No, the decision to eliminate Aero Glass was deliberate & based on a perceived need to make Windows run acceptably on tablet & netbook-class hardware, which is MASSIVELY inferior in performance to any hardware most of us would consider remotely acceptable.
GPU capabilities have improved, but back when the first netbooks & tablets appeared, there were a lot of scenarios where Windows was being run with 24-bit color, but the GPU could only do hardware-accelerated transformations in 15/16-bit color... so you had to either run Windows with 15/16-bit color all the time, do without GPU acceleration at all, or try to fudge it by dropping down to 15/16-bit color during transitions, then going back to 24-bit color afterwards. #3 ended up looking worse than you'd expect it to, #1 would have gotten the computer killed by bad reviews, so we mostly ended up with strategy #2. Aero glass with strategy #2 on a computer with wimpy hardware to begin with just wasn't viable.
Microsoft's big fuck up was killing off Aero Glass entirely, instead of keeping it as the selling point for premium hardware and letting netbooks fall back to "the ugly Windows 2000 look". By trying to keep netbooks from sucking too horribly, they killed off a major reason for average users to buy higher-end hardware and subsidize the computers WE buy. Ergo, today you can buy a shit netbook for $199, but even $1500+ laptops have optical drives that literally wear out & break after 2-3 years, shitty keyboards, and inadequate cooling (and fans that break down) that causes them to cook themselves to death within a few years.
What Microsoft REALLY needs to do is quit pandering to the low end, and start giving people compelling reasons to DEMAND more powerful hardware again... and I can think of few things better than realtime-raytraced glass effects to drive consumer demand for it, even among non-gamers.
Someone (in France, I think) came up with an entirely reasonable compromise a few years ago -- Carrier-grade NAT with a static shared public IP, and 16-1024 port addresses (out of the 65,535 possible) permanently forwarded to the private IP assigned to each customer. End users configure their router as always (except technically, now double-NAT'ing). The only difference is, ports 1-32768 are shared by everyone sharing the public IP, and only a known range of upper ports gets forwarded to you (say, 49153-50177). So you don't have *complete* freedom to run services on any arbitrary port, but you still *do* get a stable range of ports at a static IP address to do it with.
Sadly, his proposal died early in the IETF process... the ipv6 militia mobilized against it as a threat to their sense of urgency (by giving everyone another relatively painless way to hit snooze for 10-20 years), and lack of interest by the DHCP people (many of whom are ipv6 militia-members themselves) to extend it to include port-range info was the deathblow.
I personally have no problem with ipv6 per se, besides the fact that it's still dysfunctional in too many current setups. I had to disable it on my router, because having it enabled added 5 seconds to almost every DNS resolution... apparently, fsck'ing Windows refused to do both 4 and 6 lookups in parallel... it would try ipv6, wait, stall, time out after 5 seconds, THEN do the ipv4 lookup & continue instantly. Request after request.:-(
Frankly, the PLA is the LEAST of my worries. China's government has no authority to prosecute me, and I doubt whether it genuinely CARES what random Americans do. If I had family members or substantial investments in China I *might* care... but I don't.
On the other hand, bored troll losers looking for random shits & giggles scare me a lot, precisely because they DO have the potential to cause large-scale harm to random strangers at little personal risk & with minimal effort.
Let's not forget the 10-20 seconds it takes to wake up 3 sleeping monitors because Windows has to re-negotiate HDCP handshakes with each of them, one by one. Made worse by the hellbent-determination of Windows to put monitors to sleep at every possible opportunity... even IF you try disabling that behavior. The next Windows update blows all the changes you made away, and you're back to reading manuals while twiddling the mouse with one hand to trick Windows into thinking it's active.
Seriously, I think someone even made a box that sits between your mouse & computer and automatically generates fake mouse activity after some period without mouse activity for this exact purpose.
> That would likely result in fewer ad networks pushing viruses around the internet since there would actually be someone to hold responsible for it.
It would also be mostly a moot point unless the site is owned by a major corporation, since an average blogger -- even one who earns enough from it to live on -- is effectively judgment-proof by virtue of not having enough assets to sustain more than one or two losses (and that's if they lose by virtue of not hiring a lawyer to fight... if they DO get a lawyer, they won't have anything left to sue for anyway by the time their legal fees are paid).
That said, perhaps the time has come for a new type of code-signing certs that indemnify the holder & insure them for $1/10/100 million... but allow the issuers to impose their own security & auditing requirements. So you might decide to blindly allow code signed with a $100 million cert to silently execute, allow code signed by a $10 million cert to either run silently in a sandbox or prompt for approval, allow code signed by a $1 million cert to run sandboxed if it's same-origin as the site itself or fail quietly if not, and require that unsigned code be delivered via https, with appropriate content security policy headers, for it to do anything besides fail silently.
Main rationale for the final rule/exception: most javascript-enabled malware comes from third-party code hosted by someone else & used without constant scrutiny by the linking site. Forcing site owners to host it locally & configure their server to enforce content security won't eliminate the problem, but it will mitigate most of the low-hanging drive-by fruit.
With all the time & effort you'd have to invest to resell even one ticket, you'd basically have to be "homeless-person desperate" to bother. Between driving to the theater (5-15 minutes), hustling a buyer (5-10 minutes), buying the ticket (5 minutes IF you don't have to wait in line), then driving home (5-15 minutes), you've burned at least 20-45 minutes to make slightly more than minimum wage. Sure, the economics might work out if you're dirt-poor & desperate, but just barely.
That said, adding "demand pricing" is kind of a bullshit move. Regardless of whether I go to see a movie on opening weekend or 3 weeks later, the ticket cost to MoviePass is the same (since they DON'T get special pricing or rebates from theaters).
Then again, I don't understand why theaters dislike MoviePass in the first place... they're selling tickets for full price, selling more than they would have otherwise sold, AND getting the concession revenue piled on top (possibly selling more per purchase, since in the moviegoer's mind the ticket was 'free', which reduces the sting of paying $10 for popcorn and a drink).
No, but I have yet to encounter a device capable of running both Windows and Android where the performance of Windows wasn't profoundly worse than the performance of Android.
The fact is, even for undemanding and technologically-illiterate users, slow computers suck. Using touchscreens when the display keeps mutating and changing under your finger (so the screen re-composes ~200ms AFTER you start moving your finger towards the screen, but ~100ms BEFORE it actually REACHES the screen) is UNBELIEVABLY frustrating. Ditto, for times where you touch the screen and don't get instant visual feedback & have to decide between touching the screen again (and risking having the touch register & the screen re-composing again midway between starting the motion and touching the screen) or waiting.
