I feel similarly to you (I have a m4800... I really wanted a m6800, but at the time the m6800 had a "5-8 week" lead time). Current laptops just aren't inspiring.
The main things that will finally motivate me to buy my next laptop:
* 17-19" primary display... preferably 3840x2160, minimum 2560x1440... w/gSync and 144hz+ max framerate. This time around, I'm not compromising on the gSsync requirement (I'd accept freesync, but my other non-negotiable requirement is NVIDIA RTX... and AFAIK, NVIDIA are still assholes about refusing to support freesync).
* ability to run single-core sustained at 4+ GHz. Not "for 5 seconds at a time during heavy workloads before throttling because it has inadequate heat-removal", but real "4+ GHz balls-to-wall for hours at a time when plugged into a wall outlet and power is a non-issue".
* NVIDIA RTX graphics card... preferably, non-Quadro (Please. Give us at least ONE goddamn discrete video option that's a top of the line non-Quadro RTX). RTX is non-negotiable. I want hardware raytracing.
* Two 2.5" drive bays AND at least one slot (preferably two) suitable for a SSD. Or at least one 2.5" drive bay capable of taking a 15mm drive, one slot suitable for a SSD, and a removable optical bay that can take at least a 7-9mm 2.5" drive.
* Thunderbolt 3. I might never use it, but dammit, this is the one loophole that partially opens the door to a future thirdparty x16 graphics card 3 years down the road when the built-in card is hopelessly obsolete for VR development (even though x16 is itself a fairly huge limitation) & gives me a way to limp around most of the OTHER things the manufacturer decided to insidiously cut corners on that weren't necessarily obvious or deal-killers at the time of purchase.
* 802.11ad with 60GHz wi-gig certification. I do VR development, and put the likelihood at ~90% that I'll encounter a need to stream low-latency video wirelessly to a headset at some point over the next few years.
* A keyboard and pointer stick that's at least as good as the one on a m4800... which honestly, still kind of sucks compared a real Thinkpad. Sigh. Fantasy shopping list aside, I'll probably end up buying whatever is Lenovo's best w-series mobile workstation when the time comes, because they're the only laptops left with keyboards that don't completely suck (and even they're a pale shadow of their former glory).
There's a practical limit to how much info you can put in front of you at once before it either becomes too dense to easily read, or so huge that you have to start straining your neck looking up to see it. Three monitors in an arc at approximately arm's length (with hand in a fist), with the top of the monitors approximately equal to your "straight ahead" gaze and the bottom of the monitors at your comfortable "tilted downward" gaze is pretty much IT. Bigger monitors allow you to move them farther away (handy if you have presbyopia), but having three 60" monitors won't really enable you to see MORE within easy gaze-range than having three 24-27" monitors will.
Three discrete displays are handy because they allow you to easily manage maximized windows and fling them from side to side, while viewing different sets of information side-by-side. It's also the practical limit of what you can interact with using a normal mouse and a normal amount of desktop space (more monitors and/or less space requires excessive mouse acceleration, which causes more problems than it solves).
Basically, the ideal (IMHO) is a ~120-150 degree arc of three virtual monitors at roughly arm's length away, located between "straight ahead" and "comfortably downward-gazing", with optical resolution comparable to three 27" monitors with at least 2560x1440 resolution... a big, wide virtual monitor in the center for the IDE, a second virtual monitor off to the left with an aspect ratio suitable for displaying side-by-side page views from a pdf ebook, and a third virtual monitor to the right that might be 3:4 aspect ratio (mine is 16:10, but 90% of the time I have the window on it only filling 2/3 the width).
IMHO, the most viable way to limp around the focal length problem is to treat it like simulated nearsighted presbyopia (inability to focus on things closer than 2-3' due to age, blurriness beyond ~10' due to nearsightedness) and use what are -- for all intents and purposes -- some variant of bifocal lenses in the headset.
Or... simulate the same combo, but with single-vision lenses and added astigmatism. As my eye doctor explained when I had PRK years ago, a small amount of astigmatism can actually come in handy later in life by compensating for the reduced focal depth. Basically, trading text that's crisp & sharp over a limited focal depth for blurry text that's readable over a wider range.
Moore's law hasn't ended... it's just been constrained by consumer expectations that new hardware will be cheaper and use less power.
Moore said only that the number of transistors doubles every 18 months. For the past 10 years, that doubling has been held back by consumer expectations that costs, power consumption, and emitted-heat will decrease. If Dell could sell millions of computers with the inflation-adjusted price of a 1991 Amiga 3000 (~$2,500 in 1991, correlating to approx. $7,000 today) that emitted 1,800 watts of heat (using active-refrigeration chilled-water cooling to prevent them from cooking themselves, along with some kind of outdoor condenser like central AC uses to pump all that heat outside), we'd be casually playing with supercomputer hardware that would make NOAA and NASA jealous TODAY.
Seriously. Think about how many watts a 3GHz Pentium 4 emitted with its single core... and multiply that by 16-32 cores to get an appropriate brute-force modern power budget.
No, it's not really feasible, if only because most people aren't going to install (and pay to operate) the equivalent of a 2-ton central AC unit to keep their computer from turning their home office into an Easy-Bake Oven, but it should be food for thought. Moore's law is far from dead. It's just that keeping up with it isn't quite as consequence-free as it used to be.
And if you need proof about what's possible when you don't care about cost, energy use, or heat-removal, read about professional bitcoin mining operations. That's TRUE hardcore brute-force computing power.
There's a reason for the lack of AAA titles that goes beyond mere economics -- it's fucking IMPOSSIBLE to efficiently do VR development with current hardware and workflows. As the developer, you're CONSTANTLY putting the headset on... taking it off... putting it on... taking it off. And waiting... a lot. The integration of development tools into the headset environment itself is practically NONEXISTENT today. It's a definite, and very real handicap.
At least Rift has the advantage of using the same host PC to run the software that you're developing on, so the main limiting launch-to-view constraint is "how quickly can you put the headset on or remove it". With Android, VR development is downright excruciating... you can get piss-poor previews that are the equivalent of using the Android Emulator with the headset as a blurry remote display (with all timing completely shot to hell), or you can launch build+deploy and twiddle your thumbs for 30-70 seconds waiting for it to compile, upload, and launch.
What we REALLY need is a Rift-type display that can do the equivalent of overlay three virtual 27" monitors in an arc ~20" in front of you, with realtime camera-vision of the rest of the room, so you can develop without having to actually take the headset off (and sufficiently high resolution & optics so that you won't feel like someone who's legally-blind trying to sit at a desk and read those same three monitors in real life). Aside from improving game-development workflows, a capability like this would also give people who aren't even INTO gaming to invest in a headset, because it would let you have the equivalent of three large monitors in an arm's length arc around you in places where you CAN'T have three monitors (like on a plane, when traveling, etc).
What I'd really like to see: a headset that starts with the basic idea of an Oculus Rift + ZED Mini (two cameras that clip onto the front of a Rift to enable AR-type applications), but taken to the next level:
* Extremely high-resolution displays... 3840x2560 per eye... with freesync or G-sync framerates up to (and including) 100hz & 120hz.
* Equally high-resolution cameras, with two independent video paths:
1. A low-latency path that's basically analog and genlocked. This is the video of the room around the user that gets seen directly. Basically, anyplace where the GPU-generated video is "bluescreen green", this video gets overlaid.
2. A second, "digital" path that's available for the usual "AR" type purposes -- image-recognition, computer vision, environment-mapping, etc.
Path #1 allows the user to remain grounded in reality, and avoids the worst effects of "simulator sickness". Your brain can tolerate sloshing & swimming details, as long as the bulk of your surroundings are in fundamental agreement with what your other senses are telling it.
Path #2 allows for Augmented/Mixed-reality content, without compromising Path #1.
Ideally, the headset would support 802.11ad, so the relatively uncompressed digital bitstream from the cameras could be thrown directly at a server running on the local LAN to do the heavy-lifting analytical work.
Why this is important: current camera-video pipelines just plain have too much latency to use as an augmented-reality "environmental" view. I know, because I've written programs that try to take the video stream directly from the camera on an Android phone & render it straight to a Unity RenderTexture. They all lagged so badly, it was worse than having no view of the room at all. I concluded that the only solution that involves camera-to-video instead of holographic video overlaid on a directly-viewed scene (like MagicLeap or Hololens) would require cameras with an extraordinarily low-latency pipeline as described above, combined with high-resolution displays with high native framerates (ideally, 100 and 120fps... or better yet, 200fps and 240fps).
I'm highly-sensitive to simulator-sickness... but I'm pretty tolerant of traditional motion-sickness. In English, that means I have close to zero tolerance for latency, blur, and eyestrain with regard to my "environment", but can tolerate a fair amount of higher-order artifacts WITHIN that environment as long as the environment AS A WHOLE is consistent with what my other senses are telling me. The solution is to START with getting my environment to be acceptable when viewed indirectly (by looking through lenses at screens displaying realtime live video captured by cameras on the other side of the screen), and THEN overlay the synthetic video. Current VR solutions go at it "backwards".
