid-reflux).
A working battery for a seven-year-old phone may be difficult.
No.
My phone is seven years old, on its original battery. I haven't beaten the crap out of it. (Turns out, there's life after the club scene.)
My original battery only became a PITA a few months ago after accidentally discharging it to 0%.
And even so, this problem exists mainly because I have crap signal strength at home (metal isolation between suites works wonders to suppress unwanted noise, but also makes for a decent Faraday cage). Most of my calls from home involve maximum Rx power (a little less if it's nice out, and I can sit in rear patio, which is still situated at an inconvenient angle, right at the midpoint between my two local towers, one of which primarily serves a small but busy international airport which happens to be situated right beside a podunk marine airport, whose substantial bay forms the third side of my cell-phone coverage triangle, which would give me a sweet direct signal, if off-shore cell towers were all the rage).
I'm pretty sure my carrier gives me a bigger signal boost outside of the heavy usage window (from what I've read, more power for me means fewer concurrent customers in total). All my bars disappear when I'm indoors on a rainy day during mid-week office hours. Thirty character text messages take over a minute to send, if they send at all. (My wife's iPhone generally manages to maintain enough of a connection to function, but even her phone sometimes drops calls, especially on rainy days.)
I have no problem with this geriatric battery lasting for as long as I ever leave the house. But indoors, for longer calls, I usually plug it into a supplemental portable battery, and sit on the comfortable couch right beside the best window. Can't wander into the kitchen though, call loss rate is over 90%.
Works for me, though I'm sure many would trade up for a better cell phone signal, and more not-so-muffled sex noises from adjoining suites.
Additionally, my battery is replaceable, which I was thinking of doing, except that this phone is seven years old, which means it's about to explode for no particular hardware reason, as it just did today.
It's a minus in that ICS devices were still around that long.
Captain Landfill FTW.
I asked before purchase if my Galaxy S II would support the forthcoming Jelly Bean (4.1) and the kiosk staff lied, straight up. It never did.
Merely, one supposes, because my carrier (Telus) couldn't be bothered. Not a single technical issue that I'm aware of—they just didn't want to run an update cycle and then have to deal with the small number of customers (probably less than 1%) who experienced a sour patch. On that assumption, it works out to one additional, annoying support call per $30,000 worth of hardware sold. Sounds cheap to me. On the flip side, that program might have interfered with all that shiny hardware's eager landfill destiny.
My old phone still basically does its original job, just fine.
So now to faff around with a rootkit upgrade (is that even a thing any more?) or faff around trying to find something new that doesn't blow goats in some new, horrific, surprising way (aka randomly removing analog audio jacks, and other eye-popping poop in the pool that mainly traces back to a certain turtleneck's turtle head).
I know we can expect a raft of posts to follow that explain the important technical and religious differences, but the vast, vast majority of the people buying these just don't care about that stuff, they want to have what is socially considered the best.
That is not a correct recap.
In all human systems, any organized contingent of mindless behaviour is progressively and relentlessly exploited until it hits a boundary condition where the mindless behaviour reaches an unstable inflection point, and system reorganizes.
Apple, of course, has a good thing going right now and somewhere deep down, they would like to keep it going for a long time. But this wish is almost completely precluded by standard game-theoretic dynamics.
Corporations famously incentivize their internal behaviour on very short clocks (often explicitly tied to the thock-thocking quarterly report metronome).
Because of this, the time value of shearing sheep is largely valued over the next four to eight quarters—anything beyond this horizon is the dreamy domain of people who don't fully understand the implicit hazards built into their stock option packages.
So there is always internal pressure to push things as far as you can possibly get away with—not to leave a single nickel for the next guy—and to play very close to the edge of a mass customer sea change, where they all suddenly decide to pay full attention to Apple's outrageous pricing model, and to reassign their mindless behaviour to some other domain, which entirely different ventures will soon detect and strive to capture (this phase usually involves an intense tug of war over precisely where this coveted mass mindlessness finally lands; it took Jobs thirty years—including more than one brush with an existential survival risk to Apple itself, and to his own position within Apple—to finally land the meaty dinosaur on Apple's lawn).
In some cases, such a as Kodak, the long view is that there is no long view, and that you should take while you can get while you can get it. Oh, you pretend that you can be the king of digital, but no-one really believes this, because these kinds of transitions rarely happen, and the reason is not accidental, but central to what makes most corporations tick on the deepest level: those toothy short-term compensation packages embedded in the management suite. Milking the existing cow takes two years. Option package achieved.
Transitioning from chemistry to computer science takes four to eight years. If it takes four years (it won't) maybe your options vest. If it takes eight years, kiss your options goodbye. As the gorilla, you're already getting good sex. You'd like to swap out the wife with a long history of being an enthusiastic partner in the sack (photo-chemistry) for a younger-model (digital sensors) because you and the wife are of an age now where hot and bothered is a game you play merely to remember what sex used to be like.
Meanwhile, your competition to gain control over the younger model (digital sensors) has a workforce entirely composed of undercompensated spring chickens, none of whom are presently getting any, who express their frustration over this state of affairs by working 16 hour days, seven days a week. You have huge advantages: budgets (with actual funds attached), infrastructure, fiscal credibility, sales channels.
But all of these things have inertia, because people on all sides are attached to what they're presently getting, and this huge hull won't prove easy to turn, if it turns at all (around the time of Midway, both the Japanese and Americans had converted hulls originally designed to become some other kind of ship into serviceable but inefficient flat tops—and for both sides, the whole deal was compromise from stem to stern and an operational PITA throughout). Plus you could piss off the wife, and then all your chips would be in the younger model basket sooner th
Fix Firefox so that it does useful things again and tons of people would be glad to switch back.