One serious deficiency of ALL current touch-based UIs... touching the screen triggers whatever is at that precise point at the moment touch occurs, not what was at that point ~100ms or so earlier (when the muscles began moving, but before you'd have had time to notice and react to a sudden unexpected screen recomposition). Microsoft, Google, Linux in general, and Apple ALL suffer from this problem. Someone needs to come up with a way for touch UIs to gracefully deal with ambiguous touch events that occur in the 150ms or so AFTER a screen-recomposition that changes the target of a touch event. Say, maintaining a touchzone history for a few hundred milliseconds after any changes... if the system detects a touch or gesture that WOULD have had a different outcome had it occurred 150ms earlier, have it automatically ignore the gesture AND unambiguously & instantly make it clear to the user (preferably, by some tactile means... say, multiple vibrators at opposite ends of the device, pulsed in a sequence with an unmistakable feel).
And unless the Snapdragon 1000 is miraculously clocked at something impossible with present technology, like 8,000MHz, any i7 -- even a dualcore mobile one that's basically *shit* -- will completely roast it at running x86 & AMD64 apps.
x86/AMD64 have literally evolved hand-in-hand with Windows for 30 years. Today, "pure" RISC might reduce costs & allow high performance for apps built to run the "RISC" way, but when it comes to maximum balls-to-the-wall brute-force winning performance, AMD64 wins, hands down. A modern best-of-breed AMD64 chip is basically a RISC chip, augmented by thousands of additional specialized instructions to accelerate things that Windows (and Linux) PCs *do*. AMD64 processors are basically now like self-optimizing JIT compilers for x86 assembly, with plenty of additional tricks to accelerate and optimize nearly every chokepoint known to exist in Windows (and Linux).
My phone has a Snapdragon 850 with 4x1.5GHz + 4x2GHz cores, 3 gigs of RAM, and 128 gigs of eMMC SSD. Frankly, Acrobat Viewer runs faster on an ancient, creaky 15 year old Compaq Armada m700 with 500mhz Pentium III, 512mb, and a noname PATA SSD from China (that probabiy has microSD cards hand-soldered to an ASIC inside) than it does on my phone.
When comparing ARM to AMD64, specs alone don't come anywhere CLOSE to telling the real story.
What's that? It's for people who "just" use it for running browser-based apps? Ok, go to walmart.com, amazon.com, or sears.com with the fastest ARM-based phone or tablet you can find, and compare the experience to even a *shit* AMD64-based PC or laptop. The PC or laptop will win, hands-down, because Javascript performance in particular suffers *horribly* on ARM compared to AMD64. Sites like the aforementioned that build pages from the inside-out using Javascript, DOM, and AJAX work just fine on a real computer, but turn into a minefield of delayed clicks and pages that keep reflowing and recomposing more slowly than you can see, but faster than your finger can react once you go to touch something.
ARM is why we can't have nice things like Aero Glass anymore (MS took it away & forced Metro on us because otherwise, Windows would have been unusably slow on crap ARM-based tablets). Dammit, it's almost 2020... we should have UIs with butter-smooth realtime-raytraced refraction & translucency effects by now, not something that looks like a fsck'ing mess of Post-It notes.
This reminds me of the joy (not!) we had when USB replaced RS232.
In "the old days", serial ports were pretty much idiot-proof, as long as you didn't have to screw with anything where dsr, dtr, cts, or rts actually mattered (besides maybe using a cable that connected one of them to ground). The universe of potential baudrates was small and finite (as a practical matter, nearly everything used 115200, 38400, 19200, 9600, 2400, 1200, or 300, with no parity, 8 bits, and 1 stop bit). After verifying that the cable was good, all you really had to do was iterate through the aforementioned 7 baudrates & there was a 98% chance one of them would work. Best of all, once you had the baudrate and bits/parity/stopbit settings right, the hardware literally worked the instant you connected the cable.
Then came USB, and everything went to hell. Drivers broke catastrophically every time Microsoft did a major Windows release. Flaky cables & poorly-designed (or excessively value-engineered) ports had endless, seemingly-random disconnects and dysfunction. And the inevitable lag between connecting a USB device & having it actually be operational made matters even worse. We traded a cheap, simple communications bus that was easy to configure and generally worked well for one that used cheaper hardware, augmented by extremely complicated software drivers that were barely understood by their own developers (most of whom depended upon proprietary vendor-supplied toolchains working literal black magic behind a curtain).
The difference is, someone who's disabled might ALWAYS exhibit symptoms that someone who's NOT handicapped might show ONLY when drunk. If you treat every analysis as a stateless, history- and context-free event, they might look alike. Factor in "all the time" vs "between 11pm Friday and dawn Monday + located within 200 feet of a bar", and the ones likely to be drunk instead of disabled start to stand out again.
The biggest real problem with using AI as judge, jury, and executioner is the casual acceptance of unfair collateral damage, often inflicted for violating ambiguous (if not outright secret, or incomprehensible even to those who WROTE the AI) rules. Just ask anyone who innocently triggered punishment from the AI used by companies like Google, Facebook, or Amazon & had to clear up the mess... as far as THEY'RE concerned, it doesn't *matter* whether you're innocent or not, because it's perceived as being more cost-effective to permanently shed a former customer in error than to expend valuable employee resources maintaining even the illusion of justice.
A dystopian future world isn't one where police are intentionally evil... it's one where 87.3 law-abiding people per day get their lives ruined by automated, merciless bureaucracy that's 99.999% infallible... but completely fucks over 0.001% of people whose only offense was misfortune. Under the nominal supervision of 3 human employees (in a law-enforcement organization that once employed thousands) whose own supervisors are either AI or even *more* overwhelmed by the 99,999 complaints from guilty people for every 1 genuine victim.
I'm merely pointing out that the long-term feeding of an ALS patient using a sponge is FAR from being a GOOD option, let alone the BEST one. Most likely scenario is that his grandmother was frail, but intellectually-competent, and SHE refused to let them place a G-tube in favor of a sponge after thinking, "I'd rather die than be kept alive by a feeding tube" because in her mind it was some horrible, gruesome procedure rather than a minor procedure that just provides an easy, painless alternate route for food to get into the stomach.
What's truly horrifying is that lots of people needlessly suffer (slow starvation, gagging, pain swallowing, choking, fear of choking, etc) just because they have this irrational mental block over being fed through a tube that goes directly into their stomach. If my post someday spares someone from that trauma by enlightening them about g-tubes, it was worth it.