If you're talking about single-family homes on large lots vs 16-28 foot townhomes of comparable size, yes... it might increase availability and reduce costs.
If you're talking about 50+ story skyscrapers, no... it won't. Skyscrapers are SO expensive to build (per square foot of habitable dwelling), the economics just aren't there to build anything less than luxury condos for the ultra-rich. If the developer is ALREADY spending $300/sf, adding $200/sf for premium finishes and doubling the selling price is a no-brainer... if only because someone who can afford $300/sf isn't going to spend that much for the equivalent of a doublewide trailer in the sky.
The catch is, places like the Bay Area are already "built out". The swampy land adjacent to the bay that's currently mud flats is only suitable FOR building skyscrapers, because only skyscrapers have the structure & foundation to rest on deep bedrock & survive earthquakes (Florida-style landfill won't work there... the first time there's a major earthquake, the landfill gets sifted, re-compacts itself, and whatever was sitting on it collapses). People who could AFFORD to buy a hypothetical skyscraper along the bay near somewhere like Mountain View don't WANT to live in a skyscraper adjacent to the bay near Mountain View, because it smells like rotting garbage & isn't particularly attractive... they want to live in a house on the side of a cliff overlooking the whole valley, or in SF itself.
And because the whole area is already built out, the only way to get land to build townhomes and apartment buildings is to buy existing structures, knock them down, and get permission to rebuild at higher density... which the remaining neighbors will fight tooth and nail. So what REALLY happens is, someone buys a block with 24 old single-family homes, knocks them down, and replaces them with half as many mega-mansions -- effectively, making things even WORSE. Or they'll buy an old apartment building with 50 units and poor residents, demolish it, and rebuild a new apartment building with 60 units that cost twice as much as the 50 units it replaced.
Basically, the Bay Area's densification ship has already sailed, and for the most part it's too late. At least, until the next major earthquake. If a city like Palo Alto were 90% flattened by a horrific earthquake & almost everything had to be rebuilt anyway, it MIGHT be able to get residents to approve a deal that would let them have their cake & eat it too by allowing them to sell half of their lot to someone else (or replace their original one-story single-family home with a three-story duplex sharing a wall & sell the other one)... but after a catastrophic disaster, local governments are rarely in any position to pull something like that out of a hat.
True story: 50 years ago, when US-1 was six-laned through southern Miami, the section near SW 152nd Street was split into two one-way roads because there wasn't enough room to widen it to 6 lanes along the original route. In 1992, Hurricane Andrew destroyed 99% of the structures in that area beyond repair. The area was ultimately rebuilt from scratch, and US-1 is STILL a pair of one-way roads, even though the original problem (lack of land) was solved by Andrew's destruction. Why? All the people who LATER bought property along the northbound lanes (shifted a block east from the original route) paid a premium for that land because it WAS adjacent to a major highway -- shifting the northbound lanes would have profoundly harmed the value of their land (and gotten the state & county sued). And FDOT still didn't own enough land adjacent to the southbound lanes to do it, even though the land was now stripped bare. Put another way, city planning decisions made long ago tend to affect future land use, even long after the original rationale has been forgotten or ceased to matter. It's the same reason why London's streets are a medieval rat's nest, even in areas that were bombed into oblivion by the Luftwaffe during WWII, long after the original structures that dictated their layout (or cemented it in place) ceased to exist.
5mph RUBBER bumpers. Painted bumpers are BULLSHIT... graze something, $500 minimum paint job to fix. I went out of my way to get black rubber bumpers on my truck, and it's the best thing I ever did. Small scuff? Fine-grit sandpaper, 2 minutes of rubbing, fixed.
The "small base" problem is partially mitigated by Unity and Unreal. They aren't quite "write once, run everywhere", but they DO make it possible to "write once, port quickly".
That's actually a harder problem to solve than you might expect.
Problem 1: what is the camera looking at? You can't just follow what the user is looking at... it would be like watching a bad camcorder video shot by a drunk person. To write the virtual camera positioning code, you need somebody who actually understands cinematography & can apply that knowledge to the user who's lurching around inside the VR world (keeping it pointed at the most interesting content, zooming as necessary, but nevertheless keeping it relatively stable to avoid making everyone watching it nauseous).
Problem 2: you need yet another GPU pipeline to render the "public" view... in a system whose video cards are ALREADY getting pushed to the breaking point, with no ability to add more. We desperately need the ability to pack four x16 video cards into a PC. Right now, the market for higher-end video cards is too niche for anyone to make halfway-affordable higher-end cards... but to achieve what you need with only a pair of cards, you need REALLY EXPENSIVE cards. If we had 64-bit PCIe, you could instead buy four cheaper cards (with greater economies of scale) instead. Or you could buy two cards, then buy two more after you realized the first two weren't adequate.
Fort Lauderdale doesn't HAVE to be pedestrian-friendly to appeal to Brightline riders who work in Miami. Many of them might end up paying a super-premium to live at Flagler Art Village within walking distance of the Brightline station, but for everything else they'll just drive. Brightline is their loophole that enables them to have a well-paying job in downtown Miami that demands long hours, without having to further endure actually LIVING there. Don't get me wrong... Miami is fun for a few years... but after a decade or two, the endless 24/7 gridlock and general dysfunction everywhere around you just really gets tiring and old. Especially when it sinks in that it's never going to get better during our lifetime, and will only keep getting worse. Even moreso when you realize that almost all of your friends have thrown in the towel and moved to Fort Lauderdale, and you're now driving up there all the time.
Downtown Miami is "pedestrian friendly" in the sense that it's relatively easy to use Metromover to get to 80-90% of the places where well-paid people are likely to work... and people who LIVE in downtown Miami don't EXPECT to be able to do things like shop there. It's just taken for granted that if you live downtown, going shopping means driving 10 miles to Dolphin Mall or International Mall, or taking Metrorail 8 miles to Dadeland Mall.
Put another way, you don't live in downtown Miami because you crave an urban lifestyle... you live in downtown Miami because you make lots of money, have no free time because you work 60-80 hours/week, and endure it because it means you can get to work in 10 minutes instead of 90-120 minutes each way.
Moving to downtown Miami IS known to be actively harmful to your social life. Why? Stupidly-expensive parking (downtown Miami has more parking per square foot than some outlet malls, but most of the spaces are owned by a cartel that uses artificial scarcity to drive up prices... it really, really sucks paying $20 to park on a Friday night in a half-empty garage surrounded by TOTALLY empty garages). If you move to downtown Miami, none of your friends will come over to see you anymore. Or they'll do it once, spend the evening complaining about how expensive it was to park, and never do it again.
Another perk in favor of Fort Lauderdale for the car-averse... if you want to take Uber to the beach, Fort Lauderdale beach is a relatively cheap 2-3 mile trip down Las Olas or Sunrise Blvd. Getting to the beach from downtown Miami without a car is significantly more expensive (Biscayne Bay is wide, the causeway meanders, and gridlock drives up the cost even higher).
In downtown Fort Lauderdale, the transit sucks... but most of the places you'd want to get to via Uber are within 2-3 miles. In Miami, the transit sucks... but everything you'd want to get to via Uber from downtown Miami is FAR AWAY and EXPENSIVE to get to.
There IS (sort of) a way... make it look like the larger environment is stationary & perfectly-synchronized with head motion (low latency, zero jiggle/slosh), and limit the moving stuff to small elements within the larger scene. Your brain can deal with small things that wobble/jiggle/slosh, as long as the "big picture" is consistent with what you see. It's when your entire WORLD is seemingly in conflict with your senses that you get motion sickness.
Eyestrain is a related, but different, problem that's currently a lot worse than it really *has* to be, simply because most devices have piss poor calibration & lenses. Part of the calibration problem can be resolved by giving each eye its own optimally-placed display (with precise orientation), and most of the remaining problem could be resolved by involving opticians to fit & customize lenses for individual users. It could be something simple, like designing a future Oculus device to ship with standard lenses, but be able to easily swap them out with custom lenses fitted by opticians certified by Oculus as being "VR-aware".
There's also the fact that "rendering optics" today are often where "game physics" were circa 1995. Libraries try to matrix-out distortion, without fully understanding what they're doing, what has already been done, and what's going to be done further down the pipeline. That's a big reason why VR companies have finally started to hire opticians & involving them in the process (and using them to help design better lens designs).
VR requires enormous amount of integration at every level. Comparing "First Person Shooter" to "Immersive VR universe" is kind of like comparing "Java app" to "distributed J2EE app with rich native client". The problem's scope is ENORMOUSLY bigger & requires more resources & more sophisticated management.