This is true, but they can't, because Firefox elected instead to embrace the Chrome add-on model.
Extensions for Firefox are built using the WebExtensions API, a cross-browser system for developing extensions. To a large extent the system is compatible with the extension API supported by Google Chrome and Opera and the W3C Draft Community Group.
Once upon a time, Firefox had a superior extensibility model, on the primary criteria of actual extensibility, and now they don't.
I understand that there are complex issues here, but it's not a good look when you capitulate to those complexities to such a degree that you've got nothing left to substantially set yourself apart, other than that you're not actually the other guy (even though you dance to exactly the same limited API).
You can't even go to the old discussions of the old add-ons for opinions about various features, because those are no longer on line now. Plus there used to be discussion there about to cope in the barren new world (to which I'm still not totally adapted).
I guess Mozilla doesn't see the need to keep unflattering history alive.
Have you noticed that all these "AI systems" play Chess and Go? Very odd.
Thank you for spamming the entire thread with your imperceptive and unenlightened comments.
There's nothing odd about the choice of chess and Go whatsoever. Humanity has thousands of years of experience with these games. We know they aren't trivial, and we know they're not so complex that we can't understand progress, when we see it.
Additionally, the large literature of expert games was a useful hand-rail between hand-crafted and fully autonomous.
Quite apart from the neural network portion, Monte Carlo tree search (MCTS) is a fundamental algorithm in computer science, and this work demonstrates that MCTS is ready for prime time, having defeated from scratch exceptionally strong chess programs that have been painstaking hand-tuned over five decades and hundreds of man years. MCTS exists within the large and growing theory of multi-armed bandit problems. These are fundamental problems in many important industries (such as drug discovery, to name just one).
Recurrent self-learning is another important algorithm in computer science and machine learning.
And finally, the neural network portion is far closer to the human brain than the vast majority of algorithms used in computing. Without any human instruction, these neural networks are learning to detect patterns of almost arbitrary complexity (so long as they seem to help in winning games).
I was reading Galileo in the original last night (English translation, but his original prose). He knew about Kepler, but wasn't sold on elliptical motion. Then he carefully observes four previously unknown moons of Jupiter and correctly determines that they can't all be in circular orbits. The word he used (in English translation) was "oval". But he still didn't choose to accept Kepler's work (apparently, he felt that Kepler's ellipse and his oval were not the same thing).
Galileo was a giant in the history of science. But still a little wooden headed on a few points, nonetheless.
I think Odd Buster Spamalot is nuts to criticise Galileo for not being Newton. Only because Galileo sorted enough of the fundamentals out in the first place (about the proper concerns and methods of science), was it even possible for Newton to become Newton (and he knew it, himself, and he's famous for having said so).
The computers we now apply to neural networks are roughly a factor of one billion times more powerful than the computers of the 1960s (thirty doublings over 45 years gets you there at the traditional pace of Moore's law).
You could complain that neural networks are only good at this one thing, but actually no: they are now state of the art in image classification (IC), speaker-independent large-vocabulary continuous speech recognition (CSR), and machine translation (MT), as well. All of these endeavours also date back to the 1960s, and have thousands of man-years of deep research behind them. Then DNNs come along, finally on a sufficiently powerful computer, with a few small tweaks to the algorithms, and simply cleans up the state of the art with nothing more than a small team of graduate students doing a quick project within the scope of their degree program to push this along (the subsequent move to industrial scale was immediate and brisk). Traditional MT research programs would have hundreds of professional researchers, slaving away for decades, at least, and never accomplished as much.
We're all of ten years away now from the day where no competent doctor ever reads an x-ray (or other radiological image) without computer assistance (definitely including a powerful NN component).
Watson was a bit idiotic, right from the beginning. The problem was Jeopardy, itself, which was always rather facile in the nature of the questions asked, and fundamentally more a test of ridiculously wide and shallow
And before you go there, Intel was the first company to lie about node size.
In the era where there was a fairly broad consensus about how to advertise a process node size, there was only one transistor.
But then the power envelope started to matter, and signalling distance increasingly started to matter, and then there wasn't just one transistor any more, but various transistors tuned for different roles, and some transistors are contributing to peak clock frequency, while others are contributing to total integration density, while still others contribute to maintaining an acceptable power envelope.
A good process is one that enables you to strike a balance among all these trade-offs, while ultimately achieving good yield, and without requiring undue process steps.
It ceased to be a world with a single node size dimension to rule them all.
It's pretty naive to think you're going to abstract a process with this many inherent tradeoffs into a single scalar figure of merit.
Some of the same crowd is big on the idea of one master sentencing guideline to rule them all: sexual crime, white collar embezzlement, violent crime, intimidation and extortion can all be ratified into a single, unified sentencing pyramid that everyone finds just and coherent.
But sentencing is complex, and there's a deterrent component, and different kinds of criminals responding differently to the deterrence scale (few white collar criminals respond to a sentencing scale that goes from zero to infinity in one quick, irrevocable step by throwing caution to the wind and killing any cop who dares to apprehend them).
The criminal deterrence system works best when you compare nearby things to nearby things, and attempt to smooth out any ruthless punitive escalations over trivial differences (why did the man down the hall get half the sentence I got, for basically the same crime?) That kind of things throws yet more fuel on the rumour (always popular to begin with) that the law is arbitrary, capricious, vindictive, and personal (yes, it has been known to happen for real).