There's really nothing to get "rejected", because it's just a hole with a plastic tube inserted. Infection risk is minimal, because the digestive tract isn't "sterile" in any meaningful sense of the word. Your colon is about as dirty as things get, and your stomach & mouth aren't much better.
Surgically, G-tube placement is about as minor of a procedure as circumcision. Doctors who place a lot of them can get "punch & place" toolkits that make the whole process quick & easy.
The risk of a g-tube is NOTHING compared to the trauma of trying to force nutrients down the throat of someone who can't eat normally. If the patient is too frail to sedate, they'll do it with local anesthetic.
I bought Cereproc's "William" voice for my (Android) phone years ago, and still use it as my primary TTS engine.
I couldn't STAND Android's only (at the time) included TTS engine, which (to me, at least) sounded like a stoned, prepubescent elf of indeterminate gender. TTS with a posh British accent was a million times better.
Personally, I'd *love* to see Cereproc do a pair of voices (one male, one female) that sound like someone who grew up in Britain with a vaguely-posh BBC-like accent, but have nevertheless lived in the US long enough to be able to pronounce American place-names (like "Okeechobee") properly, as well as absorbed the nuances of American highway-number names (eg, "I-595" = "I five-ninety-five", "405 Freeway" = "Four-oh-five Freeway", "State road 836" = "state road eight-thirty-six", etc. In other words, British accent, but American locale rules".
Aspirational examples from TV & movies: "Jarvis" from Iron Man, and "Gideon" from "Legends of tomorrow".
Why would they bother with something as imprecise (and frankly, dangerous) as a nutrient-soaked sponge instead of just implanting a gastric feeding tube ("G-tube") with inflatable-cuff feeding port? Attach the syringe to the port, squeeze in the liquefied food over the span of 10-20 minutes, done. Repeat for each meal.
For some reason, feeding tubes freak people out and seem scary, but they're really no big deal. If someone can't swallow food, a G-tube is vastly better than almost any conceivable alternative.
1) Set aside an entire IPv6/32 prefix for North American VoIP (and other prefixes, for other regions). Instead of today's 10-digit phone numbers, you'd get a/64 prefix to subdelegate to yourself as you like... give everyone who calls you a unique 128-bit number, via some API that simultaneously whitelists incoming calls to that number and identifies them to you going forward. Start getting spammed by calls to a number you gave out to some business whose customer database got harvested? Block it, and let that business figure out how to get in touch with you going forward at their own expense.
2. Declare the aaaa:bbbb:cccc:dddd::0 number to be the "public" one, but treat it as a mere gateway to a rich API that can be used to enforce arbitrary rules dictated by the called party (with a TLS-like protocol and legally-standardized format for unambiguously communicating those rules to the calling party). For example, I might decide to charge $100 to callers who reach me by voice, $10 to callers who leave voicemail or send an "urgent" SMS-like message, or $1-5 to callers who leave "non-urgent" text messages I can peruse at my convenience (caller chooses amount they want to pay, with messages in my incoming list sorted by payment amount in descending order). Called parties CAN opt to refund those fees on a call-by-call basis, but would be obligated to do so ONLY under terms they pre-disclosed to the caller (which could legitimately include, "entirely at my personal, arbitrary discretion", leaving it up to the caller to decide whether it was worth accepting those terms & the dictated up-front charge).
This neatly solves the problem of businesses that need/demand a number from you -- generate one & whitelist it in one easy step, retaining the ultimate power to cut them off at the first hint of that number's abuse.
We'd also need laws prohibiting businesses from trying to strong-arm customers into refunding charges for marketing calls made to them. Say, allow companies to refuse to do business with anyone who doesn't provide them with a valid number, but prohibit them from demanding refunds made to the customer's public number (including calls made by the collections dept... if they have a valid claim, let the courts handle it instead of trying to pawn off the job on the debtor's friends & family members).
There's no need to trust the identity of calls to the non-public numbers... the address space is so sparse, it would be nearly impossible to discover numbers within it by blind dialing... especially if the FCC mandated exponentially-increasing delays between call attempts after dialing attempts to invalid numbers that extended (at reduced timing) to ALL numbers collectively managed by a telephony service provider (coming up with a formula that slows down all dialing attempts from customers of a service provider who fails to police and punish its own customers for spamming attempts). The idea isn't to make it impossible (which is itself impossible), but rather, to throw enough speedbumps to make it too expensive for spammers to try and evade by being shitty customers to multiple providers. Eventually, providers themselves will decide that certain customers are too toxic to keep & will shut them out of future service.
You could even set up rules to automatically hand off incoming calls to the called party's own server for custom handling. For example, I might charge $100 for blind incoming calls, but reduce the fee to $10 if the caller plays "Simon Says" with the dialpad (or some more sophisticated captcha-like time-gated barrier requiring active participation that can't be easily automated). (10 years ago, this worked brilliantly for me using Asterisk with VoIP to screen incoming calls before ringing my home phone, but would be trivial to overcome now using speech recognition if enough people did it to make telemarketers care).
With a system like this, even celebrities could openly share their phone number. Say, you call Tom Cruise, and see the following rule
And what makes you think the kids (or at least, the grandkids) of today's Muslim immigrants will be any more religious than an average European Catholic, Protestant, or Jew is today? Kids naturally rebel against their parents' beliefs... and the more over-the-top the parents are, the more forcefully their kids push back (and the less likely the kids will be to drift back towards their parents' religious beliefs as adults).
The future of Europe is a mundane, mostly token semi-religion that might as well be called "Chrislam"... the bland subset of Christianity & Islam that mostly overlap, ignores esoteric doctrine, and can be generally summed up as, "there's one God, a man who believed himself to be (and might... or might not... have been) the son of God was crucified by the Romans, and a guy many view as a prophet came later & said, "oops, Jesus was probably just confused... he was a good man who died for a noble cause, but the actual savior hasn't arrived yet. Don't lie, steal, or kill except for self-defense."
200 years from now, Ramadan will be as relevant to the daily lives of Europeans as Lent is now (ie, parental disapproval for openly flouting, but nobody REALLY cares).
Al-Eid will be a totally commercialized gift-giving holiday that'll take on the theme of whatever current holiday it's within a few weeks of. Late November to early January? It'll mark the start or end of the commercial Christmas-AlEid season. Autumn? Halloween-themed decorations, gifts appropriate for kids starting new school year. Summer? It'll come into its own & be treated like "Christmas in July". Spring? Easter-themed.
