Most of US will still have at least one "real" computer, if only because it'll be the one device we have that's not nearly as locked-down as everything else, and can be used for things that weren't necessarily anticipated at the time it was designed.
The fact is, any sealed appliance that was prematurely value-engineered for an emerging technology is doomed to end up as a paperweight and/or in a landfill within a year or two, because almost BY DEFINITION any device that's non-extensible and limited to what its creators envisioned (or was part of this year's Official Business Strategy) is bound to be made obsolete by the next generation of technology.
Imagine, for a moment, if the wet dreams of the embedded (non-Android/IOS) tablet advocates had come true, and cheap non-extensible tablet-like devices had become pervasive circa 2005 for viewing web content over wi-fi. We'd probably STILL be waiting for CSS2 (let alone 3) to become usable without major backwards-compatibility hacks & kludges. H.264 would still be "the future", and.flv would probably have been enshrined and immutable forever because nothing else would reliably work on those devices. Hell, even with relatively OPEN standards, we're still mostly stuck on 2.4GHz 802.11n for lower-cost products. I think 2-3 years ago was the first time I could run 2.4GHz 802.11n in "greenfield" mode without having to back it up with a second set of access points locked to 802.11b for the sake of a few lingering old devices that couldn't see an 802.11n access point that was operating in greenfield mode. When I redid my parents' WiFi last Christmas, I STILL had to leave their old AP set up on another channel for the sake of 3 cameras they had that could only do b/g, and couldn't do n-greenfield.
Let's not forget the arrival of HD video. Circa 2007, there were basically three ways to get HD video into a TV: a cable or satellite box (if you were really lucky), an OTA tuner, or a computer with a DVI cable. Blu-Ray was still "coming soon", D-VHS was a cruel joke that, like DAT, was ruined by dysfunctional DRM before it even GOT to the point of anybody caring how expensive it was. And the way the CE industry casually condemned two or three entire generations of HDMI-only media players into audio obsolescence by first demanding only HDMI (no Toslink or S/PDIF), then failing to have any kind of meaningful certification FOR guaranteed HDMI audio compatibility, so even people who BOUGHT new receivers circa 2012-2014 ended up getting fucked because the receiver, media player, and TV couldn't agree about the proper way to signal compatibility and gracefully fall back. And no sooner did the CE industry FINALLY start to get HDMI 1.4c working properly with surround sound and receivers, it turned around and deprecated HDCP 2.0 before the first generation of HDMI 2.0-compatible receivers even made it into retail stores, then subjected us to another 2-3 generations of at least partially-dysfunctional hardware that was literally "defective by design".
So... I still watch streaming video with an old laptop on a dock with a Toslink port and Windows 7 (running Windows Media Center as a DVR), because it's increasingly become the last refuge of Toslink surround-sound compatibility, and because I'll be DAMNED if I'm going to buy another round of new equipment only to have it all rendered obsolete within a year. When I can buy new gear that has ironclad-guaranteed gotcha-free compatibility with ATSC 3.0, 4k video, and whatever new standards they both decide to demand... I might contemplate dropping a few thousand dollars buying new equipment. Until then, they'll have to pry my HTPC from my cold, dead hands.
HAM was far from "undocumented" -- but it was very, VERY poorly UNDERSTOOD by nearly everyone at the time... mostly, because the programming community ITSELF didn't really have a shared vocabulary for even TALKING about it. We simply had no mainstream concept of full-color graphics or their realtime manipulation.
It's one thing to downconvert 24-bit images to HAM once you already HAVE them... it's another matter entirely when the entire ecosystem we have today (scanners, digital cameras, Photoshop, etc.) *barely* existed at the unfathomably-expensive exotic professional level in *any* form.
There's also the fact that HAM sucked up all of the Amiga's RAM cycles if you had only chip or OCS synchronous RAM. From what I recall, no Amiga could use sprites or the blitter in HAM (no DMA slots free), and the CPU could only execute code during VBLANK if you had only chip/synchronous ram.
Let's not go overboard, here. Most Amiga apps *I* remember grabbed total control & never gave it back upon booting into it. Most games DMA'ed raw tracks from floppy to RAM & couldn't even be LAUNCHED from a hard drive (especially European games). The closest most apps came to even *humoring* Intuition was leaving enough of it alive to keep the mouse pointer & Screens working (since they were mostly handled in VBLANK and via Copper anyway). Apps that actually respected Amiga's formal windowing system were pretty rare. At least, in the Kickstart 1.x era (after 2.x, apps weren't *quite* as blase about killing the OS & taking over the whole computer... partly, because 2.x wasn't completely butt-ugly the way 1.x was).
> As long as all the platforms do not have a standard video and audio API
errr... Have you ever heard of Unity3D? It comes pretty damn close to being a decent abstraction layer over OpenGL and OpenGL ES, plus plastering over many of the differences between Android, IOS, OS X, Windows, Linux, and many consoles. It's not great for pushing hardware to its limits, but generally more than adequate for the majority of average games (and is damn-near the only reason anyone can even *fantasize* about commercially supporting any VR platform right now that isn't either a future investment or a purely recreational endeavor).
The main problem with direct flights between the US and Australia is that there are a lot of people who travel between "the US" and "Australia" on any given day, but not a lot of people who are LITERALLY traveling between any given pair of big cities in the US. That's a problem, because the only jets with enough fuel capacity to make the flight nonstop are jets that are too big to fill with passengers.
Put another way, if 95% of the people flying between the US and Australia have to change planes at least once anyway, you might as well funnel all the traffic at the US end through LAX... it's not really out of the way, and has thousands of daily connecting flights to just about everywhere. The only other airports that are really comparable are Chicago, JFK, and *maybe* Atlanta... but once again, none of those airports are much of an improvement over LAX if you're just passing through and connecting anyway.
Supersonic does change the equation a bit... if you have to fly subsonic from New York or Miami before flying supersonic to Australia, that's a HUGE time hit... and the smaller passenger capacity isn't a huge problem, because it's more closely aligned with actual daily flight demand anyway.
For what it's worth, I've read that Fort Lauderdale International Airport's ORIGINAL motivation for spending a fortune extending and elevating its south runway was the hope of becoming South Florida's hub for Concorde flights (Miami slammed the door on Concorde in the early 1970s... by law, all arriving and departing Concorde flights at MIA had to get noise waivers because they were so loud, and only two per day were available because Dade County was afraid that Eastern Airlines was going to buy a fleet of Concorde jets and run them back & forth between JFK and Miami all day). FLL's solution was to make its southern runway long enough to allow noisy Concorde jets to make their final approach over the Atlantic & land towards the west (with a tailwind, since the sea breeze normally blows towards the west). It ended up being moot (Concorde was grounded before the airport even broke ground on the runway project), but the county went through with it anyway because it had already spent more than 20 years getting the project approved & funded and wasn't about abandon the project at the last minute. So... the supreme irony. Ten years from now, it might be possible to fly from FLL to JFK in 30 minutes... but the extra hour it'll take to DRIVE to Fort Lauderdale from Miami will probably exceed the net time savings of taking the faster flight in the first place.
Ideally, it would be something you could freely opt into or out of. Carriers have the advantage that they can anonymously scan incoming messages & keep count of similar ones, escalating the "is this spam?" judgment call to a human once some threshhold is exceeded.