Right now, Brightline charges $20-40 each way for travel between Miami and West Palm Beach, and $15-30 for travel between Fort Lauderdale and either Miami or WPB. There aren't a lot of people traveling between WPB and Miami daily, but there are a LOT of people using Brightline to travel between Fort Lauderdale and the two cities. Pretty much every affluent professional who lives near downtown Fort Lauderdale & works in downtown Miami or WPB now takes Brightline, simply because it literally cuts the travel time in half (and allows you to spend it playing with your computer, instead of sitting frustrated in gridlock).
For a better idea of how successful it'll be, check back in another year or two... after it's been around long enough for people to start making "where do I want to live" decisions based upon Brightline's existence and travel times, as opposed to "where do I want to live, if living in Fort Lauderdale means I'm going to have to endure a grueling 60-90+ minute drive to work each way, and commuting daily between Miami and WPB would take ~1.5-2 hours on a GOOD day at 8am or 5pm.
Considering how INSANELY expensive rent is anywhere in the Bay Area, I think Caltrain could easily find tens of thousands of people willing to fork over $50-100/day if it meant they could pay $2,500/month for a 3 bedroom townhome with garage and back yard, instead of paying $4,000/month for a one-room efficiency. They might still pay the same total amount per month, but they'd get a lot more for their money.
Trains don't HAVE to be faster OR cheaper than cars or planes to be used... they just need to be perceived as nicer. Consider Eurostar... planes are almost always cheaper than Eurostar (especially walk-on Eurostar fares), but anybody who can afford it unhesitatingly takes Eurostar because it's almost as fast as flying (when you consider "terminal time"), and WAY nicer & less-stressful than flying. Ditto, for Acela between DC and New York. The proles fly from Dulles to Newark & spend 2-3 hours just getting to and from the plane itself. Acela passengers spend more time in motion, but half the total time getting to wherever they're ultimately going (and enjoy substantially more unstructured free time during their trip, vs air travel's "hurry up to actively wait".
There might be a HSR track from Hong Kong into mainland China via Shenzhen (which currently has a branch line to Zhuhai & extension to Macau under construction), but even if the train were allowed to run nonstop through mainland China & skip the border formalities besides those of Hong Kong or Macau, it would STILL be a 1.5-2 hour trip because it's so far out of the way compared to the new bridge/tunnel.
Plus, a train that ran directly between Hong Kong & Macau could probably relax some of the border formalities that are seen as necessary for trains passing through mainland China. Neither Macau nor Hong Kong might want to be swarmed by mainland residents, but I doubt whether they have much to care about with regard to EACH OTHER since they're both roughly equal in terms of wealth & freedom.
In American terms, rotate the map 90 degrees clockwise, and pretend that Macau is St. Petersburg (Florida) and Hong Kong is Bradenton. Yeah, in theory, you could drive from downtown Bradenton to downtown St. Petersburg by heading east to I-75, continuing north to Brandon, then slogging your way west through downtown Tampa over to St. Petersburg & driving south... and if you were extraordinarily lucky, you might even be able to make the trip in 90-120 minutes. But if you took I-275 and the Skyway Bridge instead, it's a 20-30 minute trip. Without the Skyway Bridge, Bradenton and St. Petersburg would be practically inaccessible to each other. With it, they're next-door neighbors. All over America, there are city pairs that USED to be mutually-inaccessible (at least, in any sane amount of travel time) that turned into casual driving distance once a bridge over some wide-but-shallow body of water got built. Ditto, for Europe... consider Copenhagen vs Malmo. The ferry always existed... but back when the trip took hours, the idea of living in Malmo and commuting to Copenhagen would have simply been LUDICROUS. Now, it's something people seriously consider as a daily option. The same thing is likely to happen when/if Italy finally gets around to building a bridge to Sicily.
Do you want to know the REAL difference between "boondoggle" and "valuable transportation infrastructure"? About 50-100 years (at least, insofar as COMPLETED infrastructure projects go... projects that get started & abandoned prior to being completed to useful length don't count).
In theory, the NEC was a "boondoggle". The Pennsylvania Railroad never made a profit on it due to the astronomical cost of maintaining it as electrified railway. Yet it's the reason why today, you can casually make day trips between New York and Washington without burning most of the day. Infrastructure has a funny way of doing that... hemorrhaging money at launch, losing money for years... then gradually, over time, becoming so important that disruption for even a few weeks would be DEVASTATING and UNTHINKABLE (hence, most of Amtrak's current problems with the NEC, and the London Underground's problems... they desperately need to do basic maintenance work on century-old tunnels, but can't shut them down, so they have two choices... defer maintenance as long as they can, or build a third tunnel at staggering expense so they can always keep at least two in service at any time because total shutdown is unthinkable).
Fifty years ago, Central Florida newspapers absolutely EXCORIATED FDOT for approving Interstate 4 -- a freeway from "nowhere" (Plant City) to "nowhere" (Lakeland). Nevermind the fact that even back then, it was fully intended to eventually span from downtown Tampa to Daytona Beach, and the initial segment between Lakeland and Plant City provided a detour around two crowded small towns that would have merited the construction of at least a 4-lane divided highway as a bypass around them for Tampa-Orlando traffic ANYWAY. If you read old newspaper articles, you'd think I-4 was proposed to be just a useless freeway between two small towns in Central Florida.
Ditto, for the Florida Turnpike between Orlando and Fort Pierce (the original "Turnpike" plans were for a "Y"... Miami to Fort Pierce, then one branch to Orlando (which is now Turnpike) and one branch to Jacksonville (which ended up as I-95). Newspapers and politicians fought bitterly over it, and pretended that then-existing roads between Miami and Orlando (US-27 through Sebring & Winter Haven) were even REMOTELY adequate (they weren't).
Back when "Alligator Alley" (I-75 across the Everglades) was first proposed by FDOT, the AAA fought it BITTERLY and did everything it could to get the project canceled for some insane, unfathomable reason. If anything, FDOT's only mistake with Alligator Alley was initially building it to a relatively low-standard narrow 2-lane deathtrap, instead of going full-on total freeway from day one(*).
Even Miami's Metrorail ended up being useful, despite Miami-Dade Transit's absolutely BREATHTAKING incompetence and short-sighted decisions (like cutting back Metrorail service to once/hour on weekends... at a point when their ridership was at the highest levels in its history... then scratching their heads and wondering why literally HALF of their weekend riders said 'fuck it' and disappeared. Unbelievable.) It took almost 40 years, but today the biggest limiting constraint on daily Metrorail ridership in Miami is the system's appalling lack of parking capacity at the 3 southernmost stations (thousands of people drive to the station intending to take Metrorail to go downtown, discover that the garage was full long before 8am, and end up having to drive downtown anyway because there's nowhere else to park).
Sometimes, politicians gamble and lose, like Cincinnati's subway. But you could equally argue that if Cincinnati had actually COMPLETED it and opened it for service, it probably WOULD have ended up influencing development and becoming useful 40-50 years down the line, too, even if it ended up being mostly useless for the first 25 years of its existence. That's just how transportation infrastructure IS... you build it through hinterlands, people spend years flipping land around it, then eventually people start buil
Considering China's enthusiasm for HSR, it does seem bizarre that they didn't just design another ~25 feet into the whole thing to make room for a pair of HSR tracks while they were at it. The marginal cost would have been fairly low, and it would have turned HK & Macau into the equivalent of a casual crosstown trip for each other.
A big chunk of Phase 1's cost is the tracks between Caltrain & BART @ San Jose and somewhere around Modesto, which will get dual-use... LA-SF, and also SF-Modesto commuter rail. It's needed regardless of LA-bound trains. LA-SF is just frosting on the cake... regional commuter rail is the big prize, because it'll open up the Central Valley as East Silicon Valley.
Likewise, Brightline intends to extend Las Vegas-Victorville to Palmdale to connect with CALHSR, so you'll also be able to take the train from SF to Las Vegas (probably BEFORE it's 100% HSR the last ~50 miles into L.A.).
Worst-case, 50 years from now, CALHSR repurposes the bridge structures for vacuum supersonic maglev. The important thing will be ROW-preservation. In 50 years, central valley land will be as expensive & urban as the bay area is today, and CALHSR will be viewed as a spectacularly fortunate real-estate investment, if nothing else.
And... how many accidents do you think Tesla's system PREVENTED? Let's be honest... most of the people who use Tesla autopilot wouldn't exactly be the world's most attentive drivers WITHOUT it, either. Even IF they were going through the motions and pretending to pay marginally more attention, we're talking about a group that's generally oblivious to anything that doesn't capture their immediate interest, tends to daydream a lot, and are almost the textbook case-study poster-children for highway hypnosis.
Let me repeat. At least with Tesla Autopilot, SOMETHING is paying attention to the road while they're behind the wheel, which is generally more than you could say about the "drivers" themselves in real life. Forget driver's-ed fantasy and propaganda... in real life, Tesla Autopilot is a life-saver (or at least, a major paint-saver and dent-preventer) for the majority of its daily users. It's not 100% perfect... but the drivers themselves are FAR worse. And that's why insurance companies grudgingly tolerate it... they know that at the end of the day, drivers likely to use Autopilot are going to have fewer accidents overall than they would have had if it weren't available.