That's not to say that Catholics, Protestants, Shiites, Sunnis, etc. will disappear... they'll just be viewed as more extreme and uncompromising sects of ChrIslam (the way Americans view fire-breathing Evangelicals).
Feral cats CAN be human-socialized, but it takes YEARS & lots of effort. One of my cats was born feral & rescued by me from my office parking lot along with her kittens. I got them all spayed/neutered & adopted out the kittens, but she was too feral, so I decided to keep her as an "outside cat" (coaxing her inside into a large condo cage every night with food, letting her out in the morning... the cage was necessary because she fought with my other cat). The hardest 6 months of my life came when I moved & had to turn her into an "indoor" cat. For years, she ran away in terror whenever I came within 10 feet. Getting her into a carrier was a multi-hour endeavor, and she'd howl like I was killing her the whole time she was in it.
Then... around year 5 or 6 (it's now year 9)... she finally figured out that I wasn't trying to hurt her & started approaching me on her own terms. She STILL doesn't allow me to carry her without a fight, and tends to "bolt first, ask questions later", but within minutes of sitting on the sofa or getting in bed, she'll be parked next to me -- purring away.
I'd argue that it's MORE expensive for them if you order 3 articles of clothing (one item, three sizes) and return two than if you order 21 articles of clothing and return 18. Most of the shipping cost borne by Amazon is for the package itself... the marginal cost of adding one more items is fairly low.
To give an even more extreme example of potential clothing-related problems, suppose you bought Levi's jeans from Amazon. Levi's pair-to-pair consistency is basically nonexistent... you can try on 5 pairs of nominally-identical jeans in the same nominal size, and every single goddamn one will end up fitting differently. Which is why I don't wear Levi's, and wrote them out of my life years ago. I don't have the time or patience to try on endless pairs of jeans to find the one pair that actually fits, knowing that they probably WON'T fit after I've washed them anyway (I miss the days 20 years ago when the Gap's consistency was SO GOOD, you could literally run into the mall on you way to the airport, buy 3 pairs of jeans to wear on the trip, and get on the plane with absolute confidence that if you wore size 33/31 in regular cut, the three new pairs of 33/31 regular-cut jeans would fit exactly the same way the earlier pairs did. Sadly, Gap is almost as bad as Levi's now.)
Sony can't compete with rivals because its content-owning division is bigger than its engineering division & vetoes just about everything new and creative the engineers come up with unless it somehow involves tightening DRM even further.
Sony is the perfect example of a company where literally EVERYONE -- investors, consumers, and employees -- would be demonstrably better-off if the company were forcibly split up into a content-owning company and a consumer electronics company that couldn't veto or hamstring each other's activities.
There's one edge case in "city" driving where allowing autonomy on normal streets makes sense and still works: bumper-to-bumper stop & go gridlock where nobody is moving faster than 10mph anyway, and having AI pay attention to the traffic is probably a net improvement over the 90% of drivers who are only halfway paying attention and watching cat videos on their phone in their lap ANYWAY.
One big area for improvement... two-way communication between highway networks and vehicles. It's common for an accident to create ripples that persist for HOURS afterwards. You know... those times when you're creeping ahead, foot by foot... then suddenly, the road in front of you just clears up & you can go 80mph, and you're left wondering WTF you were stopping for (and the time it takes you to notice is why those ripples persist for so long after they have any active REASON to persist). With better communication between cars and highway networks, traffic engineers could attenuate those ripples by letting cars know they're approaching one, and inducing each car that reaches the point to begin accelerating a few milliseconds earlier than the previous car did. Individually, it wouldn't make much difference to either safety or drive time for one of the slowed-down drivers... but the payback would come for the drivers stuck in traffic BEHIND them. Within 10-20 minutes, they could eliminate the ripple and have cars flowing at full speed as if the ripple had never even happened.
We blew past 128-bit computing (or at least, 128-bit registers with 64-bit data bus and 48+ bit address bus) YEARS ago (ok, technically, the general-purpose registers RAX through R15 are "only" 64-bit... but the MMX registers ain't just for graphics anymore). Sandy Bridge bumped us up to 256-bit registers. Knights Landing raised that to 512 bit registers (AMD's best processors are comparable, and I think they doubled the number of huge registers from 8 to 16).
Power & Water is a bit of a reach, as is healthcare SERVICE. Prescription drugs, however, are ripe territory for poaching. Amazon has the capital resources to do things like find foreign suppliers for semi-commodity drugs sold in the US whose supply has tightened due to mergers & acquisitions, then pay to get those foreign suppliers FDA-approved in return for making them beholden to Amazon thereafter. Walgreens & CVS have lots of market power, but Amazon can out-spend them on capital investments with barely a blip on their balance sheet beyond the next quarter.
The UK has one advantage compared to ~80% of the continental US... it's big enough to merit at least one full-size distribution center, but compact enough to allow next-day delivery to the most urban 80-90% of the UK (plus the most urban 70-90% of Ireland) without needing to rely on expensive air cargo.
Even back in the Victorian era, cities in the UK had INSANELY fast mail delivery (compared to all but the most major of big cities in the US, like New York, Chicago, and San Francisco)... I think London had 6-12 deliveries per day (other big cities had 3-6), with 6-12 additional collections in between deliveries, so you could do insane things like post a lunch invitation to someone before bed & receive their reply by 11am (as long as you were both in the same city, or one was in London & the other was in a major city). From what I read, Victorian-era small package delivery was same-day (within a city, or to/from London) or (worst-case) next day (say, Southhampton to Edinburgh). Apparently, package delivery times took a MAJOR hit in the 20th century, though (phones rendered hourly mail service increasingly unnecessary, and packages no longer had the benefit of the staffing & infrastructure necessary for around-the-clock collection, sorting, transport, and delivery).
Simplified Chinese was a major OFFICIAL change, but at approximately half of the changes just changed the appearance of printed characters to match the way people had been writing them for YEARS using ballpoint pens and pencils.
The problem isn't ad-targeting per se, it's the things people with less-benign intentions can do with the same data. If I could go through life without ever seeing another ad for diapers, dog products, or tampons & see only ads for cool computer hardware, cat toys, and other things I'm interested in, that's fine. If someone uses the same data to assign me a "potential employee score" for HR departments that hurts my career, that's a HUGE problem.
The gray area is for advertisers to offer good deals to some people, and withhold them from others ("poverty is EXPENSIVE"), because that violates most Americans' sense of fairness (as a society, we DESPISE haggling & view it as borderline-immoral even though we grudgingly tolerate it for things like cars).