Blocking by origin number sounds nice, but doesn't really work because there's nothing to certify that a SMS sender actually IS who they claim to be. You can block spamming SMS numbers all day & ultimately accomplish nothing besides wasting your time because they probably weren't REALLY using that number anyway.
It's why Gmail is so good at catching spam... they see EVERYONE'S incoming messages & flag similar messages sent to lots of users for human scrutiny.
The representative who opposed the bill isn't entirely *wrong*, but at the moment there aren't many better options that can be implemented *quickly* to reduce sms spam. It comes down to, "is it worth the potential risk of telco tyranny to reduce our spam load NOW"? As long as it's done in a way you can freely opt into or out of, I'd say yeah.
I believe one future idea being explored is kind of a combination of OLED & LCD -- using broader-spectrum OLED that's less susceptible to burn-in, but more optimized towards proper red, green, and blue, then refined by passing it through a mask of red/green/blue filters.
The big question is whether mass-market buyers can be convinced to spend more for increased long-term color fidelity. If it becomes seen as a 'must-have' feature, it might add something like $100 to a mid-priced TV. If it were a niche feature, it might add $500-1000+ and be available only on the most expensive displays (because it would get little/no benefit from economies of scale).
Therein lies the rub in today's race to be the worst & cheapest at everything. With a global market of billions, even a tiny price difference can gain or lose millions of unit sales, amplified further by economy-of-scale difference. It's why in the past (when everything was expensive), you could spend a little more & get a much better product, but NOW, you have to spend 2-16 times as much to get any real improvement. The cheapest & shittiest product soaks up nearly 100% of the economy of scale benefit, making everything else MAGNITUDES more expensive. China hasn't eliminated the highest of high end, but it has mostly destroyed what USED to be the sane "middle end" by making the lowest end slightly better while simultaneously making it "nearly free" compared to what used to be middle-end prices. It's easy to pay $549 instead of 499 to get a better product. It's a lot harder to convince people to pay $1,999 to get something marginally better than $299 (once the $299 product soaks up 100% of the market's economies of scale). It's a development that, AFAIK, neither Smith, Marx, nor classic economics in general really anticipated (at least, not to the degree it has happened, with such a HUGE ultimate price ratio between 'shit' and 'marginally-better').
It's entirely possible to get high-quality white light from LEDs that's objectively superior to blackbody light... it's just that by the time you GET it, it requires almost as much energy, and throws off nearly the same amount of total heat as halogen (except, 100% of that heat is concentrated in tiny areas that need heatsinks to avoid melting themselves). You basically combine near-UV light (with purple, not blue) glow with the right phosphors, then fortify it with superbright near-infrared LEDs to support the red end.
You have to use superbright near-infrared to fortify the "red" end, because more "energy efficient" (visible light per watt) red LEDs spill over into the green and blue ends of the spectrum and give the light a visibly "pink" cast. Near-infrared is far enough away to stimulate your eyes' L-cones when it's sufficiently bright, but attenuate quickly enough beyond red to leave your M-cones and S-cones relatively unstimulated. By combining superbright near-infrared LEDs with high-CRI UV-seeded fluorescent light with the best phosphors available, you can achieve light that's the closest you can realistically get to "high noon" natural sunlight.
It's really a shame FED display technology was commercially choked off by LCD before it ever really had a chance to hit the market (for those who don't remember, circa 1997 "everyone" thought some variation of FED was *absolutely* the future of display technology... the rich, saturated hues of a CRT, but flat & low(-er) power (because they basically WERE flat, solid-state CRTs... except instead of shooting electrons from a point source & aiming them with a magnetic field, they just put multiple solid-state electron-emitters directly BEHIND each phosphor dot/bar and illuminated them directly). We kind of/sort of ended up getting it with OLEDs, but OLEDs shift hues as they age & basically act like burned-in high-persistence color CRTs when they're not new (OLED displays look FANTASTIC when they're brand new, but look dreadful after a few years due to uneven hue-shifting).
Part of the problem has to do with the "white" LEDs used for backlighting the LCD -- the exact nature of their "white" varies slightly from batch to batch. Expensive TVs (usually) make a point of using "white" LEDs that are from the "best" batches/bins (consistency from LED to LED, color purity, etc). Cheap TVs use "white" LEDs from the lower batches/bins. The cheapest Black Friday TVs use whatever LEDs were left over after making the "main" manufacturing run.
So... you might have a TV made by someone like Samsung with a model number like MHD4kQ62 that gets made with the best LEDs... then a model with similar (published) specs that uses the cheaper LEDs & has a model number like MHD4kQ62SXB that sells for $100 less at Best Buy, and an additional model that once again has similar published specs, but uses the cheapest/leftover LEDs, has a model number like MHD4kQ62SVW and sells for $27 less than the Best Buy version at Walmart (possibly with fewer HDMI inputs, just to further spite buyers and shave another 25 cents from the manufacturing cost).
The point is, they don't talk about THOSE performance specs, because they don't WANT to talk about those performance specs. By not talking about them, they can let Walmart have a model that looks almost the same on paper, even if it's egregiously inferior if you saw it side by side with the most expensive variant.
Another area where they often cut corners: the timing circuit that allows a TV to natively deal with 50hz and 60hz, instead of being locked to 50hz OR 60hz native. It only saves a few cents because it's mostly just a few passive components omitted from the mainboard, but they do it anyway (especially with US models) because they know that 98% of US buyers won't notice the difference anyway.
The bigger problem is that "99.99999%" of SSDs encrypt EVERYTHING at the block level, using an encryption key known only to the drive itself. So EVEN IF you can easily rip the bits from the failed drive's flash using a JTAG reader, you'll be reading what's effectively random noise.
The reason for encrypting the data itself is legit (it makes the bits look pseudorandom & improves wear), but IMHO, the fact that there's literally NO WAY to replace the drive's own encryption key with one known to the drive's owner is absolute, complete BULLSHIT.
As far as I know, it's a descendant of CPRM DRM. It's technically been a mandatory part of the ATA spec since the early 2000s... the difference is, on non-SSDs, it's disabled by default (the only devices I'm aware of that might actually use it are things like TiVO DVRs and videogame consoles). With a SSD, it's always on & can't be disabled or made to use a key known to YOU. As a result, data-recovery companies can still do recovery on a drive suffering from FILESYSTEM corruption, but they're now completely helpless if the drive ITSELF fails (even if the failure doesn't directly involve the flash memory). And unlike the old days, if the logic board is fried, you can't even solder the chips onto a sacrificial board, because the encryption key is tied to the original logic board.
Put another way, thanks to mandatory block-level black-box encryption, something that has always been a bad situation (drive failure) has NOW become insurmountably worse, even though the technical challenge of physically reading bits from failed media is arguably easier now than it has ever been in history.
^ That's usually what happens when you pair people who are TOO similar... like pairing an INTP with an INFP or ENTP (or worse yet, another INTP). They succumb to the same problems, and feel the most strongly about things they're likely to disagree about. Egos clash, and it rarely ends up being a good idea.
For pairing to really work well, you need two people who are similar enough to agree about striving toward the same goals, but follow meandering mental paths that cross back & forth along the way towards that goal. Ideally, when perfectly matched, you end up with two people who can easily solve each other's problems without making them feel insulted, and both feel like they're getting as much out of it as they're putting in.