At the end of the day, "deserving" only deserves one small wedge of the pie. Appropriate deterrence curves also have a lot to do with it (if too steep for the wrong class of criminals, cops die) and often local fairness matters more on the ground (see Bernhard Riemann) than global fairness, for which no uniform societal adjudication is possible, in any case; the one true punishment pyramid is a unicorn grail.
Uniform adjudication: sometimes a convenient fiction, but never an altogether truthful fiction.
Closer to home than cop killing, I buy shirts all the time that are an X or two fatter than I am. I have long arms, and there's no hope whatsoever for the arms unless the shoulders fit. I also have broad shoulders, and generally it turns out there's no hope for whatsoever for the shoulders to fit properly unless the man boobs are cavernous and the bat wings drape under my armpits like I'm a newly famished member of the undead. But at least my sleeves cease to grip my wrists four inches above the wrist joint, and more still, every time I raise an elbow.
All these shirts have one basically on size designation inscribed on the tag on the back of the neck (along with the odd sprinkle of T). Good to know all the tailors are on the same page, and no-one is lying their ass off: all the XLs are mildly cavernous, and all the XXLs are generously cavernous. No worries, the label never lies.
But it still doesn't help me fit a shirt, because that one reified number was bullshit to begin with. (The only shirt to ever fit me off the rack was an Italian-tapered small x LT.) American tailoring and Italian tailoring are not entirely commensurate. I don't specifically know who was to blame for this, but probably the Italians were first to lie.
That still means you have to dig it out, especially annoying in the winter. I like my keyless entry because as stated, I can just walk up to the door and press a button (on the door) and it unlocks without having to dig the key out of my pocket or jacket.
Then you deserve to have your radio signals amplified and replayed. Just because you find it annoying doesn't mean a terrible engineering solution somehow becomes a good idea.
I suppose you could two-factor with a fingerprint reader on your car door. Should work great until that surprise -50 degrees F event (climate isn't what it used to be). Then you be standing outside your car in some cold, remote place (remote from your fireplace), wishing your fob had a convenient and reliable button, rather than your car door having an erratic fingerprint reader, that reads blue fingers even less well than mostly sometimes.
Yes, real estate isn't a liquid asset, but over time, it's almost always an investment that significantly increases in value.
Historical norms should be taken with a grain of salt. Additionally, certain gambling games always go up, until the black swan comes along and multiplies by zero.
How many 2008s can the average person soak up in one lifetime?
I find Taleb to be more aggressive than necessary, and I quickly get to the point of "haven't you belaboured this enough, already?" and then I get onto the Internet, and I think "well, maybe not too belaboured in the larger context".
I tend to regard "almost always" as a dead giveaway that someone is sweeping survival risk under the handy carpet of historical norms.
There's even a worse theorem here. It's pretty much a law of human nature that if you can convince a large enough group of people to onboard some substantial exposure exclusively because it "almost always" goes well, market insiders will be quick to engineer a shocking reversal on that long-standing historical trend.
Big Finance can engineer almost any outcome these days, it just needs to be big enough to justify a systemic assault (with enough sheep on the wrong side). This kind of engineering at scale doesn't come cheap, so it tends to be carefully paced (also spacing out due to the undue-regulation minimization sheep-shear spacing theorem).
Gone are the days where it's safe to make major investments facing backwards in time.
The fastest machine you can get from Apple is some sort of iMac which costs over 7 grand.
I just don't get the love for quoting the stupid-money price ceiling.
Did speaker wire teach us nothing? Did deoxygenated digital interconnects teach us nothing? Chalk one up for "you can lead a horse to water, but you can't make him drink".
The largest price you can possibly pay is of concern to criminals, cretins, scoundrels, the insane, and marketing departments only (to spell it out in full retinal redundancy).
Anyone seeking ways Facebook abuses its dominance will lose a talking point.
Since we're into sports metaphors (life is a win/lose proposition) let's flesh that out a bit.
Is it like the Edmonton Oilers having to play a game without Connor McDavid dressed for the game (as happened recently) or is it more like the Red Army from the 1970s and 1980s having to scrape up a temporary replacement from the whole of Soviet sphere, on an express Aeroflot ticket you can't refuse?
Negative Facebook talking points comes with a depth chart attached that spans at least eleven full time zones.
I really don't "scratch one" is going to make an appreciable dent.
It's not that Google has anything in particular against paying their contract workers better.
But Brin and Page have long bucket lists of important ways they'd like to change the world, and many of these simply can't be undertaken until Alphabet reaches a $2 trillion market cap.
Pros and cons, as with all things.
As they say, you can't raise an omelette to the stratosphere without breaking eggs.
Slashdot Mirror
We're both old timers, but apparently I've kept up better than you have.
First of all, cache (and the rest of the communications fabric) is more than the half the design of a high performance CPU. Long ago now are the days where the core itself was the anchor tenant, and the rest of chip amounted to window dressing. The primacy of the core to the chip (and the ISA to the core) was the central (and false) conceit of the original RISC paradigm. If the window dressing hadn't been more important than they wished to acknowledged, there's a good chance that one of the RISC designs would have succeed in unseating Intel, long ago.
Intel was almost forced into this by accident. Starved of registers in the ISA, but having a tight read/modify/write instruction format that efficiently allowed the local stack to function as an extended register set (more efficiently than for RISC), Intel was forced to accept that their competitive foundation was memory agility (without taking this view, their ISA was the crippling liability all their RISC competitors so loudly proclaimed).