Slashdot Mirror
No, the decision to eliminate Aero Glass was deliberate & based on a perceived need to make Windows run acceptably on tablet & netbook-class hardware, which is MASSIVELY inferior in performance to any hardware most of us would consider remotely acceptable.
GPU capabilities have improved, but back when the first netbooks & tablets appeared, there were a lot of scenarios where Windows was being run with 24-bit color, but the GPU could only do hardware-accelerated transformations in 15/16-bit color... so you had to either run Windows with 15/16-bit color all the time, do without GPU acceleration at all, or try to fudge it by dropping down to 15/16-bit color during transitions, then going back to 24-bit color afterwards. #3 ended up looking worse than you'd expect it to, #1 would have gotten the computer killed by bad reviews, so we mostly ended up with strategy #2. Aero glass with strategy #2 on a computer with wimpy hardware to begin with just wasn't viable.
Microsoft's big fuck up was killing off Aero Glass entirely, instead of keeping it as the selling point for premium hardware and letting netbooks fall back to "the ugly Windows 2000 look". By trying to keep netbooks from sucking too horribly, they killed off a major reason for average users to buy higher-end hardware and subsidize the computers WE buy. Ergo, today you can buy a shit netbook for $199, but even $1500+ laptops have optical drives that literally wear out & break after 2-3 years, shitty keyboards, and inadequate cooling (and fans that break down) that causes them to cook themselves to death within a few years.
What Microsoft REALLY needs to do is quit pandering to the low end, and start giving people compelling reasons to DEMAND more powerful hardware again... and I can think of few things better than realtime-raytraced glass effects to drive consumer demand for it, even among non-gamers.
Someone (in France, I think) came up with an entirely reasonable compromise a few years ago -- Carrier-grade NAT with a static shared public IP, and 16-1024 port addresses (out of the 65,535 possible) permanently forwarded to the private IP assigned to each customer. End users configure their router as always (except technically, now double-NAT'ing). The only difference is, ports 1-32768 are shared by everyone sharing the public IP, and only a known range of upper ports gets forwarded to you (say, 49153-50177). So you don't have *complete* freedom to run services on any arbitrary port, but you still *do* get a stable range of ports at a static IP address to do it with.
Sadly, his proposal died early in the IETF process... the ipv6 militia mobilized against it as a threat to their sense of urgency (by giving everyone another relatively painless way to hit snooze for 10-20 years), and lack of interest by the DHCP people (many of whom are ipv6 militia-members themselves) to extend it to include port-range info was the deathblow.
I personally have no problem with ipv6 per se, besides the fact that it's still dysfunctional in too many current setups. I had to disable it on my router, because having it enabled added 5 seconds to almost every DNS resolution... apparently, fsck'ing Windows refused to do both 4 and 6 lookups in parallel... it would try ipv6, wait, stall, time out after 5 seconds, THEN do the ipv4 lookup & continue instantly. Request after request. :-(
Frankly, the PLA is the LEAST of my worries. China's government has no authority to prosecute me, and I doubt whether it genuinely CARES what random Americans do. If I had family members or substantial investments in China I *might* care... but I don't.
On the other hand, bored troll losers looking for random shits & giggles scare me a lot, precisely because they DO have the potential to cause large-scale harm to random strangers at little personal risk & with minimal effort.
Let's not forget the 10-20 seconds it takes to wake up 3 sleeping monitors because Windows has to re-negotiate HDCP handshakes with each of them, one by one. Made worse by the hellbent-determination of Windows to put monitors to sleep at every possible opportunity... even IF you try disabling that behavior. The next Windows update blows all the changes you made away, and you're back to reading manuals while twiddling the mouse with one hand to trick Windows into thinking it's active.
Seriously, I think someone even made a box that sits between your mouse & computer and automatically generates fake mouse activity after some period without mouse activity for this exact purpose.
> That would likely result in fewer ad networks pushing viruses around the internet since there would actually be someone to hold responsible for it.
It would also be mostly a moot point unless the site is owned by a major corporation, since an average blogger -- even one who earns enough from it to live on -- is effectively judgment-proof by virtue of not having enough assets to sustain more than one or two losses (and that's if they lose by virtue of not hiring a lawyer to fight... if they DO get a lawyer, they won't have anything left to sue for anyway by the time their legal fees are paid).
That said, perhaps the time has come for a new type of code-signing certs that indemnify the holder & insure them for $1/10/100 million... but allow the issuers to impose their own security & auditing requirements. So you might decide to blindly allow code signed with a $100 million cert to silently execute, allow code signed by a $10 million cert to either run silently in a sandbox or prompt for approval, allow code signed by a $1 million cert to run sandboxed if it's same-origin as the site itself or fail quietly if not, and require that unsigned code be delivered via https, with appropriate content security policy headers, for it to do anything besides fail silently.
Main rationale for the final rule/exception: most javascript-enabled malware comes from third-party code hosted by someone else & used without constant scrutiny by the linking site. Forcing site owners to host it locally & configure their server to enforce content security won't eliminate the problem, but it will mitigate most of the low-hanging drive-by fruit.
With all the time & effort you'd have to invest to resell even one ticket, you'd basically have to be "homeless-person desperate" to bother. Between driving to the theater (5-15 minutes), hustling a buyer (5-10 minutes), buying the ticket (5 minutes IF you don't have to wait in line), then driving home (5-15 minutes), you've burned at least 20-45 minutes to make slightly more than minimum wage. Sure, the economics might work out if you're dirt-poor & desperate, but just barely.
That said, adding "demand pricing" is kind of a bullshit move. Regardless of whether I go to see a movie on opening weekend or 3 weeks later, the ticket cost to MoviePass is the same (since they DON'T get special pricing or rebates from theaters).
Then again, I don't understand why theaters dislike MoviePass in the first place... they're selling tickets for full price, selling more than they would have otherwise sold, AND getting the concession revenue piled on top (possibly selling more per purchase, since in the moviegoer's mind the ticket was 'free', which reduces the sting of paying $10 for popcorn and a drink).
No, but I have yet to encounter a device capable of running both Windows and Android where the performance of Windows wasn't profoundly worse than the performance of Android.
The fact is, even for undemanding and technologically-illiterate users, slow computers suck. Using touchscreens when the display keeps mutating and changing under your finger (so the screen re-composes ~200ms AFTER you start moving your finger towards the screen, but ~100ms BEFORE it actually REACHES the screen) is UNBELIEVABLY frustrating. Ditto, for times where you touch the screen and don't get instant visual feedback & have to decide between touching the screen again (and risking having the touch register & the screen re-composing again midway between starting the motion and touching the screen) or waiting.