The main reason pairing works at all in the real world is due to demographics. INTP & INTJ (the golden pair) is rare overall, but account for something like 40-60% of career programmers, so the odds aren't quite random. But personality type compatibility is still just ONE requirement... they also need to be comparable in skill & genuinely regard each other as equals. Even an INTP and INTJ can resent each other or have cat fights if their skills/contributions are too lopsided.
^ oops, cut all the numbers in half, since INTP+INTJ is the same as INTJ+INTP (and so on).
Just to add, pairing two programmers with the SAME personality type is generally a waste... they'll both get stuck on the same problems, fall into the same traps, and often either bicker or tune out while the other is "in charge".
Pair an INTP with an ENTP, and they'll derail each other in no time flat. They'll probably have a great time, and might come up with a brilliant, creative solution to the wrong problem, but are unlikely to accomplish their goal (or even remember what it WAS).
Pairing a "N" type with a complementary "S" type might help when it comes down to things like "AI vs UI work", but you'd arguably get more done with a xNxx (AI) and a xSxx (UI) if you had a good project manager just split up the task & had both programmers simultaneously working on the sub-task they enjoy & excel at (N-types tend to be theoreticians who LOVE things like AI, but come up with the ugliest & most dysfunctional UIs on earth... S-types tend to be artisans who get immense satisfaction out of making a beautiful, polished UI, but get frustrated with more abstract things like AI. HOWEVER, N and S types CAN complement each other well in areas like computer vision... just not necessarily touching the same code simultaneously. Put them together to brainstorm, then let them be inspired & go explore on their own for a while.
The main problem with pairing ENxx types with INxx types is that it tends to get the INxx type mowed over and bulldozed... at best, the INxx gets shoved aside, bored, and tunes out. Both groups add insight to each other (the ENxx types expand the INxx types' horizons & inspire them to "think big"), but they work better as teammates than pairs.
Getting back to INTP+INTJ... it works so well because both types easily see around each other's roadblocks, effortlessly spot each other's typos + oversights, and generally follow meandering paths towards the same goal. It's important that they regard each other as equivalently-competent, but ideally each will be an expert in a slightly different aspect of the problem.
INTP+INFJ or INFP+INTJ can work, but the 'F' partner is more likely to be unintentionally insulted by the 'T' partner, or read into motives that don't actually exist. Likewise, a badly-frustrated 'F' can sink a verbal dagger pretty deeply into their 'N' complement and *really* hurt their feelings in a way few other types are capable of (INTx types get accused of having no feelings, but the reality is that they have a Jeckyl-Hyde relationship with their feelings... they're rightfully scared *shitless* of them, and work hard to keep them safely locked away for everyone's safety & well-being lest they explode out like a shaken bottle of Diet Coke).
The point is, different types all have valuable insight, but not all of them can intimately work *together* as a pair. Some combos are better left as teammates & coworkers for everyone's sake, and others are better off collaborating briefly, in small chunks, before going back to working on their own.
If you assume that there are approximately 16 general hardware configurations humans tend to cluster around, there are approximately:
* 16 combinations that are *golden* for pair programming. Say, INTP + INTJ (probably the most ideal, golden pairing of all)
* 16-48 combinations that are likely to be subpar at best. Say, INFP + INTJ, or ENTJ + INTP
* at least 128 combinations likely to be flat-out dysfunctional, like INTP + ESTJ
The point being, you can't pair by HR or management decree or force it to work. When you match up two programmers with complementary personality types, of roughly equal skill, it's magic. Nearly every other combo ends up being a net drawback, handicap, and liability.
Actually, government & commercial are additive in both countries, not mutually exclusive.
China: companies officially beholden to CCP, AND they want to know everything about you to sell you more stuff (though both are largely indifferent to you unless you're either a threat to the CCP *or* a potential consumer in China)
US: companies can be compelled to secretly do the federal government's bidding (upon the relevant law enforcement agency or intelligence agency getting a court order), and they want to know everything about you (regardless of where you are) so they can sell you stuff.
To be honest, China worries me less. China's government wants to exercise total control over people in China, but doesn't give much of a shit about anyone else. The US's government wants to exercise jurisdiction over everyone on earth, and has the de-facto power to at least indirectly impose it upon a large plurality of earth's inhabitants.
"Wants to" + "sort of able to" is FAR more dangerous than "Indifferent to" + "generally incapable of (unless you're Chinese)"
Will people please stop fantasizing about cities like NY, DC, and SF flooding? It won't be allowed to happen, regardless of what might happen elsewhere. All of those cities have relatively narrow paths to the open ocean that can, and certainly will, be blocked with dams to hold back rising sea levels.
And most of Florida's urban coastline will be fortified & raised with new crushed limestone and/or concrete as hurricanes progressively destroy it storm by storm, until Miami (and much of the rest of Florida) literally IS a "concrete jungle" (on a manmade hill).
Will sea-level rise devastate natural ecosystems? Of course. Will it destroy cities? No. Or at least, not permanently... new stuff fortified against climate change will replace whatever gets destroyed, and big cities will remain largely where they are... just taller, on raised terrain, with a hell of a lot more concrete.
No, I'm not contradicting myself. You're twisting my words.
A jackknifed 18-wheeler perpendicular to oncoming traffic is not equivalent to an 18-wheeler driving across a normal 4-6 lane road at an intersection. The former is a rare edge case that isn't supposed to happen, and when it DOES, it's almost always a freak accident that creates a situation so dangerous, EVERYONE in the area is at risk of injury or death. The latter is something that happens CONTINUOUSLY whenever an 18-wheeler is driving down a normal road.
The Tesla accident involving the autopilot slamming into a stopped emergency vehicle parked along the shoulder actually strengthens my original point... stopped vehicles alongside a limited-access freeway -- even in the emergency lanes -- CREATE a significant hazard for both the parked vehicle's occupants AND people driving on the road. Some parked vehicles are necessary, like emergency vehicles. Nevertheless, that doesn't mean we're safer if we intentionally create situations where MORE cars are encouraged or required to pull over and stop. If a guy with semi-autonomous autopilot falls asleep at the wheel, the safest course of action is to just keep driving while attempting to wake him up (preferably, while buffering the car's controls from sudden movements he might make if startled), even if he wakes up to find himself 42 miles past his intended exit. Everyone -- human and autonomous vehicle -- is safer when cars on a freeway behave in an outwardly-rational, predictable manner. Everyone is in danger when vehicles deviate from the predicted norm in shocking or surprising ways.
US-27 is most assuredly NOT a limited-access freeway. The truck Joshua Brown's car skidded under was perpendicular to oncoming traffic. That would not happen on a real freeway (or at least, if it did, it would only be if there were some horrific accident involving a jackknifed truck that would probably result in the deaths of HUMAN drivers, too).
Repeat after me: constraints. A limited-access freeway is a relatively easy case to design for. A rural highway with high-speed traffic and traffic crossing at grade is a safety nightmare for autonomous vehicles.
On a limited-access highway, "coming to a stop" is usually WORSE than "staying in the lane & maintaining a normal cruising speed". Autopilot lanekeeping & collision-avoidance is now better on average at avoiding accidents than most human drivers. It's UNUSUAL situations that create the danger.
A car staying in its lane & moving appropriately is the norm on a freeway. A car stopped on the shoulder creates an active road hazard for everyone else. Remember, his autonomous vehicle wasn't the only one on the road. It's not a safety improvement if your "safety feature" creates a situation that's MORE dangerous than letting the car just do what it does best -- follow the lane, avoid collisions, and behave in a predictable manner.