When Intel's first OOO chip came out in the mid 1990s with the first Pentium Pro there was the great day of reckoning in the RISC camp. They had all naively assumed that x86 would never achieve those kinds of performance numbers on heavy, server workloads. RISC people read the numbers and muttered under their breath "oh, shit, we're doomed". And they were right.
RISC still easily won single threaded workloads, and floating point workloads by a factor of 2:1, but on a heavily loaded server, the P6 simply never caved. Small register sets make for faster task switching. Intel had provisioned several layers of cache, with lots of internal concurrency, and an external split-transaction data bus. Departmental file and mail servers all went straight to the P6, while dedicated COTS workstations, especially engineering workstations, went in for Alpha or MIPS (you could obtain Windows NT in a variety of flavours back then). Which market would you rather have? COTS Windows NT workstations were a niche market poaching from Sun's well-defended back yard.
The press roundly thrashed the P6 because it wasn't very good at running Window 95. Talk about short-term small-minded priorities. Meanwhile, it ran 32-bit protected OSes like a champ. Most important chip in Intel's history, in my opinion, and the one true reason why x86.die.die.die never came to pass as confidently foretold by every enlightened RISC chip-head to ever awaken under a juniper bush after eating way too much majestic, desert-sunset peyote.
Except for the Pentium IV debacle, every major chip Intel has released since is basically just a P6 fitted out with a king cab and jacked suspension. AMD kindly contributed an expanded ISA with more and wider registers. Intel gradually provided wider decoders, more dispatch paths, more execution units, more in-flight instructions, better branch prediction, larger TLBs, larger caches, better cache prediction, some fancy new SIMD instructions, etc. but it was all just more of the same.
As the multicore era progressed, an actual new technology was the invisible core added to manage the thermal envelope. This was not something the P6 needed to do. There was no instruction mix that would burn the chip out, if it didn't self limit (though some especially pernicious instruction mixes would separate the men from the boys in your CPU's cooling system.)
This ushered in a new design regime where peak performance (aka bragging rights) had to compromise with performance/watt. Just because a clever design would make some subsystem faster, didn't mean that design would win (you had to also look at the thermal cost). Gradually, the performance/watt criteria became the senior cook in the kitchen.
Performance/watt is joined at the hip with your fabrication node. Modern nodes don't offer just a single transistor dimension, but multiple choices of transistor dimension, depending on whether you wish to emphasize speed or thermal efficie
China doesn't need to import foreign brain power any more than any other country does. Sheesh. (Every economy benefits from mixing the pot to some degree, no matter how plush their own labour pool.)
Have you ever looked at the highschool math education data? (In related news, New York is going to found a new law school with not a single slack-ass Jewish law professor passing over the admittance bar of the inaugural faculty.)
Nor is there any great lack of commercial opportunity with Baidu and Alibaba rivaling any company in Silicon Valley (with even more lax data collection laws—lax to such a degree that collection verges on mandatory).
And, of course, the German auto sector is asleep at the wheel, as always. It's just not in the German national character to invest in or perfect shiny new technologies.
I suspect the main reason no German engineer is ever found asleep at the wheel is because, by the time the steering wheel is completely unnecessary, the steering wheel has already been removed altogether.
Germans are pretty big on shit or get off the pot. Does it work, or doesn't it? Very direct. If it works, amputate the vestigial wheel already. And if it doesn't work, don't show your flaccid, disgraceful face in the beer hall until you've got the situation under control, like a good German should—which doesn't even need to be stated, it runs so deep.
Winston tastes good like a cigarette should
Interesting advertising tactic: "I'll have what she's having!" [points at Daisy May "Granny" Moses]
You know, to hell with celebrity endorsements from the stacked Donna Douglas, or the studly Max Baer Jr. We've got our eyes on the prize: the old, the ugly, and the infirm.
So which is it? Are the Germans good engineers who generally stay on top of modern technology, or asleep at the wheel, which isn't even there any more, because they got rid of that useless appendage long ago?
In any version of the story I've ever heard, the only correct answer is both of the above.
It wasn't a big jump from Quake to Counter-Strike, and Counter-Strike (or near replicas) ended up training real troops. For real combat. With real weapons. Real wounds. Real blood. Real gravestones.
What hasn't been pointed out here is that Quake wasn't just innovative for true 3D, but had pretty good 3D physics as well. In particular, you could bounce the grenade off walls and ceilings at different angles and velocities in all kinds of unpredictable ways. I wasn't that big into rocket jumping, but I think that was a thing, too.
The physics alone basically changed the game from a finite universe (the usual stream of self-contained rabbit warrens with endless texture make-overs) into an effectively open-ended combat dynamic, even if it was open-ended in a fairly small way.
The biggest limitation with single-player Quake was the polygon budget. The game had to ration your opponents in every region. You could have maybe a dozen standard monsters, or up to three boss monsters in the current zone (if they were carefully limited in graphic complexity). The whole place ended up feeling more like a crypt frozen in time, or a haunted temple only activated by Indiana Jones daring to enter the portal. There was no sense that the place had its own daily rhythm, or economic rhythm, or even a religious or cult rhythm.
Making a big noise here rarely had any sustained effect on what woke up there.
The game wasn't horrible for alternative routes, but the alternate routes always felt explicitly designed (or accessible only to controller monkeys). You always felt the walls funnelling you back into the standard sequence.
This was not far from the era of the infamous FreeBSD GIANT lock. I'm pretty sure it was Daniel Kahneman's book where I learned that the number one predictor of success in the Israeli Air Force was the ability to quickly dial into a cognitive task and then filter out everything not relevant to that task, and then to shift focus in an instant, choose a completely different cognitive task, and again filter out everything not relevant to the new task (activating a completely different set of filters).