One serious deficiency of ALL current touch-based UIs... touching the screen triggers whatever is at that precise point at the moment touch occurs, not what was at that point ~100ms or so earlier (when the muscles began moving, but before you'd have had time to notice and react to a sudden unexpected screen recomposition). Microsoft, Google, Linux in general, and Apple ALL suffer from this problem. Someone needs to come up with a way for touch UIs to gracefully deal with ambiguous touch events that occur in the 150ms or so AFTER a screen-recomposition that changes the target of a touch event. Say, maintaining a touchzone history for a few hundred milliseconds after any changes... if the system detects a touch or gesture that WOULD have had a different outcome had it occurred 150ms earlier, have it automatically ignore the gesture AND unambiguously & instantly make it clear to the user (preferably, by some tactile means... say, multiple vibrators at opposite ends of the device, pulsed in a sequence with an unmistakable feel).
And unless the Snapdragon 1000 is miraculously clocked at something impossible with present technology, like 8,000MHz, any i7 -- even a dualcore mobile one that's basically *shit* -- will completely roast it at running x86 & AMD64 apps.
x86/AMD64 have literally evolved hand-in-hand with Windows for 30 years. Today, "pure" RISC might reduce costs & allow high performance for apps built to run the "RISC" way, but when it comes to maximum balls-to-the-wall brute-force winning performance, AMD64 wins, hands down. A modern best-of-breed AMD64 chip is basically a RISC chip, augmented by thousands of additional specialized instructions to accelerate things that Windows (and Linux) PCs *do*. AMD64 processors are basically now like self-optimizing JIT compilers for x86 assembly, with plenty of additional tricks to accelerate and optimize nearly every chokepoint known to exist in Windows (and Linux).
My phone has a Snapdragon 850 with 4x1.5GHz + 4x2GHz cores, 3 gigs of RAM, and 128 gigs of eMMC SSD. Frankly, Acrobat Viewer runs faster on an ancient, creaky 15 year old Compaq Armada m700 with 500mhz Pentium III, 512mb, and a noname PATA SSD from China (that probabiy has microSD cards hand-soldered to an ASIC inside) than it does on my phone.
When comparing ARM to AMD64, specs alone don't come anywhere CLOSE to telling the real story.
What's that? It's for people who "just" use it for running browser-based apps? Ok, go to walmart.com, amazon.com, or sears.com with the fastest ARM-based phone or tablet you can find, and compare the experience to even a *shit* AMD64-based PC or laptop. The PC or laptop will win, hands-down, because Javascript performance in particular suffers *horribly* on ARM compared to AMD64. Sites like the aforementioned that build pages from the inside-out using Javascript, DOM, and AJAX work just fine on a real computer, but turn into a minefield of delayed clicks and pages that keep reflowing and recomposing more slowly than you can see, but faster than your finger can react once you go to touch something.
ARM is why we can't have nice things like Aero Glass anymore (MS took it away & forced Metro on us because otherwise, Windows would have been unusably slow on crap ARM-based tablets). Dammit, it's almost 2020... we should have UIs with butter-smooth realtime-raytraced refraction & translucency effects by now, not something that looks like a fsck'ing mess of Post-It notes.
This reminds me of the joy (not!) we had when USB replaced RS232.
In "the old days", serial ports were pretty much idiot-proof, as long as you didn't have to screw with anything where dsr, dtr, cts, or rts actually mattered (besides maybe using a cable that connected one of them to ground). The universe of potential baudrates was small and finite (as a practical matter, nearly everything used 115200, 38400, 19200, 9600, 2400, 1200, or 300, with no parity, 8 bits, and 1 stop bit). After verifying that the cable was good, all you really had to do was iterate through the aforementioned 7 baudrates & there was a 98% chance one of them would work. Best of all, once you had the baudrate and bits/parity/stopbit settings right, the hardware literally worked the instant you connected the cable.
Then came USB, and everything went to hell. Drivers broke catastrophically every time Microsoft did a major Windows release. Flaky cables & poorly-designed (or excessively value-engineered) ports had endless, seemingly-random disconnects and dysfunction. And the inevitable lag between connecting a USB device & having it actually be operational made matters even worse. We traded a cheap, simple communications bus that was easy to configure and generally worked well for one that used cheaper hardware, augmented by extremely complicated software drivers that were barely understood by their own developers (most of whom depended upon proprietary vendor-supplied toolchains working literal black magic behind a curtain).
The difference is, someone who's disabled might ALWAYS exhibit symptoms that someone who's NOT handicapped might show ONLY when drunk. If you treat every analysis as a stateless, history- and context-free event, they might look alike. Factor in "all the time" vs "between 11pm Friday and dawn Monday + located within 200 feet of a bar", and the ones likely to be drunk instead of disabled start to stand out again.
The biggest real problem with using AI as judge, jury, and executioner is the casual acceptance of unfair collateral damage, often inflicted for violating ambiguous (if not outright secret, or incomprehensible even to those who WROTE the AI) rules. Just ask anyone who innocently triggered punishment from the AI used by companies like Google, Facebook, or Amazon & had to clear up the mess... as far as THEY'RE concerned, it doesn't *matter* whether you're innocent or not, because it's perceived as being more cost-effective to permanently shed a former customer in error than to expend valuable employee resources maintaining even the illusion of justice.
A dystopian future world isn't one where police are intentionally evil... it's one where 87.3 law-abiding people per day get their lives ruined by automated, merciless bureaucracy that's 99.999% infallible... but completely fucks over 0.001% of people whose only offense was misfortune. Under the nominal supervision of 3 human employees (in a law-enforcement organization that once employed thousands) whose own supervisors are either AI or even *more* overwhelmed by the 99,999 complaints from guilty people for every 1 genuine victim.
I'm merely pointing out that the long-term feeding of an ALS patient using a sponge is FAR from being a GOOD option, let alone the BEST one. Most likely scenario is that his grandmother was frail, but intellectually-competent, and SHE refused to let them place a G-tube in favor of a sponge after thinking, "I'd rather die than be kept alive by a feeding tube" because in her mind it was some horrible, gruesome procedure rather than a minor procedure that just provides an easy, painless alternate route for food to get into the stomach.
What's truly horrifying is that lots of people needlessly suffer (slow starvation, gagging, pain swallowing, choking, fear of choking, etc) just because they have this irrational mental block over being fed through a tube that goes directly into their stomach. If my post someday spares someone from that trauma by enlightening them about g-tubes, it was worth it.