An individual IC might not be "repairable" in any meaningful way, but there are also very few modes of failure where a literal single IC fails independently of nearly everything around it. Unlike, say, the screen on a phone, tablet, or laptop (which can break, suffer cable or interconnect failure, etc), batteries (which can, do, and are 100% EXPECTED to degrade over time), mechanical switches & fans, anything directly exposed to 100-240v AC or involving electrolytic capacitors, etc.
If a vendor wants to use a proprietary, purpose-built part... fine. As long as they publicly document enough of it to enable a comparably-sophisticated manufacturer to make compatible second-source replacement parts, with straightforward FRAND licensing terms. Maybe even require vendors to disclose itemized BOM costs per production run (registered with the department of Commerce, or its equivalent in China or wherever) & give statutory immunity from infringement lawsuits if a company manufactures & sells replacement parts when the original vendor either can't, won't, or charges more that 400% of its own BOM cost and/or unreasonably restricts availability.
A good start would be to look at laws passed in the 1960s & 70s to fight the auto industry's abuses... and the ways the auto industry initially tried to skirt them before largely settling down into the new regulatory regime. The electronics industry isn't identical to the auto industry... but it's not all that DIFFERENT from it, either. 98% of the issues consumers have now were issues faced by US car buyers in the past -- complex supply chains, proprietary parts, vendors seeking lock-in, restrictions on repair, and all. Regulations didn't solve 100% of the problems... but they DID fix the most egregious 90% pretty quickly, and the next 9% within a decade or so. And today, even automakers like Tesla (probably the most proprietary car in existence today) have made peace (or at least ambivalent detente) with third-party repair shops.
That would be silly... but what if you took the basic idea and applied it to laser-exposed microfilm? Sure, it might be slow to expose & probably require off-site developing... but once you had it done, you'd have a storage medium that could passively survive decades, maybe centuries.
The fundamental problem with any data-storage mechanism that requires active handling every few years to preserve it going forward is the fact that such data is unlikely to survive long enough to BE of interest to future historians. Businesses pay to retain records long enough to satisfy regulatory and immediate business needs, and that's IT. Governments are only slightly better at long-term storage for posterity. Just look at 1970s-era NASA and the BBC.
Old data that HAS survived to the present time almost ALWAYS survived ONLY because it was in a form that could be put somewhere, forgotten about, and rediscovered decades after it would have otherwise been systematically destroyed... and often, only because it WAS someplace that the people who would have wanted to see it get destroyed either couldn't get at it, or didn't realize it even existed. Like old letters in an old chest that went from attic to attic for a century or more. Or old tax records from Edwardian England or the Roman Empire. Or crates of punch cards from some long-gone business that were stored in an old office building in Detroit that was abandoned and left to rot decades ago.
Consider for a moment the fate of really, really old films from the early 20th century. Most of them were recorded on nitrate film that was literally dangerous. Few of them were considered valuable enough at the time to be worth transferring to "safety" film. The copies that exist today exist primarily because of a few film-buffs who were willing to disregard safety, grabbed them from trash heaps, and stored them in backyard sheds, until their grandkids discovered them decades later while cleaning out the house after their death & realized that they were a) dangerous hazmat that could literally go up in flames at any moment, and b) now worth a fortune, because all known surviving copies of them had been (intentionally or unintentionally) destroyed decades ago.
Museums and universities routinely have to deal with archived data that they don't have the resources to deal with RIGHT NOW, but nevertheless would LOVE to have some way to put "on ice" and just passively & compactly store somewhere indefinitely on the off-chance that some future historian might find it interesting. Or until AI gets to the point where historians can use computers to analyze, catalog, and categorize data that would take a hundred human lifetimes to meaningfully sort and categorize by conventional means. At the time the decision is made to archive it, the data might be very low-value, and won't BE sufficiently valuable to expend resources re-archiving for several human lifetimes. But if it's stored on a medium that can literally be packed into a mountain and forgotten about for a century or two, that's ok... it'll still be there for future historians when they eventually DO care about it.
I'd argue that a fair reform would be one that slightly raised the amount charged by USPS to deliver packages from a customs point... then made the same rate available to US companies willing to handle their own logistics up to that point. So if USPS charges $1.50 to deliver a package from their facility near JFK to someone in Miami, someone like Amazon or Walmart should be able to get the exact same rate by showing up at the USPS facility near JFK with 40 million packages per day.
Ultimately, it IS kind of fucked up that it costs less to mail a half-pound package from Wuhan to Miami than it costs to mail a fucking POSTCARD from Miami to Fort Lauderdale. A fair rate structure would be one that would be available to domestic mailers under identical terms, but wouldn't induce domestic shippers to do logistically-silly things to take advantage of it.
Right now, USPS rates are lopsided at BOTH ends... USPS *radically* increased the cost of mailing small, light padded envelopes ~2 years ago. Prior to that, you could mail almost anything physically-capable of fitting into an envelope that's within weight limits for the cost of first-class letter-rate postage. NOW, you get hit with multiple surcharges for "non-flat" and "more than 1/8" thick" packages that make it 2-4x more expensive. This new rate policy has utterly DESTROYED the ability of Americans to compete with Chinese sellers of things like small USB cables, microUSB-C adapters, etc, by driving the cheapest possible shipping cost for anything that's thicker than 1/8" and not consistently-flat up to $2-3/envelope. Merely raising the amount charged to Chinese shippers by 10-25c & using it to abolish the non-thin/flat surcharge charged to US shippers would go a long way towards restoring fairness & parity.
Is there any particular reason why nobody makes a product that's basically like non-LTH (phase-change/magneto-optical) BD-R, but on a flexible film substrate stored on reels instead of bulky discs?
The main problem I see with magnetic tape is that it's inherently susceptible to stray magnetic fields (including the Earth's poles). In contrast, phase-change media can theoretically have a passive lifespan that's measured in decades (centuries, if "being able to read it with normal, consumer-grade hardware as a normal OS filesystem" isn't a hard requirement, and you can deal with forensic data-recovery using exotic purpose-built hardware).
It won't eliminate celebrities... but it will reduce the number of living celebrities, and turn their kids almost into literal "royalty". Why? The film industry is weird... when an actor gets paid $50 million to make a movie, he or she isn't just providing $50 million worth of acting-in-front-of-the-camera value... the mere fact that they were PAID $50 million adds value to the film by getting others to take the film seriously. Kids from "Hollywood families" will go to the bargaining table able to offer not only their own services, but licenses to use the likenesses of their long-dead (or retired, if their parents want to help bootstrap their kids' acting careers) parents as well. Win-win for studio execs... they can brag about paying huge amounts of money to "an actor", but metaphorically get two (or three, or four) for the price of one.
Another possibility is that studios will still use live actors for much of the filming, but use post-production CGI to make their mouths match whatever language they've been dubbed into for a particular audience. This is a particularly big deal for high-framerate video. At 24fps, you can take a lot of liberties with lip-sync and mouthing of words. At 48, 50, and 60fps... it's more noticeable. At 100 & 120fps, any disconnect between what a native-speaker expects the lips and mouth to look like and what they see is going to be instantly and painfully obvious. With the right software and CGI, you could have a workflow like this:
* Shoot an episode of some TV show at 50-120fps in English, Mandarin, or whatever other primary language you choose.
* Record all the localized dialogue in different languages... quite possibly while recording the faces and mouths of the voice actor to capture additional nuance.
* Using the appropriate software, match up the origin actor's spoken dialogue with his or her mouth movements. Then, after analyzing the mouth movements of the voice actor for the target language, re-render the original actor's mouth to match.
Once again, lip-sync of dubbed dialogue is something that Hollywood and the TV industry have traditionally been able to somewhat dance around at 24/25/30fps (noticeable, but not necessarily a HORRIFIC problem), but becomes something you simply can't ignore at higher framerates and higher-resolution... and ABSOLUTELY can't ignore when you combine high-framerate AND high-definition. At 3840x2160@120fps, you'd better make sure not only that the spoken dialogue matches the mouth movements perfectly, but that the entire audio pipeline is PRECISELY synchronized with the video and room acoustics as well, or you'll find yet another novel way to fall into the "uncanny valley".
I'd argue the opposite point... the constraints imposed by present-day AR/MR systems generally match the limits imposed by age-induced presbyopia on someone who was myopic to begin with. If you can't comfortably focus on things that are closer than 15-20 inches without eyestrain (or at all), a 1-meter "near" focal plane isn't much of a limit anyway. Likewise, if you can't clearly see things that are more than 10 feet away without glasses due to residual myopia, a 3-meter "distant" focal plane is about as far away as you can see clearly without glasses anyway.
It's not middle-aged men and women who'll have a problem with the ML1's optics, it's KIDS who'll likely find it to be the most limiting. Which for now, is kind of ok, because middle-aged adults are just about the only people who can AFFORD to buy a ML1 or Hololens as a recreational toy.