If you really need to punish drivers, add warning lights to visually communicate to other drivers that a car is operating without active human control. On a limited-access highway not under construction in good weather, inattentiveness with autopilot is a statistical non-issue... and in the real world, it's probably a net improvement over human drivers who are semi-distracted. On a non-freeway that has at-grade cross traffic, it's a genuine hazard. In bumper-to-bumper city-street gridlock, it's a non-issue (the car can stop within inches if necessary). On city streets at 30mph, it's likely to be dangerous.
The point is, there is no blanket "one size fits all" rule. Autonomous systems have known constraints. Stay within them, and you're fine. Deviate, and you create problems. Ignore the constraints entirely, and you're in uncharted territory risk-wise. Following lines & not colliding is easy. Finding a safe place to pull over is enormously harder.
The thing is, if you're a business, depositing cash in a business checking account usually incurs fees of its own:
* Do you have an employee carry the cash to the bank? HUGE insurance premium, because you're going to get sued for a small fortune if that employee gets mugged on the way there.
* Do you pay an armored car company to pick up the deposit? This costs money.
* Are you a sole proprietor who carries your own deposits to the bank? You're probably insane, because eventually you WILL get mugged by someone.
OK, one way or another, you get the cash to the bank and hand it to the teller. Guess what? Counting fees, cash-deposit fees, and additional surcharge fees for coins. If you're a business, banks constantly hit you up for fees in BOTH directions.
Cash is only "free" for a business to accept if it turns around and directly uses that cash to pay its employees, vendors, rent, and the owner takes the remainder home and deposits it straight into his own personal bank account. And if you do this, your accountant is going to absolutely hate you & charge you more for their services.
That's interesting, I didn't know the Volt worked that way.
Does it somehow make NORMAL gas last a year, or does the "one year" assume you added stabilizer at the same time as the fill-up? Or does it have a separate tank for stabilizer, and add it automatically a few days after your last fill-up?
Or worse, they'll deploy 5G radio technology that only works on a few AT&T-branded phones, without meaningfully increasing the amount of fiber backhaul available to back up the new RF bandwidth.
It's kind of like if you had 50 people connected to an 802.11ac access point, but the access point itself were connected to a DSL modem with bottom-tier service.
Except... it's actually worse. At least with the 802.11ac access point, users on the same LAN could get 100+mbps speeds peer to peer with each other, regardless of how fast or slow the internet connectivity is, because the local access point is (usually) smart enough to RECOGNIZE traffic that doesn't really have to touch "the internet" itself. AFAIK, *no* mobile network even tries to route peer to peer traffic among users connected to the same tower site directly (without involving a round trip to the off-site data center), so EVERYTHING ends up being backhaul-constrained, even when there's technically local wireless bandwidth to burn.
Being Devil's Advocate for a moment... you might get 54mpg "average" when driving under mixed urban-suburban typical daily conditions, but what kind of mileage would you SPECIFICALLY see on days 2 and 3 of a drive from Miami to Los Angeles where you spent basically 100% of your driving time on an interstate doing 80mph... the kind of drive that's long enough to deplete the Prius' batteries, without really giving them a good opportunity to recharge (because the engine is running at full efficiency keeping the car moving horizontally at 80mph, and doesn't really HAVE any surplus energy to divert to battery charging under those conditions... especially if the air conditioner is running on max & you have a car stereo system with a huge amp(*)).
(*) The "huge amp" part isn't entirely hypothetical... when I was younger, I destroyed my first car's battery on day 2 of my first cross-country road trip. The amp was a 50x4+250 watt (RMS) class-D amp, and basically it drew more power than the stock alternator could actually supply. On day #2, I stopped to get dinner, and discovered that the battery was too drained to start the engine afterwards (and in fact, was already damaged from being severely discharged past the point of no return).
The problem isn't that people aren't buying passenger cars per se, the problem is that SUVs and trucks are the only market where American carmakers can profitably compete against foreign competitors. In countries where "normal" people drive tiny econoboxes, an American F350 pickup truck or a Cadillac Escalade has EXACTLY the same kind of prestige and brand image that a BMW or Lexus does in the US.
This is the exact problem Chrysler ran up against ~10 years ago when Daimler owned them. Within a few years, Daimler transformed Chrysler into a company whose cars were every bit as good as the best German cars, both in quality and design. The problem was, Chrysler had close to zero brand prestige in America OR abroad (at least, for passenger cars), so they were in a no-win situation... they had cars that were as good as a premium German car, but couldn't sell them for enough money to justify that level of design, performance, and quality. So they went back to making mediocre cars with tolerable profitability.
Chrysler DID come up with a very cool, uniquely American application for hybrid electric-ICE traction... instead of using the electric motor solely to improve fuel economy, they ALSO used it to provide a short burst of additional horsepower in conjunction with a turbocharger. See, Chrysler has always been a big fan of turbochargers, but there's a reason why they've never been terribly popular with the mainstream... it takes a second or two to kick in when activated. Someone at Chrysler realized that electric traction motors aren't really useful for SUSTAINED augmenting of an ICE's horsepower (it overheats & damages the battery), but you could still use it to BRIEFLY provide surge power to offset the turbo lag (ie, slam the pedal, electric traction kicks in to add power while the turbocharger spins up, then the electric traction motor backs off as the turbocharger kicks in).
The idea was so cool, Porsche licensed it to use with THEIR OWN future hybrids. But Chrysler just couldn't pull off selling it in America. In America, Chrysler's big money-maker is muscle cars... a market segment that only cares about turbocharging as a way to make an absurdly-overpowered 12-cylinder engine act like a 16-cylinder engine. For that segment, eliminating turbo lag is almost moot, because their cars have so much power to begin with, the turbocharger is just another bullet point on the 'brag' list to add another hundred horsepower or so that will never actually be used.
Personally, I think one area where American automakers could have pulled off "mostly" electric cars would have been to make one with an all-electric drivetrain, but a small diesel engine with generator whose only purpose was to run at constant speed, with maximum efficiency, and provide enough electricity to augment the battery & double the car's range on a full charge.
Carrying your own diesel-powered battery charger isn't particularly efficient (roughly half the mileage per gallon as burning it directly for locomotion), so you wouldn't want to use it as your car's primary source of energy (vs charging), but it would have been a great stopgap measure for people who wanted to drive an EV across America back when rapid charging stations were few and far between... you might have gone for weeks or months at a time without ever firing up the diesel charger, but when you had to make that trip from California to Texas (for example), you could have fired it up and extended your EV range enough to not have to plan your entire trip and meal/rest-stop schedule around your car's battery capacity.
Why diesel? Because regular gasoline turns into varnish after a few weeks... and by definition, this would be a feature you'd really only want to use every few weeks at most (and possibly just once or twice a year). Diesel can survive months of disuse. Unleaded gasoline can't (at least, not without the kind of active effort and ongoing maintenance that experience shows American car owners simply can't be expected to perform with any degree of re
Not a 6502, but I'm pretty sure the TI99/4a used SRAM for its 256-byte "scratchpad" RAM (the only RAM its CPU could access directly).
I know the Amiga used "Static Column" RAM, but I think SC ram was what we'd NOW consider to be "PSRAM" -- DRAM with extra onboard circuitry to do its own refreshing automatically so it "looks" (and behaves) like SRAM as far as the outside world is concerned.