There's really nothing to get "rejected", because it's just a hole with a plastic tube inserted. Infection risk is minimal, because the digestive tract isn't "sterile" in any meaningful sense of the word. Your colon is about as dirty as things get, and your stomach & mouth aren't much better.
Surgically, G-tube placement is about as minor of a procedure as circumcision. Doctors who place a lot of them can get "punch & place" toolkits that make the whole process quick & easy.
The risk of a g-tube is NOTHING compared to the trauma of trying to force nutrients down the throat of someone who can't eat normally. If the patient is too frail to sedate, they'll do it with local anesthetic.
I bought Cereproc's "William" voice for my (Android) phone years ago, and still use it as my primary TTS engine.
I couldn't STAND Android's only (at the time) included TTS engine, which (to me, at least) sounded like a stoned, prepubescent elf of indeterminate gender. TTS with a posh British accent was a million times better.
Personally, I'd *love* to see Cereproc do a pair of voices (one male, one female) that sound like someone who grew up in Britain with a vaguely-posh BBC-like accent, but have nevertheless lived in the US long enough to be able to pronounce American place-names (like "Okeechobee") properly, as well as absorbed the nuances of American highway-number names (eg, "I-595" = "I five-ninety-five", "405 Freeway" = "Four-oh-five Freeway", "State road 836" = "state road eight-thirty-six", etc. In other words, British accent, but American locale rules".
Aspirational examples from TV & movies: "Jarvis" from Iron Man, and "Gideon" from "Legends of tomorrow".
Why would they bother with something as imprecise (and frankly, dangerous) as a nutrient-soaked sponge instead of just implanting a gastric feeding tube ("G-tube") with inflatable-cuff feeding port? Attach the syringe to the port, squeeze in the liquefied food over the span of 10-20 minutes, done. Repeat for each meal.
For some reason, feeding tubes freak people out and seem scary, but they're really no big deal. If someone can't swallow food, a G-tube is vastly better than almost any conceivable alternative.
1) Set aside an entire IPv6/32 prefix for North American VoIP (and other prefixes, for other regions). Instead of today's 10-digit phone numbers, you'd get a /64 prefix to subdelegate to yourself as you like... give everyone who calls you a unique 128-bit number, via some API that simultaneously whitelists incoming calls to that number and identifies them to you going forward. Start getting spammed by calls to a number you gave out to some business whose customer database got harvested? Block it, and let that business figure out how to get in touch with you going forward at their own expense.
2. Declare the aaaa:bbbb:cccc:dddd::0 number to be the "public" one, but treat it as a mere gateway to a rich API that can be used to enforce arbitrary rules dictated by the called party (with a TLS-like protocol and legally-standardized format for unambiguously communicating those rules to the calling party). For example, I might decide to charge $100 to callers who reach me by voice, $10 to callers who leave voicemail or send an "urgent" SMS-like message, or $1-5 to callers who leave "non-urgent" text messages I can peruse at my convenience (caller chooses amount they want to pay, with messages in my incoming list sorted by payment amount in descending order). Called parties CAN opt to refund those fees on a call-by-call basis, but would be obligated to do so ONLY under terms they pre-disclosed to the caller (which could legitimately include, "entirely at my personal, arbitrary discretion", leaving it up to the caller to decide whether it was worth accepting those terms & the dictated up-front charge).
This neatly solves the problem of businesses that need/demand a number from you -- generate one & whitelist it in one easy step, retaining the ultimate power to cut them off at the first hint of that number's abuse.
We'd also need laws prohibiting businesses from trying to strong-arm customers into refunding charges for marketing calls made to them. Say, allow companies to refuse to do business with anyone who doesn't provide them with a valid number, but prohibit them from demanding refunds made to the customer's public number (including calls made by the collections dept... if they have a valid claim, let the courts handle it instead of trying to pawn off the job on the debtor's friends & family members).
There's no need to trust the identity of calls to the non-public numbers... the address space is so sparse, it would be nearly impossible to discover numbers within it by blind dialing... especially if the FCC mandated exponentially-increasing delays between call attempts after dialing attempts to invalid numbers that extended (at reduced timing) to ALL numbers collectively managed by a telephony service provider (coming up with a formula that slows down all dialing attempts from customers of a service provider who fails to police and punish its own customers for spamming attempts). The idea isn't to make it impossible (which is itself impossible), but rather, to throw enough speedbumps to make it too expensive for spammers to try and evade by being shitty customers to multiple providers. Eventually, providers themselves will decide that certain customers are too toxic to keep & will shut them out of future service.
You could even set up rules to automatically hand off incoming calls to the called party's own server for custom handling. For example, I might charge $100 for blind incoming calls, but reduce the fee to $10 if the caller plays "Simon Says" with the dialpad (or some more sophisticated captcha-like time-gated barrier requiring active participation that can't be easily automated). (10 years ago, this worked brilliantly for me using Asterisk with VoIP to screen incoming calls before ringing my home phone, but would be trivial to overcome now using speech recognition if enough people did it to make telemarketers care).
With a system like this, even celebrities could openly share their phone number. Say, you call Tom Cruise, and see the following rule
And what makes you think the kids (or at least, the grandkids) of today's Muslim immigrants will be any more religious than an average European Catholic, Protestant, or Jew is today? Kids naturally rebel against their parents' beliefs... and the more over-the-top the parents are, the more forcefully their kids push back (and the less likely the kids will be to drift back towards their parents' religious beliefs as adults).
The future of Europe is a mundane, mostly token semi-religion that might as well be called "Chrislam"... the bland subset of Christianity & Islam that mostly overlap, ignores esoteric doctrine, and can be generally summed up as, "there's one God, a man who believed himself to be (and might... or might not... have been) the son of God was crucified by the Romans, and a guy many view as a prophet came later & said, "oops, Jesus was probably just confused... he was a good man who died for a noble cause, but the actual savior hasn't arrived yet. Don't lie, steal, or kill except for self-defense."
200 years from now, Ramadan will be as relevant to the daily lives of Europeans as Lent is now (ie, parental disapproval for openly flouting, but nobody REALLY cares).
Al-Eid will be a totally commercialized gift-giving holiday that'll take on the theme of whatever current holiday it's within a few weeks of. Late November to early January? It'll mark the start or end of the commercial Christmas-AlEid season. Autumn? Halloween-themed decorations, gifts appropriate for kids starting new school year. Summer? It'll come into its own & be treated like "Christmas in July". Spring? Easter-themed.