Slashdot Mirror
I feel similarly to you (I have a m4800... I really wanted a m6800, but at the time the m6800 had a "5-8 week" lead time). Current laptops just aren't inspiring.
The main things that will finally motivate me to buy my next laptop:
* 17-19" primary display... preferably 3840x2160, minimum 2560x1440... w/gSync and 144hz+ max framerate. This time around, I'm not compromising on the gSsync requirement (I'd accept freesync, but my other non-negotiable requirement is NVIDIA RTX... and AFAIK, NVIDIA are still assholes about refusing to support freesync).
* ability to run single-core sustained at 4+ GHz. Not "for 5 seconds at a time during heavy workloads before throttling because it has inadequate heat-removal", but real "4+ GHz balls-to-wall for hours at a time when plugged into a wall outlet and power is a non-issue".
* NVIDIA RTX graphics card... preferably, non-Quadro (Please. Give us at least ONE goddamn discrete video option that's a top of the line non-Quadro RTX). RTX is non-negotiable. I want hardware raytracing.
* Two 2.5" drive bays AND at least one slot (preferably two) suitable for a SSD. Or at least one 2.5" drive bay capable of taking a 15mm drive, one slot suitable for a SSD, and a removable optical bay that can take at least a 7-9mm 2.5" drive.
* Thunderbolt 3. I might never use it, but dammit, this is the one loophole that partially opens the door to a future thirdparty x16 graphics card 3 years down the road when the built-in card is hopelessly obsolete for VR development (even though x16 is itself a fairly huge limitation) & gives me a way to limp around most of the OTHER things the manufacturer decided to insidiously cut corners on that weren't necessarily obvious or deal-killers at the time of purchase.
* 802.11ad with 60GHz wi-gig certification. I do VR development, and put the likelihood at ~90% that I'll encounter a need to stream low-latency video wirelessly to a headset at some point over the next few years.
* A keyboard and pointer stick that's at least as good as the one on a m4800... which honestly, still kind of sucks compared a real Thinkpad. Sigh. Fantasy shopping list aside, I'll probably end up buying whatever is Lenovo's best w-series mobile workstation when the time comes, because they're the only laptops left with keyboards that don't completely suck (and even they're a pale shadow of their former glory).
There's a practical limit to how much info you can put in front of you at once before it either becomes too dense to easily read, or so huge that you have to start straining your neck looking up to see it. Three monitors in an arc at approximately arm's length (with hand in a fist), with the top of the monitors approximately equal to your "straight ahead" gaze and the bottom of the monitors at your comfortable "tilted downward" gaze is pretty much IT. Bigger monitors allow you to move them farther away (handy if you have presbyopia), but having three 60" monitors won't really enable you to see MORE within easy gaze-range than having three 24-27" monitors will.
Three discrete displays are handy because they allow you to easily manage maximized windows and fling them from side to side, while viewing different sets of information side-by-side. It's also the practical limit of what you can interact with using a normal mouse and a normal amount of desktop space (more monitors and/or less space requires excessive mouse acceleration, which causes more problems than it solves).
Basically, the ideal (IMHO) is a ~120-150 degree arc of three virtual monitors at roughly arm's length away, located between "straight ahead" and "comfortably downward-gazing", with optical resolution comparable to three 27" monitors with at least 2560x1440 resolution... a big, wide virtual monitor in the center for the IDE, a second virtual monitor off to the left with an aspect ratio suitable for displaying side-by-side page views from a pdf ebook, and a third virtual monitor to the right that might be 3:4 aspect ratio (mine is 16:10, but 90% of the time I have the window on it only filling 2/3 the width).
IMHO, the most viable way to limp around the focal length problem is to treat it like simulated nearsighted presbyopia (inability to focus on things closer than 2-3' due to age, blurriness beyond ~10' due to nearsightedness) and use what are -- for all intents and purposes -- some variant of bifocal lenses in the headset.
Or... simulate the same combo, but with single-vision lenses and added astigmatism. As my eye doctor explained when I had PRK years ago, a small amount of astigmatism can actually come in handy later in life by compensating for the reduced focal depth. Basically, trading text that's crisp & sharp over a limited focal depth for blurry text that's readable over a wider range.
Moore's law hasn't ended... it's just been constrained by consumer expectations that new hardware will be cheaper and use less power.
Moore said only that the number of transistors doubles every 18 months. For the past 10 years, that doubling has been held back by consumer expectations that costs, power consumption, and emitted-heat will decrease. If Dell could sell millions of computers with the inflation-adjusted price of a 1991 Amiga 3000 (~$2,500 in 1991, correlating to approx. $7,000 today) that emitted 1,800 watts of heat (using active-refrigeration chilled-water cooling to prevent them from cooking themselves, along with some kind of outdoor condenser like central AC uses to pump all that heat outside), we'd be casually playing with supercomputer hardware that would make NOAA and NASA jealous TODAY.
Seriously. Think about how many watts a 3GHz Pentium 4 emitted with its single core... and multiply that by 16-32 cores to get an appropriate brute-force modern power budget.
No, it's not really feasible, if only because most people aren't going to install (and pay to operate) the equivalent of a 2-ton central AC unit to keep their computer from turning their home office into an Easy-Bake Oven, but it should be food for thought. Moore's law is far from dead. It's just that keeping up with it isn't quite as consequence-free as it used to be.
And if you need proof about what's possible when you don't care about cost, energy use, or heat-removal, read about professional bitcoin mining operations. That's TRUE hardcore brute-force computing power.
You mean, like this?
https://www.youtube.com/watch?...
There's a reason for the lack of AAA titles that goes beyond mere economics -- it's fucking IMPOSSIBLE to efficiently do VR development with current hardware and workflows. As the developer, you're CONSTANTLY putting the headset on... taking it off... putting it on... taking it off. And waiting... a lot. The integration of development tools into the headset environment itself is practically NONEXISTENT today. It's a definite, and very real handicap.
At least Rift has the advantage of using the same host PC to run the software that you're developing on, so the main limiting launch-to-view constraint is "how quickly can you put the headset on or remove it". With Android, VR development is downright excruciating... you can get piss-poor previews that are the equivalent of using the Android Emulator with the headset as a blurry remote display (with all timing completely shot to hell), or you can launch build+deploy and twiddle your thumbs for 30-70 seconds waiting for it to compile, upload, and launch.
What we REALLY need is a Rift-type display that can do the equivalent of overlay three virtual 27" monitors in an arc ~20" in front of you, with realtime camera-vision of the rest of the room, so you can develop without having to actually take the headset off (and sufficiently high resolution & optics so that you won't feel like someone who's legally-blind trying to sit at a desk and read those same three monitors in real life). Aside from improving game-development workflows, a capability like this would also give people who aren't even INTO gaming to invest in a headset, because it would let you have the equivalent of three large monitors in an arm's length arc around you in places where you CAN'T have three monitors (like on a plane, when traveling, etc).
What I'd really like to see: a headset that starts with the basic idea of an Oculus Rift + ZED Mini (two cameras that clip onto the front of a Rift to enable AR-type applications), but taken to the next level:
* Extremely high-resolution displays... 3840x2560 per eye... with freesync or G-sync framerates up to (and including) 100hz & 120hz.
* Equally high-resolution cameras, with two independent video paths:
1. A low-latency path that's basically analog and genlocked. This is the video of the room around the user that gets seen directly. Basically, anyplace where the GPU-generated video is "bluescreen green", this video gets overlaid.
2. A second, "digital" path that's available for the usual "AR" type purposes -- image-recognition, computer vision, environment-mapping, etc.
Path #1 allows the user to remain grounded in reality, and avoids the worst effects of "simulator sickness". Your brain can tolerate sloshing & swimming details, as long as the bulk of your surroundings are in fundamental agreement with what your other senses are telling it.
Path #2 allows for Augmented/Mixed-reality content, without compromising Path #1.
Ideally, the headset would support 802.11ad, so the relatively uncompressed digital bitstream from the cameras could be thrown directly at a server running on the local LAN to do the heavy-lifting analytical work.
Why this is important: current camera-video pipelines just plain have too much latency to use as an augmented-reality "environmental" view. I know, because I've written programs that try to take the video stream directly from the camera on an Android phone & render it straight to a Unity RenderTexture. They all lagged so badly, it was worse than having no view of the room at all. I concluded that the only solution that involves camera-to-video instead of holographic video overlaid on a directly-viewed scene (like MagicLeap or Hololens) would require cameras with an extraordinarily low-latency pipeline as described above, combined with high-resolution displays with high native framerates (ideally, 100 and 120fps... or better yet, 200fps and 240fps).
I'm highly-sensitive to simulator-sickness... but I'm pretty tolerant of traditional motion-sickness. In English, that means I have close to zero tolerance for latency, blur, and eyestrain with regard to my "environment", but can tolerate a fair amount of higher-order artifacts WITHIN that environment as long as the environment AS A WHOLE is consistent with what my other senses are telling me. The solution is to START with getting my environment to be acceptable when viewed indirectly (by looking through lenses at screens displaying realtime live video captured by cameras on the other side of the screen), and THEN overlay the synthetic video. Current VR solutions go at it "backwards".