Slashdot Mirror
Most of US will still have at least one "real" computer, if only because it'll be the one device we have that's not nearly as locked-down as everything else, and can be used for things that weren't necessarily anticipated at the time it was designed.
The fact is, any sealed appliance that was prematurely value-engineered for an emerging technology is doomed to end up as a paperweight and/or in a landfill within a year or two, because almost BY DEFINITION any device that's non-extensible and limited to what its creators envisioned (or was part of this year's Official Business Strategy) is bound to be made obsolete by the next generation of technology.
Imagine, for a moment, if the wet dreams of the embedded (non-Android/IOS) tablet advocates had come true, and cheap non-extensible tablet-like devices had become pervasive circa 2005 for viewing web content over wi-fi. We'd probably STILL be waiting for CSS2 (let alone 3) to become usable without major backwards-compatibility hacks & kludges. H.264 would still be "the future", and .flv would probably have been enshrined and immutable forever because nothing else would reliably work on those devices. Hell, even with relatively OPEN standards, we're still mostly stuck on 2.4GHz 802.11n for lower-cost products. I think 2-3 years ago was the first time I could run 2.4GHz 802.11n in "greenfield" mode without having to back it up with a second set of access points locked to 802.11b for the sake of a few lingering old devices that couldn't see an 802.11n access point that was operating in greenfield mode. When I redid my parents' WiFi last Christmas, I STILL had to leave their old AP set up on another channel for the sake of 3 cameras they had that could only do b/g, and couldn't do n-greenfield.
Let's not forget the arrival of HD video. Circa 2007, there were basically three ways to get HD video into a TV: a cable or satellite box (if you were really lucky), an OTA tuner, or a computer with a DVI cable. Blu-Ray was still "coming soon", D-VHS was a cruel joke that, like DAT, was ruined by dysfunctional DRM before it even GOT to the point of anybody caring how expensive it was. And the way the CE industry casually condemned two or three entire generations of HDMI-only media players into audio obsolescence by first demanding only HDMI (no Toslink or S/PDIF), then failing to have any kind of meaningful certification FOR guaranteed HDMI audio compatibility, so even people who BOUGHT new receivers circa 2012-2014 ended up getting fucked because the receiver, media player, and TV couldn't agree about the proper way to signal compatibility and gracefully fall back. And no sooner did the CE industry FINALLY start to get HDMI 1.4c working properly with surround sound and receivers, it turned around and deprecated HDCP 2.0 before the first generation of HDMI 2.0-compatible receivers even made it into retail stores, then subjected us to another 2-3 generations of at least partially-dysfunctional hardware that was literally "defective by design".
So... I still watch streaming video with an old laptop on a dock with a Toslink port and Windows 7 (running Windows Media Center as a DVR), because it's increasingly become the last refuge of Toslink surround-sound compatibility, and because I'll be DAMNED if I'm going to buy another round of new equipment only to have it all rendered obsolete within a year. When I can buy new gear that has ironclad-guaranteed gotcha-free compatibility with ATSC 3.0, 4k video, and whatever new standards they both decide to demand... I might contemplate dropping a few thousand dollars buying new equipment. Until then, they'll have to pry my HTPC from my cold, dead hands.
HAM was far from "undocumented" -- but it was very, VERY poorly UNDERSTOOD by nearly everyone at the time... mostly, because the programming community ITSELF didn't really have a shared vocabulary for even TALKING about it. We simply had no mainstream concept of full-color graphics or their realtime manipulation.
It's one thing to downconvert 24-bit images to HAM once you already HAVE them... it's another matter entirely when the entire ecosystem we have today (scanners, digital cameras, Photoshop, etc.) *barely* existed at the unfathomably-expensive exotic professional level in *any* form.
There's also the fact that HAM sucked up all of the Amiga's RAM cycles if you had only chip or OCS synchronous RAM. From what I recall, no Amiga could use sprites or the blitter in HAM (no DMA slots free), and the CPU could only execute code during VBLANK if you had only chip/synchronous ram.
Let's not go overboard, here. Most Amiga apps *I* remember grabbed total control & never gave it back upon booting into it. Most games DMA'ed raw tracks from floppy to RAM & couldn't even be LAUNCHED from a hard drive (especially European games). The closest most apps came to even *humoring* Intuition was leaving enough of it alive to keep the mouse pointer & Screens working (since they were mostly handled in VBLANK and via Copper anyway). Apps that actually respected Amiga's formal windowing system were pretty rare. At least, in the Kickstart 1.x era (after 2.x, apps weren't *quite* as blase about killing the OS & taking over the whole computer... partly, because 2.x wasn't completely butt-ugly the way 1.x was).
> As long as all the platforms do not have a standard video and audio API
errr... Have you ever heard of Unity3D? It comes pretty damn close to being a decent abstraction layer over OpenGL and OpenGL ES, plus plastering over many of the differences between Android, IOS, OS X, Windows, Linux, and many consoles. It's not great for pushing hardware to its limits, but generally more than adequate for the majority of average games (and is damn-near the only reason anyone can even *fantasize* about commercially supporting any VR platform right now that isn't either a future investment or a purely recreational endeavor).
The main problem with direct flights between the US and Australia is that there are a lot of people who travel between "the US" and "Australia" on any given day, but not a lot of people who are LITERALLY traveling between any given pair of big cities in the US. That's a problem, because the only jets with enough fuel capacity to make the flight nonstop are jets that are too big to fill with passengers.
Put another way, if 95% of the people flying between the US and Australia have to change planes at least once anyway, you might as well funnel all the traffic at the US end through LAX... it's not really out of the way, and has thousands of daily connecting flights to just about everywhere. The only other airports that are really comparable are Chicago, JFK, and *maybe* Atlanta... but once again, none of those airports are much of an improvement over LAX if you're just passing through and connecting anyway.
Supersonic does change the equation a bit... if you have to fly subsonic from New York or Miami before flying supersonic to Australia, that's a HUGE time hit... and the smaller passenger capacity isn't a huge problem, because it's more closely aligned with actual daily flight demand anyway.
For what it's worth, I've read that Fort Lauderdale International Airport's ORIGINAL motivation for spending a fortune extending and elevating its south runway was the hope of becoming South Florida's hub for Concorde flights (Miami slammed the door on Concorde in the early 1970s... by law, all arriving and departing Concorde flights at MIA had to get noise waivers because they were so loud, and only two per day were available because Dade County was afraid that Eastern Airlines was going to buy a fleet of Concorde jets and run them back & forth between JFK and Miami all day). FLL's solution was to make its southern runway long enough to allow noisy Concorde jets to make their final approach over the Atlantic & land towards the west (with a tailwind, since the sea breeze normally blows towards the west). It ended up being moot (Concorde was grounded before the airport even broke ground on the runway project), but the county went through with it anyway because it had already spent more than 20 years getting the project approved & funded and wasn't about abandon the project at the last minute. So... the supreme irony. Ten years from now, it might be possible to fly from FLL to JFK in 30 minutes... but the extra hour it'll take to DRIVE to Fort Lauderdale from Miami will probably exceed the net time savings of taking the faster flight in the first place.
Ideally, it would be something you could freely opt into or out of. Carriers have the advantage that they can anonymously scan incoming messages & keep count of similar ones, escalating the "is this spam?" judgment call to a human once some threshhold is exceeded.
Blocking by origin number sounds nice, but doesn't really work because there's nothing to certify that a SMS sender actually IS who they claim to be. You can block spamming SMS numbers all day & ultimately accomplish nothing besides wasting your time because they probably weren't REALLY using that number anyway.