That's really hard to do, so most people who aren't experience in that kind of dynamic environment learn to coordinate themselves through a cognitive GIANT lock. At certain key points of the process, your behaviours are rigidly serialized. The brain is inherently SMP, but learning how to coordinate several concurrent processes all at once without them stepping on each other is extremely difficult.
It was through videogames that I sandpapered away the GIANT lock out of my own cognitive codebase. I basically played from the dual-joystick Robotron era through to Half-Life (with a sprinkle of Diablo and Age of Empires on the side). Then the games started to become more vested in faux realism (aka fantasy gratification) and more like skateboards and less like harried brain-teasers, and I quickly lost interest in the entire scene. (Rail guns and mental GIANT locks are like chocolate and peanut butter, so even Half-Life was already on the downslope; lumbering around the entire maze fumigating snipers in single-file was less enjoyable to me than the "real" work I was probably avoiding, so I just ignored the snipers and went full on into the nearest exposed melee and rarely lasted more than 40 seconds before my head popped off, yet again, no matter how well I meleed—presumably by some mono-tasking 12-year-old rail gunner who probably chuckled about how clueless and stupid the other players were to get shot over and over and over again in the same way.)
Eradicating your personal GIANT lock is exactly what you need to do if you want to become Robin Williams or Tina Fey. At the height of improv, you need your wits online across the board all at once. Railgunnery hails from the tired, gaseous province of scripted of fart jokes.
Improv, itself, is an essential skill if you want to contemplate how language really works, at its deepest level. Your highschool grammar teac
Professional psychometricians don't take Myers–Briggs all that seriously these days.
Most of these systems are beloved of consultants and sages because they're good conversation starters.
The best supported factorizations are the Big Five personality traits (often seen as OCEAN).
Ordered from most likely to improve your career, to least likely:
* conscientiousness
* openness to experience
* agreeableness
* extraversion
* neuroticism
(In some career tracks, you'd reverse openness and agreeableness, but not computer science, in my opinion.)
I've had two experiences with pair programming, and in both cases my partner hated it, because they were extreme introverts, and they liked to burrow into the details. The first guy liked to achieve working code at a breakneck speed, and would usually cut many corners along the way. We had a problem where we had two linguistic input spaces, and we formed a conceptual superset, into which both of these could be mapped one-to-one.
And I would watch him sail into an edge case, and not really deal with the problem. And I would interject, "uh, if you do that, our conceptual mathematical function isn't one-to-one anymore". But then we would discover that he was trying to work around previous work that hadn't gotten this entirely straight, and I would want to refactor everything in sight simultaneously, and then he would slap my fingers, and perform all the refactoring in prudent single file. Except that he wanted to constantly quit, before the job was done, so I kept slapping his fingers. Probably the longest he ever endured between successful compiles in his entire professional life.
We moved more dirty rocks that day working together than we had in the previous week. The problem was that my partner couldn't see the big picture. He felt like it had been the slowest most aggravating day of his life. What I saw was a system which finally had all its formal ducks in a row, one that never gave us any trouble again, ever after. But his dopamine loop didn't work on that scale. I seriously cramped his cruddy kLOC style, and he avoided sitting beside me ever again, even though he was far better at properly pacing complex code refactoring than I'll ever be (because I tend to think about the big picture all the time) and those complementary skills made our partnership highly effective.
The second person hated it because I kept reminding him of how much work remained to be done. Every time he made a change, I would go "oh, this is going to impact that" and I would go running off to check out that on another screen. It's funny, because this time I was the one catching loose stitches, but my my way of doing this was mostly leaving source code comments in related modules about how "on such and such a day, we changed X and Y, and even though we tested U and V, we couldn't figure out how to test Z quickly". It's not that I'm low on conscientiousness, it's that my conscientiousness is always global in nature.
It's actually pretty hard, as a solitary person, to be working that hard on the central task, and to simultaneously leave so many useful breadcrumbs scattered around.
The funny thing is, you bump into those comments a few weeks or months later, and you're surprised, because in your present state of mind, it's completely slipped your consciousness that Z is even something that needs to be tested as a result of your local changes. The comment mutates from "sorry, we skipped Z under pressure in the trenches" to "OMG, I had better pay attention to Z, which had almost completely slipped my mind".
It's always a counterfactual proposition, but I'm sure that a prescient comment of this sort has saved me hours, or even half a day compared to how I might otherwise have proceeded (until the light dawned).
My experience is that people have different emotional management loops. People who cope with stress by trying not to consider the wh
Ah, the world-famous escape hatch "just", wherein "you get what you pay for" can pound sand, no questions asked.
It's a big jump—fraught with insecurity—to go from living at home to living on your own. Who's going to tell you you're still awesome when the best laid plans of callow fledglings screw the pooch?
Apple Inc.: I will! Me, me, me, meeeeeee!
Oh, what a complex world you weave when Fiction as a Service becomes the killer application of sugar-laced independence.
Any explanation which drops the "simple" bomb at word four sets itself up for an ankle-high bar, but that was awesome.
Left unexamined: why does Microsoft care enough about the desktop experience to bother doing this? Governments have nowhere else to go, anyway (not since Linux came down with Munich syndrome). Or is Chromebook ultimately a viable threat in the hidebound arena of the public purse? If so, that would be the real story here.
All that is true, but normally you aren't hedging into a tulip euphoria where you pony up seven figures for a strike price that winds up 14x out of the strike zone, though results may vary in Japan, where Godzilla is DH eligible.