Higher-density can help... but only to an extent.
If you're talking about single-family homes on large lots vs 16-28 foot townhomes of comparable size, yes... it might increase availability and reduce costs.
If you're talking about 50+ story skyscrapers, no... it won't. Skyscrapers are SO expensive to build (per square foot of habitable dwelling), the economics just aren't there to build anything less than luxury condos for the ultra-rich. If the developer is ALREADY spending $300/sf, adding $200/sf for premium finishes and doubling the selling price is a no-brainer... if only because someone who can afford $300/sf isn't going to spend that much for the equivalent of a doublewide trailer in the sky.
The catch is, places like the Bay Area are already "built out". The swampy land adjacent to the bay that's currently mud flats is only suitable FOR building skyscrapers, because only skyscrapers have the structure & foundation to rest on deep bedrock & survive earthquakes (Florida-style landfill won't work there... the first time there's a major earthquake, the landfill gets sifted, re-compacts itself, and whatever was sitting on it collapses). People who could AFFORD to buy a hypothetical skyscraper along the bay near somewhere like Mountain View don't WANT to live in a skyscraper adjacent to the bay near Mountain View, because it smells like rotting garbage & isn't particularly attractive... they want to live in a house on the side of a cliff overlooking the whole valley, or in SF itself.
And because the whole area is already built out, the only way to get land to build townhomes and apartment buildings is to buy existing structures, knock them down, and get permission to rebuild at higher density... which the remaining neighbors will fight tooth and nail. So what REALLY happens is, someone buys a block with 24 old single-family homes, knocks them down, and replaces them with half as many mega-mansions -- effectively, making things even WORSE. Or they'll buy an old apartment building with 50 units and poor residents, demolish it, and rebuild a new apartment building with 60 units that cost twice as much as the 50 units it replaced.
Basically, the Bay Area's densification ship has already sailed, and for the most part it's too late. At least, until the next major earthquake. If a city like Palo Alto were 90% flattened by a horrific earthquake & almost everything had to be rebuilt anyway, it MIGHT be able to get residents to approve a deal that would let them have their cake & eat it too by allowing them to sell half of their lot to someone else (or replace their original one-story single-family home with a three-story duplex sharing a wall & sell the other one)... but after a catastrophic disaster, local governments are rarely in any position to pull something like that out of a hat.
True story: 50 years ago, when US-1 was six-laned through southern Miami, the section near SW 152nd Street was split into two one-way roads because there wasn't enough room to widen it to 6 lanes along the original route. In 1992, Hurricane Andrew destroyed 99% of the structures in that area beyond repair. The area was ultimately rebuilt from scratch, and US-1 is STILL a pair of one-way roads, even though the original problem (lack of land) was solved by Andrew's destruction. Why? All the people who LATER bought property along the northbound lanes (shifted a block east from the original route) paid a premium for that land because it WAS adjacent to a major highway -- shifting the northbound lanes would have profoundly harmed the value of their land (and gotten the state & county sued). And FDOT still didn't own enough land adjacent to the southbound lanes to do it, even though the land was now stripped bare. Put another way, city planning decisions made long ago tend to affect future land use, even long after the original rationale has been forgotten or ceased to matter. It's the same reason why London's streets are a medieval rat's nest, even in areas that were bombed into oblivion by the Luftwaffe during WWII, long after the original structures that dictated their layout (or cemented it in place) ceased to exist.
5mph RUBBER bumpers. Painted bumpers are BULLSHIT... graze something, $500 minimum paint job to fix. I went out of my way to get black rubber bumpers on my truck, and it's the best thing I ever did. Small scuff? Fine-grit sandpaper, 2 minutes of rubbing, fixed.
The "small base" problem is partially mitigated by Unity and Unreal. They aren't quite "write once, run everywhere", but they DO make it possible to "write once, port quickly".
That's actually a harder problem to solve than you might expect.
Problem 1: what is the camera looking at? You can't just follow what the user is looking at... it would be like watching a bad camcorder video shot by a drunk person. To write the virtual camera positioning code, you need somebody who actually understands cinematography & can apply that knowledge to the user who's lurching around inside the VR world (keeping it pointed at the most interesting content, zooming as necessary, but nevertheless keeping it relatively stable to avoid making everyone watching it nauseous).
Problem 2: you need yet another GPU pipeline to render the "public" view... in a system whose video cards are ALREADY getting pushed to the breaking point, with no ability to add more. We desperately need the ability to pack four x16 video cards into a PC. Right now, the market for higher-end video cards is too niche for anyone to make halfway-affordable higher-end cards... but to achieve what you need with only a pair of cards, you need REALLY EXPENSIVE cards. If we had 64-bit PCIe, you could instead buy four cheaper cards (with greater economies of scale) instead. Or you could buy two cards, then buy two more after you realized the first two weren't adequate.
Fort Lauderdale doesn't HAVE to be pedestrian-friendly to appeal to Brightline riders who work in Miami. Many of them might end up paying a super-premium to live at Flagler Art Village within walking distance of the Brightline station, but for everything else they'll just drive. Brightline is their loophole that enables them to have a well-paying job in downtown Miami that demands long hours, without having to further endure actually LIVING there. Don't get me wrong... Miami is fun for a few years... but after a decade or two, the endless 24/7 gridlock and general dysfunction everywhere around you just really gets tiring and old. Especially when it sinks in that it's never going to get better during our lifetime, and will only keep getting worse. Even moreso when you realize that almost all of your friends have thrown in the towel and moved to Fort Lauderdale, and you're now driving up there all the time.
Downtown Miami is "pedestrian friendly" in the sense that it's relatively easy to use Metromover to get to 80-90% of the places where well-paid people are likely to work... and people who LIVE in downtown Miami don't EXPECT to be able to do things like shop there. It's just taken for granted that if you live downtown, going shopping means driving 10 miles to Dolphin Mall or International Mall, or taking Metrorail 8 miles to Dadeland Mall.
Put another way, you don't live in downtown Miami because you crave an urban lifestyle... you live in downtown Miami because you make lots of money, have no free time because you work 60-80 hours/week, and endure it because it means you can get to work in 10 minutes instead of 90-120 minutes each way.
Moving to downtown Miami IS known to be actively harmful to your social life. Why? Stupidly-expensive parking (downtown Miami has more parking per square foot than some outlet malls, but most of the spaces are owned by a cartel that uses artificial scarcity to drive up prices... it really, really sucks paying $20 to park on a Friday night in a half-empty garage surrounded by TOTALLY empty garages). If you move to downtown Miami, none of your friends will come over to see you anymore. Or they'll do it once, spend the evening complaining about how expensive it was to park, and never do it again.
Another perk in favor of Fort Lauderdale for the car-averse... if you want to take Uber to the beach, Fort Lauderdale beach is a relatively cheap 2-3 mile trip down Las Olas or Sunrise Blvd. Getting to the beach from downtown Miami without a car is significantly more expensive (Biscayne Bay is wide, the causeway meanders, and gridlock drives up the cost even higher).
In downtown Fort Lauderdale, the transit sucks... but most of the places you'd want to get to via Uber are within 2-3 miles. In Miami, the transit sucks... but everything you'd want to get to via Uber from downtown Miami is FAR AWAY and EXPENSIVE to get to.
There IS (sort of) a way... make it look like the larger environment is stationary & perfectly-synchronized with head motion (low latency, zero jiggle/slosh), and limit the moving stuff to small elements within the larger scene. Your brain can deal with small things that wobble/jiggle/slosh, as long as the "big picture" is consistent with what you see. It's when your entire WORLD is seemingly in conflict with your senses that you get motion sickness.
Eyestrain is a related, but different, problem that's currently a lot worse than it really *has* to be, simply because most devices have piss poor calibration & lenses. Part of the calibration problem can be resolved by giving each eye its own optimally-placed display (with precise orientation), and most of the remaining problem could be resolved by involving opticians to fit & customize lenses for individual users. It could be something simple, like designing a future Oculus device to ship with standard lenses, but be able to easily swap them out with custom lenses fitted by opticians certified by Oculus as being "VR-aware".
There's also the fact that "rendering optics" today are often where "game physics" were circa 1995. Libraries try to matrix-out distortion, without fully understanding what they're doing, what has already been done, and what's going to be done further down the pipeline. That's a big reason why VR companies have finally started to hire opticians & involving them in the process (and using them to help design better lens designs).
VR requires enormous amount of integration at every level. Comparing "First Person Shooter" to "Immersive VR universe" is kind of like comparing "Java app" to "distributed J2EE app with rich native client". The problem's scope is ENORMOUSLY bigger & requires more resources & more sophisticated management.