It's why Gmail is so good at catching spam... they see EVERYONE'S incoming messages & flag similar messages sent to lots of users for human scrutiny.
The representative who opposed the bill isn't entirely *wrong*, but at the moment there aren't many better options that can be implemented *quickly* to reduce sms spam. It comes down to, "is it worth the potential risk of telco tyranny to reduce our spam load NOW"? As long as it's done in a way you can freely opt into or out of, I'd say yeah.
I believe one future idea being explored is kind of a combination of OLED & LCD -- using broader-spectrum OLED that's less susceptible to burn-in, but more optimized towards proper red, green, and blue, then refined by passing it through a mask of red/green/blue filters.
The big question is whether mass-market buyers can be convinced to spend more for increased long-term color fidelity. If it becomes seen as a 'must-have' feature, it might add something like $100 to a mid-priced TV. If it were a niche feature, it might add $500-1000+ and be available only on the most expensive displays (because it would get little/no benefit from economies of scale).
Therein lies the rub in today's race to be the worst & cheapest at everything. With a global market of billions, even a tiny price difference can gain or lose millions of unit sales, amplified further by economy-of-scale difference. It's why in the past (when everything was expensive), you could spend a little more & get a much better product, but NOW, you have to spend 2-16 times as much to get any real improvement. The cheapest & shittiest product soaks up nearly 100% of the economy of scale benefit, making everything else MAGNITUDES more expensive. China hasn't eliminated the highest of high end, but it has mostly destroyed what USED to be the sane "middle end" by making the lowest end slightly better while simultaneously making it "nearly free" compared to what used to be middle-end prices. It's easy to pay $549 instead of 499 to get a better product. It's a lot harder to convince people to pay $1,999 to get something marginally better than $299 (once the $299 product soaks up 100% of the market's economies of scale). It's a development that, AFAIK, neither Smith, Marx, nor classic economics in general really anticipated (at least, not to the degree it has happened, with such a HUGE ultimate price ratio between 'shit' and 'marginally-better').
It's entirely possible to get high-quality white light from LEDs that's objectively superior to blackbody light... it's just that by the time you GET it, it requires almost as much energy, and throws off nearly the same amount of total heat as halogen (except, 100% of that heat is concentrated in tiny areas that need heatsinks to avoid melting themselves). You basically combine near-UV light (with purple, not blue) glow with the right phosphors, then fortify it with superbright near-infrared LEDs to support the red end.
You have to use superbright near-infrared to fortify the "red" end, because more "energy efficient" (visible light per watt) red LEDs spill over into the green and blue ends of the spectrum and give the light a visibly "pink" cast. Near-infrared is far enough away to stimulate your eyes' L-cones when it's sufficiently bright, but attenuate quickly enough beyond red to leave your M-cones and S-cones relatively unstimulated. By combining superbright near-infrared LEDs with high-CRI UV-seeded fluorescent light with the best phosphors available, you can achieve light that's the closest you can realistically get to "high noon" natural sunlight.
It's really a shame FED display technology was commercially choked off by LCD before it ever really had a chance to hit the market (for those who don't remember, circa 1997 "everyone" thought some variation of FED was *absolutely* the future of display technology... the rich, saturated hues of a CRT, but flat & low(-er) power (because they basically WERE flat, solid-state CRTs... except instead of shooting electrons from a point source & aiming them with a magnetic field, they just put multiple solid-state electron-emitters directly BEHIND each phosphor dot/bar and illuminated them directly). We kind of/sort of ended up getting it with OLEDs, but OLEDs shift hues as they age & basically act like burned-in high-persistence color CRTs when they're not new (OLED displays look FANTASTIC when they're brand new, but look dreadful after a few years due to uneven hue-shifting).
Part of the problem has to do with the "white" LEDs used for backlighting the LCD -- the exact nature of their "white" varies slightly from batch to batch. Expensive TVs (usually) make a point of using "white" LEDs that are from the "best" batches/bins (consistency from LED to LED, color purity, etc). Cheap TVs use "white" LEDs from the lower batches/bins. The cheapest Black Friday TVs use whatever LEDs were left over after making the "main" manufacturing run.
So... you might have a TV made by someone like Samsung with a model number like MHD4kQ62 that gets made with the best LEDs... then a model with similar (published) specs that uses the cheaper LEDs & has a model number like MHD4kQ62SXB that sells for $100 less at Best Buy, and an additional model that once again has similar published specs, but uses the cheapest/leftover LEDs, has a model number like MHD4kQ62SVW and sells for $27 less than the Best Buy version at Walmart (possibly with fewer HDMI inputs, just to further spite buyers and shave another 25 cents from the manufacturing cost).
The point is, they don't talk about THOSE performance specs, because they don't WANT to talk about those performance specs. By not talking about them, they can let Walmart have a model that looks almost the same on paper, even if it's egregiously inferior if you saw it side by side with the most expensive variant.
Another area where they often cut corners: the timing circuit that allows a TV to natively deal with 50hz and 60hz, instead of being locked to 50hz OR 60hz native. It only saves a few cents because it's mostly just a few passive components omitted from the mainboard, but they do it anyway (especially with US models) because they know that 98% of US buyers won't notice the difference anyway.
The bigger problem is that "99.99999%" of SSDs encrypt EVERYTHING at the block level, using an encryption key known only to the drive itself. So EVEN IF you can easily rip the bits from the failed drive's flash using a JTAG reader, you'll be reading what's effectively random noise.
The reason for encrypting the data itself is legit (it makes the bits look pseudorandom & improves wear), but IMHO, the fact that there's literally NO WAY to replace the drive's own encryption key with one known to the drive's owner is absolute, complete BULLSHIT.
As far as I know, it's a descendant of CPRM DRM. It's technically been a mandatory part of the ATA spec since the early 2000s... the difference is, on non-SSDs, it's disabled by default (the only devices I'm aware of that might actually use it are things like TiVO DVRs and videogame consoles). With a SSD, it's always on & can't be disabled or made to use a key known to YOU. As a result, data-recovery companies can still do recovery on a drive suffering from FILESYSTEM corruption, but they're now completely helpless if the drive ITSELF fails (even if the failure doesn't directly involve the flash memory). And unlike the old days, if the logic board is fried, you can't even solder the chips onto a sacrificial board, because the encryption key is tied to the original logic board.
Put another way, thanks to mandatory block-level black-box encryption, something that has always been a bad situation (drive failure) has NOW become insurmountably worse, even though the technical challenge of physically reading bits from failed media is arguably easier now than it has ever been in history.
^ That's usually what happens when you pair people who are TOO similar... like pairing an INTP with an INFP or ENTP (or worse yet, another INTP). They succumb to the same problems, and feel the most strongly about things they're likely to disagree about. Egos clash, and it rarely ends up being a good idea.
For pairing to really work well, you need two people who are similar enough to agree about striving toward the same goals, but follow meandering mental paths that cross back & forth along the way towards that goal. Ideally, when perfectly matched, you end up with two people who can easily solve each other's problems without making them feel insulted, and both feel like they're getting as much out of it as they're putting in.
The main reason pairing works at all in the real world is due to demographics. INTP & INTJ (the golden pair) is rare overall, but account for something like 40-60% of career programmers, so the odds aren't quite random. But personality type compatibility is still just ONE requirement... they also need to be comparable in skill & genuinely regard each other as equals. Even an INTP and INTJ can resent each other or have cat fights if their skills/contributions are too lopsided.