With a proper security model, suspect Android downloads could be sandboxed with permissions to do SFA, and all the IP endpoints it initiates could be thoroughly logged.
Who else out there is breathlessly awaiting Comcast to bring us the next AlphaZero?
They say that a watched pot never boils, without mentioning that an unwatched pot can't boil, for the type of pot where the boiling point goes up as fast as the market will bear.
Funny, I would have guessed that the polar opposite of network neutrality was chilling effect.
Perhaps you could argue that. But the other side is that without network neutrality, no-one will want to use all the shitty, overpriced applications, even if all the bits arrive like a [highly metered] bat out of hell.
A tremendous amount of innovation derives directly from the old adage "too cheap to meter".
There are things you can build in the absence of friction, that simply won't fly with friction added (nor does it help that this friction is almost always capricious and unpredictable—if we can agree to term a consistent program of sudden, unannounced price jacks "unpredictable").
Innovation is fundamentally an ecological good. "Ecology" is a term of art meaning "you can try, but you'll never manage to attach oversimplified, hard numbers that don't make you look stupid the first time a non-idiot raises a hand in challenge."
The Tyranny of Metrics (2018) by Jerry Z. Muller is a good antidote, if you're stuck in a work culture that worships only that which emerges from a trance of numbers.
So along comes a guy like William P. Barr, prince of the silver lining: that the measurement of ecological goods (such as innovation) ranges from difficult to impossible is easily reconstrued by the glib as a feature, not a bug, this line of argument always working out to "physical-asset sunk-cost is king" (kings being known for their palatial spreads paved with acres of green).
Hard innovation is bone, soft innovation is flesh (so easily melted away with a bathtub full of neocon acid-reflux).
No.
My phone is seven years old, on its original battery. I haven't beaten the crap out of it. (Turns out, there's life after the club scene.)
My original battery only became a PITA a few months ago after accidentally discharging it to 0%.
And even so, this problem exists mainly because I have crap signal strength at home (metal isolation between suites works wonders to suppress unwanted noise, but also makes for a decent Faraday cage). Most of my calls from home involve maximum Rx power (a little less if it's nice out, and I can sit in rear patio, which is still situated at an inconvenient angle, right at the midpoint between my two local towers, one of which primarily serves a small but busy international airport which happens to be situated right beside a podunk marine airport, whose substantial bay forms the third side of my cell-phone coverage triangle, which would give me a sweet direct signal, if off-shore cell towers were all the rage).
I'm pretty sure my carrier gives me a bigger signal boost outside of the heavy usage window (from what I've read, more power for me means fewer concurrent customers in total). All my bars disappear when I'm indoors on a rainy day during mid-week office hours. Thirty character text messages take over a minute to send, if they send at all. (My wife's iPhone generally manages to maintain enough of a connection to function, but even her phone sometimes drops calls, especially on rainy days.)
I have no problem with this geriatric battery lasting for as long as I ever leave the house. But indoors, for longer calls, I usually plug it into a supplemental portable battery, and sit on the comfortable couch right beside the best window. Can't wander into the kitchen though, call loss rate is over 90%.
Works for me, though I'm sure many would trade up for a better cell phone signal, and more not-so-muffled sex noises from adjoining suites.
Additionally, my battery is replaceable, which I was thinking of doing, except that this phone is seven years old, which means it's about to explode for no particular hardware reason, as it just did today.
Captain Landfill FTW.
I asked before purchase if my Galaxy S II would support the forthcoming Jelly Bean (4.1) and the kiosk staff lied, straight up. It never did.
Merely, one supposes, because my carrier (Telus) couldn't be bothered. Not a single technical issue that I'm aware of—they just didn't want to run an update cycle and then have to deal with the small number of customers (probably less than 1%) who experienced a sour patch. On that assumption, it works out to one additional, annoying support call per $30,000 worth of hardware sold. Sounds cheap to me. On the flip side, that program might have interfered with all that shiny hardware's eager landfill destiny.
My old phone still basically does its original job, just fine.
So now to faff around with a rootkit upgrade (is that even a thing any more?) or faff around trying to find something new that doesn't blow goats in some new, horrific, surprising way (aka randomly removing analog audio jacks, and other eye-popping poop in the pool that mainly traces back to a certain turtleneck's turtle head).
That is not a correct recap.
In all human systems, any organized contingent of mindless behaviour is progressively and relentlessly exploited until it hits a boundary condition where the mindless behaviour reaches an unstable inflection point, and system reorganizes.
Apple, of course, has a good thing going right now and somewhere deep down, they would like to keep it going for a long time. But this wish is almost completely precluded by standard game-theoretic dynamics.
Corporations famously incentivize their internal behaviour on very short clocks (often explicitly tied to the thock-thocking quarterly report metronome).
Because of this, the time value of shearing sheep is largely valued over the next four to eight quarters—anything beyond this horizon is the dreamy domain of people who don't fully understand the implicit hazards built into their stock option packages.
So there is always internal pressure to push things as far as you can possibly get away with—not to leave a single nickel for the next guy—and to play very close to the edge of a mass customer sea change, where they all suddenly decide to pay full attention to Apple's outrageous pricing model, and to reassign their mindless behaviour to some other domain, which entirely different ventures will soon detect and strive to capture (this phase usually involves an intense tug of war over precisely where this coveted mass mindlessness finally lands; it took Jobs thirty years—including more than one brush with an existential survival risk to Apple itself, and to his own position within Apple—to finally land the meaty dinosaur on Apple's lawn).