Right now, Brightline charges $20-40 each way for travel between Miami and West Palm Beach, and $15-30 for travel between Fort Lauderdale and either Miami or WPB. There aren't a lot of people traveling between WPB and Miami daily, but there are a LOT of people using Brightline to travel between Fort Lauderdale and the two cities. Pretty much every affluent professional who lives near downtown Fort Lauderdale & works in downtown Miami or WPB now takes Brightline, simply because it literally cuts the travel time in half (and allows you to spend it playing with your computer, instead of sitting frustrated in gridlock).
For a better idea of how successful it'll be, check back in another year or two... after it's been around long enough for people to start making "where do I want to live" decisions based upon Brightline's existence and travel times, as opposed to "where do I want to live, if living in Fort Lauderdale means I'm going to have to endure a grueling 60-90+ minute drive to work each way, and commuting daily between Miami and WPB would take ~1.5-2 hours on a GOOD day at 8am or 5pm.
Considering how INSANELY expensive rent is anywhere in the Bay Area, I think Caltrain could easily find tens of thousands of people willing to fork over $50-100/day if it meant they could pay $2,500/month for a 3 bedroom townhome with garage and back yard, instead of paying $4,000/month for a one-room efficiency. They might still pay the same total amount per month, but they'd get a lot more for their money.
Trains don't HAVE to be faster OR cheaper than cars or planes to be used... they just need to be perceived as nicer. Consider Eurostar... planes are almost always cheaper than Eurostar (especially walk-on Eurostar fares), but anybody who can afford it unhesitatingly takes Eurostar because it's almost as fast as flying (when you consider "terminal time"), and WAY nicer & less-stressful than flying. Ditto, for Acela between DC and New York. The proles fly from Dulles to Newark & spend 2-3 hours just getting to and from the plane itself. Acela passengers spend more time in motion, but half the total time getting to wherever they're ultimately going (and enjoy substantially more unstructured free time during their trip, vs air travel's "hurry up to actively wait".
There might be a HSR track from Hong Kong into mainland China via Shenzhen (which currently has a branch line to Zhuhai & extension to Macau under construction), but even if the train were allowed to run nonstop through mainland China & skip the border formalities besides those of Hong Kong or Macau, it would STILL be a 1.5-2 hour trip because it's so far out of the way compared to the new bridge/tunnel.
Plus, a train that ran directly between Hong Kong & Macau could probably relax some of the border formalities that are seen as necessary for trains passing through mainland China. Neither Macau nor Hong Kong might want to be swarmed by mainland residents, but I doubt whether they have much to care about with regard to EACH OTHER since they're both roughly equal in terms of wealth & freedom.
In American terms, rotate the map 90 degrees clockwise, and pretend that Macau is St. Petersburg (Florida) and Hong Kong is Bradenton. Yeah, in theory, you could drive from downtown Bradenton to downtown St. Petersburg by heading east to I-75, continuing north to Brandon, then slogging your way west through downtown Tampa over to St. Petersburg & driving south... and if you were extraordinarily lucky, you might even be able to make the trip in 90-120 minutes. But if you took I-275 and the Skyway Bridge instead, it's a 20-30 minute trip. Without the Skyway Bridge, Bradenton and St. Petersburg would be practically inaccessible to each other. With it, they're next-door neighbors. All over America, there are city pairs that USED to be mutually-inaccessible (at least, in any sane amount of travel time) that turned into casual driving distance once a bridge over some wide-but-shallow body of water got built. Ditto, for Europe... consider Copenhagen vs Malmo. The ferry always existed... but back when the trip took hours, the idea of living in Malmo and commuting to Copenhagen would have simply been LUDICROUS. Now, it's something people seriously consider as a daily option. The same thing is likely to happen when/if Italy finally gets around to building a bridge to Sicily.
Do you want to know the REAL difference between "boondoggle" and "valuable transportation infrastructure"? About 50-100 years (at least, insofar as COMPLETED infrastructure projects go... projects that get started & abandoned prior to being completed to useful length don't count).
In theory, the NEC was a "boondoggle". The Pennsylvania Railroad never made a profit on it due to the astronomical cost of maintaining it as electrified railway. Yet it's the reason why today, you can casually make day trips between New York and Washington without burning most of the day. Infrastructure has a funny way of doing that... hemorrhaging money at launch, losing money for years... then gradually, over time, becoming so important that disruption for even a few weeks would be DEVASTATING and UNTHINKABLE (hence, most of Amtrak's current problems with the NEC, and the London Underground's problems... they desperately need to do basic maintenance work on century-old tunnels, but can't shut them down, so they have two choices... defer maintenance as long as they can, or build a third tunnel at staggering expense so they can always keep at least two in service at any time because total shutdown is unthinkable).
Fifty years ago, Central Florida newspapers absolutely EXCORIATED FDOT for approving Interstate 4 -- a freeway from "nowhere" (Plant City) to "nowhere" (Lakeland). Nevermind the fact that even back then, it was fully intended to eventually span from downtown Tampa to Daytona Beach, and the initial segment between Lakeland and Plant City provided a detour around two crowded small towns that would have merited the construction of at least a 4-lane divided highway as a bypass around them for Tampa-Orlando traffic ANYWAY. If you read old newspaper articles, you'd think I-4 was proposed to be just a useless freeway between two small towns in Central Florida.
Ditto, for the Florida Turnpike between Orlando and Fort Pierce (the original "Turnpike" plans were for a "Y"... Miami to Fort Pierce, then one branch to Orlando (which is now Turnpike) and one branch to Jacksonville (which ended up as I-95). Newspapers and politicians fought bitterly over it, and pretended that then-existing roads between Miami and Orlando (US-27 through Sebring & Winter Haven) were even REMOTELY adequate (they weren't).
Back when "Alligator Alley" (I-75 across the Everglades) was first proposed by FDOT, the AAA fought it BITTERLY and did everything it could to get the project canceled for some insane, unfathomable reason. If anything, FDOT's only mistake with Alligator Alley was initially building it to a relatively low-standard narrow 2-lane deathtrap, instead of going full-on total freeway from day one(*).
Even Miami's Metrorail ended up being useful, despite Miami-Dade Transit's absolutely BREATHTAKING incompetence and short-sighted decisions (like cutting back Metrorail service to once/hour on weekends... at a point when their ridership was at the highest levels in its history... then scratching their heads and wondering why literally HALF of their weekend riders said 'fuck it' and disappeared. Unbelievable.) It took almost 40 years, but today the biggest limiting constraint on daily Metrorail ridership in Miami is the system's appalling lack of parking capacity at the 3 southernmost stations (thousands of people drive to the station intending to take Metrorail to go downtown, discover that the garage was full long before 8am, and end up having to drive downtown anyway because there's nowhere else to park).
Sometimes, politicians gamble and lose, like Cincinnati's subway. But you could equally argue that if Cincinnati had actually COMPLETED it and opened it for service, it probably WOULD have ended up influencing development and becoming useful 40-50 years down the line, too, even if it ended up being mostly useless for the first 25 years of its existence. That's just how transportation infrastructure IS... you build it through hinterlands, people spend years flipping land around it, then eventually people start buil
Considering China's enthusiasm for HSR, it does seem bizarre that they didn't just design another ~25 feet into the whole thing to make room for a pair of HSR tracks while they were at it. The marginal cost would have been fairly low, and it would have turned HK & Macau into the equivalent of a casual crosstown trip for each other.
A big chunk of Phase 1's cost is the tracks between Caltrain & BART @ San Jose and somewhere around Modesto, which will get dual-use... LA-SF, and also SF-Modesto commuter rail. It's needed regardless of LA-bound trains. LA-SF is just frosting on the cake... regional commuter rail is the big prize, because it'll open up the Central Valley as East Silicon Valley.
Likewise, Brightline intends to extend Las Vegas-Victorville to Palmdale to connect with CALHSR, so you'll also be able to take the train from SF to Las Vegas (probably BEFORE it's 100% HSR the last ~50 miles into L.A.).
Worst-case, 50 years from now, CALHSR repurposes the bridge structures for vacuum supersonic maglev. The important thing will be ROW-preservation. In 50 years, central valley land will be as expensive & urban as the bay area is today, and CALHSR will be viewed as a spectacularly fortunate real-estate investment, if nothing else.
And... how many accidents do you think Tesla's system PREVENTED? Let's be honest... most of the people who use Tesla autopilot wouldn't exactly be the world's most attentive drivers WITHOUT it, either. Even IF they were going through the motions and pretending to pay marginally more attention, we're talking about a group that's generally oblivious to anything that doesn't capture their immediate interest, tends to daydream a lot, and are almost the textbook case-study poster-children for highway hypnosis.
Let me repeat. At least with Tesla Autopilot, SOMETHING is paying attention to the road while they're behind the wheel, which is generally more than you could say about the "drivers" themselves in real life. Forget driver's-ed fantasy and propaganda... in real life, Tesla Autopilot is a life-saver (or at least, a major paint-saver and dent-preventer) for the majority of its daily users. It's not 100% perfect... but the drivers themselves are FAR worse. And that's why insurance companies grudgingly tolerate it... they know that at the end of the day, drivers likely to use Autopilot are going to have fewer accidents overall than they would have had if it weren't available.