^ oops, cut all the numbers in half, since INTP+INTJ is the same as INTJ+INTP (and so on).
Just to add, pairing two programmers with the SAME personality type is generally a waste... they'll both get stuck on the same problems, fall into the same traps, and often either bicker or tune out while the other is "in charge".
Pair an INTP with an ENTP, and they'll derail each other in no time flat. They'll probably have a great time, and might come up with a brilliant, creative solution to the wrong problem, but are unlikely to accomplish their goal (or even remember what it WAS).
Pairing a "N" type with a complementary "S" type might help when it comes down to things like "AI vs UI work", but you'd arguably get more done with a xNxx (AI) and a xSxx (UI) if you had a good project manager just split up the task & had both programmers simultaneously working on the sub-task they enjoy & excel at (N-types tend to be theoreticians who LOVE things like AI, but come up with the ugliest & most dysfunctional UIs on earth... S-types tend to be artisans who get immense satisfaction out of making a beautiful, polished UI, but get frustrated with more abstract things like AI. HOWEVER, N and S types CAN complement each other well in areas like computer vision... just not necessarily touching the same code simultaneously. Put them together to brainstorm, then let them be inspired & go explore on their own for a while.
The main problem with pairing ENxx types with INxx types is that it tends to get the INxx type mowed over and bulldozed... at best, the INxx gets shoved aside, bored, and tunes out. Both groups add insight to each other (the ENxx types expand the INxx types' horizons & inspire them to "think big"), but they work better as teammates than pairs.
Getting back to INTP+INTJ... it works so well because both types easily see around each other's roadblocks, effortlessly spot each other's typos + oversights, and generally follow meandering paths towards the same goal. It's important that they regard each other as equivalently-competent, but ideally each will be an expert in a slightly different aspect of the problem.
INTP+INFJ or INFP+INTJ can work, but the 'F' partner is more likely to be unintentionally insulted by the 'T' partner, or read into motives that don't actually exist. Likewise, a badly-frustrated 'F' can sink a verbal dagger pretty deeply into their 'N' complement and *really* hurt their feelings in a way few other types are capable of (INTx types get accused of having no feelings, but the reality is that they have a Jeckyl-Hyde relationship with their feelings... they're rightfully scared *shitless* of them, and work hard to keep them safely locked away for everyone's safety & well-being lest they explode out like a shaken bottle of Diet Coke).
The point is, different types all have valuable insight, but not all of them can intimately work *together* as a pair. Some combos are better left as teammates & coworkers for everyone's sake, and others are better off collaborating briefly, in small chunks, before going back to working on their own.
If you assume that there are approximately 16 general hardware configurations humans tend to cluster around, there are approximately:
* 16 combinations that are *golden* for pair programming. Say, INTP + INTJ (probably the most ideal, golden pairing of all)
* 16-48 combinations that are likely to be subpar at best. Say, INFP + INTJ, or ENTJ + INTP
* at least 128 combinations likely to be flat-out dysfunctional, like INTP + ESTJ
The point being, you can't pair by HR or management decree or force it to work. When you match up two programmers with complementary personality types, of roughly equal skill, it's magic. Nearly every other combo ends up being a net drawback, handicap, and liability.
Actually, government & commercial are additive in both countries, not mutually exclusive.
China: companies officially beholden to CCP, AND they want to know everything about you to sell you more stuff (though both are largely indifferent to you unless you're either a threat to the CCP *or* a potential consumer in China)
US: companies can be compelled to secretly do the federal government's bidding (upon the relevant law enforcement agency or intelligence agency getting a court order), and they want to know everything about you (regardless of where you are) so they can sell you stuff.
To be honest, China worries me less. China's government wants to exercise total control over people in China, but doesn't give much of a shit about anyone else. The US's government wants to exercise jurisdiction over everyone on earth, and has the de-facto power to at least indirectly impose it upon a large plurality of earth's inhabitants.
"Wants to" + "sort of able to" is FAR more dangerous than "Indifferent to" + "generally incapable of (unless you're Chinese)"
$300G? That's basically one NY to LA round trip on NetJets. Or 3-6 regional one-way flights.
Will people please stop fantasizing about cities like NY, DC, and SF flooding? It won't be allowed to happen, regardless of what might happen elsewhere. All of those cities have relatively narrow paths to the open ocean that can, and certainly will, be blocked with dams to hold back rising sea levels.
And most of Florida's urban coastline will be fortified & raised with new crushed limestone and/or concrete as hurricanes progressively destroy it storm by storm, until Miami (and much of the rest of Florida) literally IS a "concrete jungle" (on a manmade hill).
Will sea-level rise devastate natural ecosystems? Of course. Will it destroy cities? No. Or at least, not permanently... new stuff fortified against climate change will replace whatever gets destroyed, and big cities will remain largely where they are... just taller, on raised terrain, with a hell of a lot more concrete.
No, I'm not contradicting myself. You're twisting my words.
A jackknifed 18-wheeler perpendicular to oncoming traffic is not equivalent to an 18-wheeler driving across a normal 4-6 lane road at an intersection. The former is a rare edge case that isn't supposed to happen, and when it DOES, it's almost always a freak accident that creates a situation so dangerous, EVERYONE in the area is at risk of injury or death. The latter is something that happens CONTINUOUSLY whenever an 18-wheeler is driving down a normal road.
The Tesla accident involving the autopilot slamming into a stopped emergency vehicle parked along the shoulder actually strengthens my original point... stopped vehicles alongside a limited-access freeway -- even in the emergency lanes -- CREATE a significant hazard for both the parked vehicle's occupants AND people driving on the road. Some parked vehicles are necessary, like emergency vehicles. Nevertheless, that doesn't mean we're safer if we intentionally create situations where MORE cars are encouraged or required to pull over and stop. If a guy with semi-autonomous autopilot falls asleep at the wheel, the safest course of action is to just keep driving while attempting to wake him up (preferably, while buffering the car's controls from sudden movements he might make if startled), even if he wakes up to find himself 42 miles past his intended exit. Everyone -- human and autonomous vehicle -- is safer when cars on a freeway behave in an outwardly-rational, predictable manner. Everyone is in danger when vehicles deviate from the predicted norm in shocking or surprising ways.
US-27 is most assuredly NOT a limited-access freeway. The truck Joshua Brown's car skidded under was perpendicular to oncoming traffic. That would not happen on a real freeway (or at least, if it did, it would only be if there were some horrific accident involving a jackknifed truck that would probably result in the deaths of HUMAN drivers, too).
Repeat after me: constraints. A limited-access freeway is a relatively easy case to design for. A rural highway with high-speed traffic and traffic crossing at grade is a safety nightmare for autonomous vehicles.
On a limited-access highway, "coming to a stop" is usually WORSE than "staying in the lane & maintaining a normal cruising speed". Autopilot lanekeeping & collision-avoidance is now better on average at avoiding accidents than most human drivers. It's UNUSUAL situations that create the danger.
A car staying in its lane & moving appropriately is the norm on a freeway. A car stopped on the shoulder creates an active road hazard for everyone else. Remember, his autonomous vehicle wasn't the only one on the road. It's not a safety improvement if your "safety feature" creates a situation that's MORE dangerous than letting the car just do what it does best -- follow the lane, avoid collisions, and behave in a predictable manner.