That's not to say that Catholics, Protestants, Shiites, Sunnis, etc. will disappear... they'll just be viewed as more extreme and uncompromising sects of ChrIslam (the way Americans view fire-breathing Evangelicals).
Feral cats CAN be human-socialized, but it takes YEARS & lots of effort. One of my cats was born feral & rescued by me from my office parking lot along with her kittens. I got them all spayed/neutered & adopted out the kittens, but she was too feral, so I decided to keep her as an "outside cat" (coaxing her inside into a large condo cage every night with food, letting her out in the morning... the cage was necessary because she fought with my other cat). The hardest 6 months of my life came when I moved & had to turn her into an "indoor" cat. For years, she ran away in terror whenever I came within 10 feet. Getting her into a carrier was a multi-hour endeavor, and she'd howl like I was killing her the whole time she was in it.
Then... around year 5 or 6 (it's now year 9)... she finally figured out that I wasn't trying to hurt her & started approaching me on her own terms. She STILL doesn't allow me to carry her without a fight, and tends to "bolt first, ask questions later", but within minutes of sitting on the sofa or getting in bed, she'll be parked next to me -- purring away.
I'd argue that it's MORE expensive for them if you order 3 articles of clothing (one item, three sizes) and return two than if you order 21 articles of clothing and return 18. Most of the shipping cost borne by Amazon is for the package itself... the marginal cost of adding one more items is fairly low.
To give an even more extreme example of potential clothing-related problems, suppose you bought Levi's jeans from Amazon. Levi's pair-to-pair consistency is basically nonexistent... you can try on 5 pairs of nominally-identical jeans in the same nominal size, and every single goddamn one will end up fitting differently. Which is why I don't wear Levi's, and wrote them out of my life years ago. I don't have the time or patience to try on endless pairs of jeans to find the one pair that actually fits, knowing that they probably WON'T fit after I've washed them anyway (I miss the days 20 years ago when the Gap's consistency was SO GOOD, you could literally run into the mall on you way to the airport, buy 3 pairs of jeans to wear on the trip, and get on the plane with absolute confidence that if you wore size 33/31 in regular cut, the three new pairs of 33/31 regular-cut jeans would fit exactly the same way the earlier pairs did. Sadly, Gap is almost as bad as Levi's now.)
Sony can't compete with rivals because its content-owning division is bigger than its engineering division & vetoes just about everything new and creative the engineers come up with unless it somehow involves tightening DRM even further.
Sony is the perfect example of a company where literally EVERYONE -- investors, consumers, and employees -- would be demonstrably better-off if the company were forcibly split up into a content-owning company and a consumer electronics company that couldn't veto or hamstring each other's activities.
There's one edge case in "city" driving where allowing autonomy on normal streets makes sense and still works: bumper-to-bumper stop & go gridlock where nobody is moving faster than 10mph anyway, and having AI pay attention to the traffic is probably a net improvement over the 90% of drivers who are only halfway paying attention and watching cat videos on their phone in their lap ANYWAY.
One big area for improvement... two-way communication between highway networks and vehicles. It's common for an accident to create ripples that persist for HOURS afterwards. You know... those times when you're creeping ahead, foot by foot... then suddenly, the road in front of you just clears up & you can go 80mph, and you're left wondering WTF you were stopping for (and the time it takes you to notice is why those ripples persist for so long after they have any active REASON to persist). With better communication between cars and highway networks, traffic engineers could attenuate those ripples by letting cars know they're approaching one, and inducing each car that reaches the point to begin accelerating a few milliseconds earlier than the previous car did. Individually, it wouldn't make much difference to either safety or drive time for one of the slowed-down drivers... but the payback would come for the drivers stuck in traffic BEHIND them. Within 10-20 minutes, they could eliminate the ripple and have cars flowing at full speed as if the ripple had never even happened.
We blew past 128-bit computing (or at least, 128-bit registers with 64-bit data bus and 48+ bit address bus) YEARS ago (ok, technically, the general-purpose registers RAX through R15 are "only" 64-bit... but the MMX registers ain't just for graphics anymore). Sandy Bridge bumped us up to 256-bit registers. Knights Landing raised that to 512 bit registers (AMD's best processors are comparable, and I think they doubled the number of huge registers from 8 to 16).
Power & Water is a bit of a reach, as is healthcare SERVICE. Prescription drugs, however, are ripe territory for poaching. Amazon has the capital resources to do things like find foreign suppliers for semi-commodity drugs sold in the US whose supply has tightened due to mergers & acquisitions, then pay to get those foreign suppliers FDA-approved in return for making them beholden to Amazon thereafter. Walgreens & CVS have lots of market power, but Amazon can out-spend them on capital investments with barely a blip on their balance sheet beyond the next quarter.
The UK has one advantage compared to ~80% of the continental US... it's big enough to merit at least one full-size distribution center, but compact enough to allow next-day delivery to the most urban 80-90% of the UK (plus the most urban 70-90% of Ireland) without needing to rely on expensive air cargo.
Even back in the Victorian era, cities in the UK had INSANELY fast mail delivery (compared to all but the most major of big cities in the US, like New York, Chicago, and San Francisco)... I think London had 6-12 deliveries per day (other big cities had 3-6), with 6-12 additional collections in between deliveries, so you could do insane things like post a lunch invitation to someone before bed & receive their reply by 11am (as long as you were both in the same city, or one was in London & the other was in a major city). From what I read, Victorian-era small package delivery was same-day (within a city, or to/from London) or (worst-case) next day (say, Southhampton to Edinburgh). Apparently, package delivery times took a MAJOR hit in the 20th century, though (phones rendered hourly mail service increasingly unnecessary, and packages no longer had the benefit of the staffing & infrastructure necessary for around-the-clock collection, sorting, transport, and delivery).
Simplified Chinese was a major OFFICIAL change, but at approximately half of the changes just changed the appearance of printed characters to match the way people had been writing them for YEARS using ballpoint pens and pencils.
The problem isn't ad-targeting per se, it's the things people with less-benign intentions can do with the same data. If I could go through life without ever seeing another ad for diapers, dog products, or tampons & see only ads for cool computer hardware, cat toys, and other things I'm interested in, that's fine. If someone uses the same data to assign me a "potential employee score" for HR departments that hurts my career, that's a HUGE problem.
The gray area is for advertisers to offer good deals to some people, and withhold them from others ("poverty is EXPENSIVE"), because that violates most Americans' sense of fairness (as a society, we DESPISE haggling & view it as borderline-immoral even though we grudgingly tolerate it for things like cars).