An individual IC might not be "repairable" in any meaningful way, but there are also very few modes of failure where a literal single IC fails independently of nearly everything around it. Unlike, say, the screen on a phone, tablet, or laptop (which can break, suffer cable or interconnect failure, etc), batteries (which can, do, and are 100% EXPECTED to degrade over time), mechanical switches & fans, anything directly exposed to 100-240v AC or involving electrolytic capacitors, etc.
If a vendor wants to use a proprietary, purpose-built part... fine. As long as they publicly document enough of it to enable a comparably-sophisticated manufacturer to make compatible second-source replacement parts, with straightforward FRAND licensing terms. Maybe even require vendors to disclose itemized BOM costs per production run (registered with the department of Commerce, or its equivalent in China or wherever) & give statutory immunity from infringement lawsuits if a company manufactures & sells replacement parts when the original vendor either can't, won't, or charges more that 400% of its own BOM cost and/or unreasonably restricts availability.
A good start would be to look at laws passed in the 1960s & 70s to fight the auto industry's abuses... and the ways the auto industry initially tried to skirt them before largely settling down into the new regulatory regime. The electronics industry isn't identical to the auto industry... but it's not all that DIFFERENT from it, either. 98% of the issues consumers have now were issues faced by US car buyers in the past -- complex supply chains, proprietary parts, vendors seeking lock-in, restrictions on repair, and all. Regulations didn't solve 100% of the problems... but they DID fix the most egregious 90% pretty quickly, and the next 9% within a decade or so. And today, even automakers like Tesla (probably the most proprietary car in existence today) have made peace (or at least ambivalent detente) with third-party repair shops.
That would be silly... but what if you took the basic idea and applied it to laser-exposed microfilm? Sure, it might be slow to expose & probably require off-site developing... but once you had it done, you'd have a storage medium that could passively survive decades, maybe centuries.
The fundamental problem with any data-storage mechanism that requires active handling every few years to preserve it going forward is the fact that such data is unlikely to survive long enough to BE of interest to future historians. Businesses pay to retain records long enough to satisfy regulatory and immediate business needs, and that's IT. Governments are only slightly better at long-term storage for posterity. Just look at 1970s-era NASA and the BBC.
Old data that HAS survived to the present time almost ALWAYS survived ONLY because it was in a form that could be put somewhere, forgotten about, and rediscovered decades after it would have otherwise been systematically destroyed... and often, only because it WAS someplace that the people who would have wanted to see it get destroyed either couldn't get at it, or didn't realize it even existed. Like old letters in an old chest that went from attic to attic for a century or more. Or old tax records from Edwardian England or the Roman Empire. Or crates of punch cards from some long-gone business that were stored in an old office building in Detroit that was abandoned and left to rot decades ago.
Consider for a moment the fate of really, really old films from the early 20th century. Most of them were recorded on nitrate film that was literally dangerous. Few of them were considered valuable enough at the time to be worth transferring to "safety" film. The copies that exist today exist primarily because of a few film-buffs who were willing to disregard safety, grabbed them from trash heaps, and stored them in backyard sheds, until their grandkids discovered them decades later while cleaning out the house after their death & realized that they were a) dangerous hazmat that could literally go up in flames at any moment, and b) now worth a fortune, because all known surviving copies of them had been (intentionally or unintentionally) destroyed decades ago.
Museums and universities routinely have to deal with archived data that they don't have the resources to deal with RIGHT NOW, but nevertheless would LOVE to have some way to put "on ice" and just passively & compactly store somewhere indefinitely on the off-chance that some future historian might find it interesting. Or until AI gets to the point where historians can use computers to analyze, catalog, and categorize data that would take a hundred human lifetimes to meaningfully sort and categorize by conventional means. At the time the decision is made to archive it, the data might be very low-value, and won't BE sufficiently valuable to expend resources re-archiving for several human lifetimes. But if it's stored on a medium that can literally be packed into a mountain and forgotten about for a century or two, that's ok... it'll still be there for future historians when they eventually DO care about it.
I'd argue that a fair reform would be one that slightly raised the amount charged by USPS to deliver packages from a customs point... then made the same rate available to US companies willing to handle their own logistics up to that point. So if USPS charges $1.50 to deliver a package from their facility near JFK to someone in Miami, someone like Amazon or Walmart should be able to get the exact same rate by showing up at the USPS facility near JFK with 40 million packages per day.
Ultimately, it IS kind of fucked up that it costs less to mail a half-pound package from Wuhan to Miami than it costs to mail a fucking POSTCARD from Miami to Fort Lauderdale. A fair rate structure would be one that would be available to domestic mailers under identical terms, but wouldn't induce domestic shippers to do logistically-silly things to take advantage of it.
Right now, USPS rates are lopsided at BOTH ends... USPS *radically* increased the cost of mailing small, light padded envelopes ~2 years ago. Prior to that, you could mail almost anything physically-capable of fitting into an envelope that's within weight limits for the cost of first-class letter-rate postage. NOW, you get hit with multiple surcharges for "non-flat" and "more than 1/8" thick" packages that make it 2-4x more expensive. This new rate policy has utterly DESTROYED the ability of Americans to compete with Chinese sellers of things like small USB cables, microUSB-C adapters, etc, by driving the cheapest possible shipping cost for anything that's thicker than 1/8" and not consistently-flat up to $2-3/envelope. Merely raising the amount charged to Chinese shippers by 10-25c & using it to abolish the non-thin/flat surcharge charged to US shippers would go a long way towards restoring fairness & parity.
Is there any particular reason why nobody makes a product that's basically like non-LTH (phase-change/magneto-optical) BD-R, but on a flexible film substrate stored on reels instead of bulky discs?
The main problem I see with magnetic tape is that it's inherently susceptible to stray magnetic fields (including the Earth's poles). In contrast, phase-change media can theoretically have a passive lifespan that's measured in decades (centuries, if "being able to read it with normal, consumer-grade hardware as a normal OS filesystem" isn't a hard requirement, and you can deal with forensic data-recovery using exotic purpose-built hardware).
It won't eliminate celebrities... but it will reduce the number of living celebrities, and turn their kids almost into literal "royalty". Why? The film industry is weird... when an actor gets paid $50 million to make a movie, he or she isn't just providing $50 million worth of acting-in-front-of-the-camera value... the mere fact that they were PAID $50 million adds value to the film by getting others to take the film seriously. Kids from "Hollywood families" will go to the bargaining table able to offer not only their own services, but licenses to use the likenesses of their long-dead (or retired, if their parents want to help bootstrap their kids' acting careers) parents as well. Win-win for studio execs... they can brag about paying huge amounts of money to "an actor", but metaphorically get two (or three, or four) for the price of one.
Another possibility is that studios will still use live actors for much of the filming, but use post-production CGI to make their mouths match whatever language they've been dubbed into for a particular audience. This is a particularly big deal for high-framerate video. At 24fps, you can take a lot of liberties with lip-sync and mouthing of words. At 48, 50, and 60fps... it's more noticeable. At 100 & 120fps, any disconnect between what a native-speaker expects the lips and mouth to look like and what they see is going to be instantly and painfully obvious. With the right software and CGI, you could have a workflow like this:
* Shoot an episode of some TV show at 50-120fps in English, Mandarin, or whatever other primary language you choose.
* Record all the localized dialogue in different languages... quite possibly while recording the faces and mouths of the voice actor to capture additional nuance.
* Using the appropriate software, match up the origin actor's spoken dialogue with his or her mouth movements. Then, after analyzing the mouth movements of the voice actor for the target language, re-render the original actor's mouth to match.
Once again, lip-sync of dubbed dialogue is something that Hollywood and the TV industry have traditionally been able to somewhat dance around at 24/25/30fps (noticeable, but not necessarily a HORRIFIC problem), but becomes something you simply can't ignore at higher framerates and higher-resolution... and ABSOLUTELY can't ignore when you combine high-framerate AND high-definition. At 3840x2160@120fps, you'd better make sure not only that the spoken dialogue matches the mouth movements perfectly, but that the entire audio pipeline is PRECISELY synchronized with the video and room acoustics as well, or you'll find yet another novel way to fall into the "uncanny valley".
I'd argue the opposite point... the constraints imposed by present-day AR/MR systems generally match the limits imposed by age-induced presbyopia on someone who was myopic to begin with. If you can't comfortably focus on things that are closer than 15-20 inches without eyestrain (or at all), a 1-meter "near" focal plane isn't much of a limit anyway. Likewise, if you can't clearly see things that are more than 10 feet away without glasses due to residual myopia, a 3-meter "distant" focal plane is about as far away as you can see clearly without glasses anyway.
It's not middle-aged men and women who'll have a problem with the ML1's optics, it's KIDS who'll likely find it to be the most limiting. Which for now, is kind of ok, because middle-aged adults are just about the only people who can AFFORD to buy a ML1 or Hololens as a recreational toy.