If you really need to punish drivers, add warning lights to visually communicate to other drivers that a car is operating without active human control. On a limited-access highway not under construction in good weather, inattentiveness with autopilot is a statistical non-issue... and in the real world, it's probably a net improvement over human drivers who are semi-distracted. On a non-freeway that has at-grade cross traffic, it's a genuine hazard. In bumper-to-bumper city-street gridlock, it's a non-issue (the car can stop within inches if necessary). On city streets at 30mph, it's likely to be dangerous.
The point is, there is no blanket "one size fits all" rule. Autonomous systems have known constraints. Stay within them, and you're fine. Deviate, and you create problems. Ignore the constraints entirely, and you're in uncharted territory risk-wise. Following lines & not colliding is easy. Finding a safe place to pull over is enormously harder.
The thing is, if you're a business, depositing cash in a business checking account usually incurs fees of its own:
* Do you have an employee carry the cash to the bank? HUGE insurance premium, because you're going to get sued for a small fortune if that employee gets mugged on the way there.
* Do you pay an armored car company to pick up the deposit? This costs money.
* Are you a sole proprietor who carries your own deposits to the bank? You're probably insane, because eventually you WILL get mugged by someone.
OK, one way or another, you get the cash to the bank and hand it to the teller. Guess what? Counting fees, cash-deposit fees, and additional surcharge fees for coins. If you're a business, banks constantly hit you up for fees in BOTH directions.
Cash is only "free" for a business to accept if it turns around and directly uses that cash to pay its employees, vendors, rent, and the owner takes the remainder home and deposits it straight into his own personal bank account. And if you do this, your accountant is going to absolutely hate you & charge you more for their services.
That's interesting, I didn't know the Volt worked that way.
Does it somehow make NORMAL gas last a year, or does the "one year" assume you added stabilizer at the same time as the fill-up? Or does it have a separate tank for stabilizer, and add it automatically a few days after your last fill-up?
Or worse, they'll deploy 5G radio technology that only works on a few AT&T-branded phones, without meaningfully increasing the amount of fiber backhaul available to back up the new RF bandwidth.
It's kind of like if you had 50 people connected to an 802.11ac access point, but the access point itself were connected to a DSL modem with bottom-tier service.
Except... it's actually worse. At least with the 802.11ac access point, users on the same LAN could get 100+mbps speeds peer to peer with each other, regardless of how fast or slow the internet connectivity is, because the local access point is (usually) smart enough to RECOGNIZE traffic that doesn't really have to touch "the internet" itself. AFAIK, *no* mobile network even tries to route peer to peer traffic among users connected to the same tower site directly (without involving a round trip to the off-site data center), so EVERYTHING ends up being backhaul-constrained, even when there's technically local wireless bandwidth to burn.
Being Devil's Advocate for a moment... you might get 54mpg "average" when driving under mixed urban-suburban typical daily conditions, but what kind of mileage would you SPECIFICALLY see on days 2 and 3 of a drive from Miami to Los Angeles where you spent basically 100% of your driving time on an interstate doing 80mph... the kind of drive that's long enough to deplete the Prius' batteries, without really giving them a good opportunity to recharge (because the engine is running at full efficiency keeping the car moving horizontally at 80mph, and doesn't really HAVE any surplus energy to divert to battery charging under those conditions... especially if the air conditioner is running on max & you have a car stereo system with a huge amp(*)).
(*) The "huge amp" part isn't entirely hypothetical... when I was younger, I destroyed my first car's battery on day 2 of my first cross-country road trip. The amp was a 50x4+250 watt (RMS) class-D amp, and basically it drew more power than the stock alternator could actually supply. On day #2, I stopped to get dinner, and discovered that the battery was too drained to start the engine afterwards (and in fact, was already damaged from being severely discharged past the point of no return).
The problem isn't that people aren't buying passenger cars per se, the problem is that SUVs and trucks are the only market where American carmakers can profitably compete against foreign competitors. In countries where "normal" people drive tiny econoboxes, an American F350 pickup truck or a Cadillac Escalade has EXACTLY the same kind of prestige and brand image that a BMW or Lexus does in the US.
This is the exact problem Chrysler ran up against ~10 years ago when Daimler owned them. Within a few years, Daimler transformed Chrysler into a company whose cars were every bit as good as the best German cars, both in quality and design. The problem was, Chrysler had close to zero brand prestige in America OR abroad (at least, for passenger cars), so they were in a no-win situation... they had cars that were as good as a premium German car, but couldn't sell them for enough money to justify that level of design, performance, and quality. So they went back to making mediocre cars with tolerable profitability.
Chrysler DID come up with a very cool, uniquely American application for hybrid electric-ICE traction... instead of using the electric motor solely to improve fuel economy, they ALSO used it to provide a short burst of additional horsepower in conjunction with a turbocharger. See, Chrysler has always been a big fan of turbochargers, but there's a reason why they've never been terribly popular with the mainstream... it takes a second or two to kick in when activated. Someone at Chrysler realized that electric traction motors aren't really useful for SUSTAINED augmenting of an ICE's horsepower (it overheats & damages the battery), but you could still use it to BRIEFLY provide surge power to offset the turbo lag (ie, slam the pedal, electric traction kicks in to add power while the turbocharger spins up, then the electric traction motor backs off as the turbocharger kicks in).
The idea was so cool, Porsche licensed it to use with THEIR OWN future hybrids. But Chrysler just couldn't pull off selling it in America. In America, Chrysler's big money-maker is muscle cars... a market segment that only cares about turbocharging as a way to make an absurdly-overpowered 12-cylinder engine act like a 16-cylinder engine. For that segment, eliminating turbo lag is almost moot, because their cars have so much power to begin with, the turbocharger is just another bullet point on the 'brag' list to add another hundred horsepower or so that will never actually be used.
Personally, I think one area where American automakers could have pulled off "mostly" electric cars would have been to make one with an all-electric drivetrain, but a small diesel engine with generator whose only purpose was to run at constant speed, with maximum efficiency, and provide enough electricity to augment the battery & double the car's range on a full charge.
Carrying your own diesel-powered battery charger isn't particularly efficient (roughly half the mileage per gallon as burning it directly for locomotion), so you wouldn't want to use it as your car's primary source of energy (vs charging), but it would have been a great stopgap measure for people who wanted to drive an EV across America back when rapid charging stations were few and far between... you might have gone for weeks or months at a time without ever firing up the diesel charger, but when you had to make that trip from California to Texas (for example), you could have fired it up and extended your EV range enough to not have to plan your entire trip and meal/rest-stop schedule around your car's battery capacity.
Why diesel? Because regular gasoline turns into varnish after a few weeks... and by definition, this would be a feature you'd really only want to use every few weeks at most (and possibly just once or twice a year). Diesel can survive months of disuse. Unleaded gasoline can't (at least, not without the kind of active effort and ongoing maintenance that experience shows American car owners simply can't be expected to perform with any degree of re
Not a 6502, but I'm pretty sure the TI99/4a used SRAM for its 256-byte "scratchpad" RAM (the only RAM its CPU could access directly).
I know the Amiga used "Static Column" RAM, but I think SC ram was what we'd NOW consider to be "PSRAM" -- DRAM with extra onboard circuitry to do its own refreshing automatically so it "looks" (and behaves) like SRAM as far as the outside world is concerned.