In some cases, such a as Kodak, the long view is that there is no long view, and that you should take while you can get while you can get it. Oh, you pretend that you can be the king of digital, but no-one really believes this, because these kinds of transitions rarely happen, and the reason is not accidental, but central to what makes most corporations tick on the deepest level: those toothy short-term compensation packages embedded in the management suite. Milking the existing cow takes two years. Option package achieved.
Transitioning from chemistry to computer science takes four to eight years. If it takes four years (it won't) maybe your options vest. If it takes eight years, kiss your options goodbye. As the gorilla, you're already getting good sex. You'd like to swap out the wife with a long history of being an enthusiastic partner in the sack (photo-chemistry) for a younger-model (digital sensors) because you and the wife are of an age now where hot and bothered is a game you play merely to remember what sex used to be like.
Meanwhile, your competition to gain control over the younger model (digital sensors) has a workforce entirely composed of undercompensated spring chickens, none of whom are presently getting any, who express their frustration over this state of affairs by working 16 hour days, seven days a week. You have huge advantages: budgets (with actual funds attached), infrastructure, fiscal credibility, sales channels.
But all of these things have inertia, because people on all sides are attached to what they're presently getting, and this huge hull won't prove easy to turn, if it turns at all (around the time of Midway, both the Japanese and Americans had converted hulls originally designed to become some other kind of ship into serviceable but inefficient flat tops—and for both sides, the whole deal was compromise from stem to stern and an operational PITA throughout). Plus you could piss off the wife, and then all your chips would be in the younger model basket sooner th
This is true, but they can't, because Firefox elected instead to embrace the Chrome add-on model.
Once upon a time, Firefox had a superior extensibility model, on the primary criteria of actual extensibility, and now they don't.
I understand that there are complex issues here, but it's not a good look when you capitulate to those complexities to such a degree that you've got nothing left to substantially set yourself apart, other than that you're not actually the other guy (even though you dance to exactly the same limited API).
You can't even go to the old discussions of the old add-ons for opinions about various features, because those are no longer on line now. Plus there used to be discussion there about to cope in the barren new world (to which I'm still not totally adapted).
I guess Mozilla doesn't see the need to keep unflattering history alive.
Thank you for spamming the entire thread with your imperceptive and unenlightened comments.
There's nothing odd about the choice of chess and Go whatsoever. Humanity has thousands of years of experience with these games. We know they aren't trivial, and we know they're not so complex that we can't understand progress, when we see it.
Additionally, the large literature of expert games was a useful hand-rail between hand-crafted and fully autonomous.
Quite apart from the neural network portion, Monte Carlo tree search (MCTS) is a fundamental algorithm in computer science, and this work demonstrates that MCTS is ready for prime time, having defeated from scratch exceptionally strong chess programs that have been painstaking hand-tuned over five decades and hundreds of man years. MCTS exists within the large and growing theory of multi-armed bandit problems. These are fundamental problems in many important industries (such as drug discovery, to name just one).
Multi-armed bandit
Recurrent self-learning is another important algorithm in computer science and machine learning.
And finally, the neural network portion is far closer to the human brain than the vast majority of algorithms used in computing. Without any human instruction, these neural networks are learning to detect patterns of almost arbitrary complexity (so long as they seem to help in winning games).
I was reading Galileo in the original last night (English translation, but his original prose). He knew about Kepler, but wasn't sold on elliptical motion. Then he carefully observes four previously unknown moons of Jupiter and correctly determines that they can't all be in circular orbits. The word he used (in English translation) was "oval". But he still didn't choose to accept Kepler's work (apparently, he felt that Kepler's ellipse and his oval were not the same thing).
Galileo was a giant in the history of science. But still a little wooden headed on a few points, nonetheless.
I think Odd Buster Spamalot is nuts to criticise Galileo for not being Newton. Only because Galileo sorted enough of the fundamentals out in the first place (about the proper concerns and methods of science), was it even possible for Newton to become Newton (and he knew it, himself, and he's famous for having said so).
The computers we now apply to neural networks are roughly a factor of one billion times more powerful than the computers of the 1960s (thirty doublings over 45 years gets you there at the traditional pace of Moore's law).
You could complain that neural networks are only good at this one thing, but actually no: they are now state of the art in image classification (IC), speaker-independent large-vocabulary continuous speech recognition (CSR), and machine translation (MT), as well. All of these endeavours also date back to the 1960s, and have thousands of man-years of deep research behind them. Then DNNs come along, finally on a sufficiently powerful computer, with a few small tweaks to the algorithms, and simply cleans up the state of the art with nothing more than a small team of graduate students doing a quick project within the scope of their degree program to push this along (the subsequent move to industrial scale was immediate and brisk). Traditional MT research programs would have hundreds of professional researchers, slaving away for decades, at least, and never accomplished as much.
We're all of ten years away now from the day where no competent doctor ever reads an x-ray (or other radiological image) without computer assistance (definitely including a powerful NN component).
Watson was a bit idiotic, right from the beginning. The problem was Jeopardy, itself, which was always rather facile in the nature of the questions asked, and fundamentally more a test of ridiculously wide and shallow
No idea what happened there. Truly spastic.
In the era where there was a fairly broad consensus about how to advertise a process node size, there was only one transistor.
But then the power envelope started to matter, and signalling distance increasingly started to matter, and then there wasn't just one transistor any more, but various transistors tuned for different roles, and some transistors are contributing to peak clock frequency, while others are contributing to total integration density, while still others contribute to maintaining an acceptable power envelope.
A good process is one that enables you to strike a balance among all these trade-offs, while ultimately achieving good yield, and without requiring undue process steps.
It ceased to be a world with a single node size dimension to rule them all.
It's pretty naive to think you're going to abstract a process with this many inherent tradeoffs into a single scalar figure of merit.
Some of the same crowd is big on the idea of one master sentencing guideline to rule them all: sexual crime, white collar embezzlement, violent crime, intimidation and extortion can all be ratified into a single, unified sentencing pyramid that everyone finds just and coherent.
But sentencing is complex, and there's a deterrent component, and different kinds of criminals responding differently to the deterrence scale (few white collar criminals respond to a sentencing scale that goes from zero to infinity in one quick, irrevocable step by throwing caution to the wind and killing any cop who dares to apprehend them).
The criminal deterrence system works best when you compare nearby things to nearby things, and attempt to smooth out any ruthless punitive escalations over trivial differences (why did the man down the hall get half the sentence I got, for basically the same crime?) That kind of things throws yet more fuel on the rumour (always popular to begin with) that the law is arbitrary, capricious, vindictive, and personal (yes, it has been known to happen for real).
At the end of the day, "deserving" only deserves one small wedge of the pie. Appropriate deterrence curves also have a lot to do with it (if too steep for the wrong class of criminals, cops die) and often local fairness matters more on the ground (see Bernhard Riemann) than global fairness, for which no uniform societal adjudication is possible, in any case; the one true punishment pyramid is a unicorn grail.
Uniform adjudication: sometimes a convenient fiction, but never an altogether truthful fiction.
Closer to home than cop killing, I buy shirts all the time that are an X or two fatter than I am. I have long arms, and there's no hope whatsoever for the arms unless the shoulders fit. I also have broad shoulders, and generally it turns out there's no hope for whatsoever for the shoulders to fit properly unless the man boobs are cavernous and the bat wings drape under my armpits like I'm a newly famished member of the undead. But at least my sleeves cease to grip my wrists four inches above the wrist joint, and more still, every time I raise an elbow.
All these shirts have one basically on size designation inscribed on the tag on the back of the neck (along with the odd sprinkle of T). Good to know all the tailors are on the same page, and no-one is lying their ass off: all the XLs are mildly cavernous, and all the XXLs are generously cavernous. No worries, the label never lies.
But it still doesn't help me fit a shirt, because that one reified number was bullshit to begin with. (The only shirt to ever fit me off the rack was an Italian-tapered small x LT.) American tailoring and Italian tailoring are not entirely commensurate. I don't specifically know who was to blame for this, but probably the Italians were first to lie.
Why Zuckerbergâ(TM)s 14-Year Apology Tour Hasnâ(TM)t Fixed Facebook — April 2018
The Apology API is known to be of limited ultimate effect when called in a continuous digging motion.
The bunker is real, and it's spectacular.
Then you deserve to have your radio signals amplified and replayed. Just because you find it annoying doesn't mean a terrible engineering solution somehow becomes a good idea.
I suppose you could two-factor with a fingerprint reader on your car door. Should work great until that surprise -50 degrees F event (climate isn't what it used to be). Then you be standing outside your car in some cold, remote place (remote from your fireplace), wishing your fob had a convenient and reliable button, rather than your car door having an erratic fingerprint reader, that reads blue fingers even less well than mostly sometimes.
Historical norms should be taken with a grain of salt. Additionally, certain gambling games always go up, until the black swan comes along and multiplies by zero.
How many 2008s can the average person soak up in one lifetime?
Russ Roberts interviews Nassim Nicholas Taleb on rationality, risk, and skin in the game — 5 March 2008
I find Taleb to be more aggressive than necessary, and I quickly get to the point of "haven't you belaboured this enough, already?" and then I get onto the Internet, and I think "well, maybe not too belaboured in the larger context".
I tend to regard "almost always" as a dead giveaway that someone is sweeping survival risk under the handy carpet of historical norms.
There's even a worse theorem here. It's pretty much a law of human nature that if you can convince a large enough group of people to onboard some substantial exposure exclusively because it "almost always" goes well, market insiders will be quick to engineer a shocking reversal on that long-standing historical trend.
Big Finance can engineer almost any outcome these days, it just needs to be big enough to justify a systemic assault (with enough sheep on the wrong side). This kind of engineering at scale doesn't come cheap, so it tends to be carefully paced (also spacing out due to the undue-regulation minimization sheep-shear spacing theorem).
Gone are the days where it's safe to make major investments facing backwards in time.
I just don't get the love for quoting the stupid-money price ceiling.
Did speaker wire teach us nothing? Did deoxygenated digital interconnects teach us nothing? Chalk one up for "you can lead a horse to water, but you can't make him drink".
The largest price you can possibly pay is of concern to criminals, cretins, scoundrels, the insane, and marketing departments only (to spell it out in full retinal redundancy).
Since we're into sports metaphors (life is a win/lose proposition) let's flesh that out a bit.
Is it like the Edmonton Oilers having to play a game without Connor McDavid dressed for the game (as happened recently) or is it more like the Red Army from the 1970s and 1980s having to scrape up a temporary replacement from the whole of Soviet sphere, on an express Aeroflot ticket you can't refuse?
Negative Facebook talking points comes with a depth chart attached that spans at least eleven full time zones.
I really don't "scratch one" is going to make an appreciable dent.
It's not that Google has anything in particular against paying their contract workers better.
But Brin and Page have long bucket lists of important ways they'd like to change the world, and many of these simply can't be undertaken until Alphabet reaches a $2 trillion market cap.
Pros and cons, as with all things.
As they say, you can't raise an omelette to the stratosphere without breaking eggs.