Starlink uses highly directional beams from phased array antennas. The footprint from each beam is very small; there's not going to be meaningful bleed. They can target what areas do (and do not) get service quite precisely.

In fact, one of the reasons they need so many satellites is that each one can only put out so many beams at once, and each beam can only serve a tiny area, so they need a lot of satellites just to get full coverage of, say, any given large city. These are highly directional beams.

$2300 is way overpriced for VR. High-quality AR - where it projects onto a clear visor - is much more expensive, but good VR systems can be had for $300-$500, depending on sales and the exact setup you want. Even the bleeding-edge, top-of-the-line nothing-takes-full-advantage-of-this-hardware-yet commercial VR is usually under $1000.

If cost was no object at all? Starliner, honestly; Boeing has an awe-inspiring capacity to burn money but their rockets are top-notch safety-wise (perfect record on the Atlas V) and I'm hopeful that their capsules take more from their rocketry than their airliners in this regard (while the failed parachute deployment was concerning, parachutes are hard; SpaceX has had plenty of trouble with them too).

If I had to pay even 0.1% of the total cost myself, Crew Dragon. Heck, if you raise the cost to a full 1%, I can't afford Starliner!

Symmetric gigabit is slow? You think Ku-band radio is blocked by clouds? Damn, you don't even know what a phased array antenna looks like (it's a flat box, not a disk). What a moron.

There absolutely are licensing requirements for high-bandwidth purposes like communication satellites. It's not a problem, though, because they use directional antennas that aren't just indiscriminately spraying the planet, and they can (and must) steer those antennas as they pass overhead (or, if they're in GEO, they aim at a particular region).

There are some bands which are, by international agreement, unlicensed - WiFi uses these, for example - but even those aren't fully unregulated (there are still prohibitions on things like deliberate interference).

I have a friend who works on Starlink; according to him, getting licenses to operate in the various countries is one of the biggest impediments to creating a global network. It's a big part of why the service will initially roll out in the US only, before expanding to the rest of the world.

You're aware that LEO isn't a single altitude, right? That not all satellites are in the same "shell"? There's a lot more room in space once you remember that it's three-dimensional.

Also, rockets don't go straight up; most of them will be entering a parking orbit well below the Starlink satellites (just like they do today) and then they (or their spacecraft payloads) will raise their orbit along a path similar to the path that the satellites are taking. It's like complaining that freeways are too dangerous to allow because you'll have cars whizzing by every five seconds, without considering that you'll be going the same speed and direction as them.

Finally, "instantaneous" one-second launch windows are already quite commonplace. Every single space station launch is an instantaneous window, for example. If you don't make it, you try again another day when it passes overhead again.

24 launches have been stated as the minimum for global coverage. The rest are just to ensure there's enough bandwidth for high-traffic areas. This system isn't (just) intended to compete with geostationary satellite connectivity for people in the middle of nowhere; it's also meant to be an alternative to your local ISP.

Starlink uses highly-directional beams to connect a user to the satellite. Both the user terminal and the satellite have to know where the other is to enable high-speed connectivity. For typical high-bandwidth communication satellites in geostationary orbit, the "footprint" per beam is relatively large, but those satellites have terrible latency (the distance to GEO is roughly the circumference of the planet, so your round-trip time, even at light speed, is quite significant). Starlink uses a low earth orbit, only a few hundred miles (or KM) up. This means that the footprint of each beam is tiny, so it needs to know fairly precisely where the users are.

Why does this matter? Well, if China wants to say "no using Starlink in this country until they let us censor everything" and Starlink doesn't enable that censorship, then they won't be licensed to operate in China. That doesn't physically stop them from aiming beams at China, of course, but that would be detectable (and in violation of the bandwidth allocation by the Chinese equivalent of the FCC). Because the beam footprints are so tiny, you can't really say "Oh sorry, we have a customer in Nepal, we didn't mean to let people in Tibet access the system"; that excuse just wouldn't fly. If Starlink doesn't have approval from China, they probably just won't aim any beams at it.

Also, China could just arrest anybody with a Starlink user terminal. Phased array antennas are not small or particularly subtle; they're roughly the size of a pizza box. Even if SpaceX were willing to allow service in China without government approval, and deal with the resultant incident (which would probably get them in hot water with the US government, for violating international agreements about national sovereignty and spectrum use and so on), the user on the ground still needs hardware.

Or, to put it differently: Starlink doesn't enable connectivity anywhere that Iridium doesn't already provide, it's just much faster and (hopefully) cheaper. Iridium user terminals are small enough to hold in one hand, though, because they don't use phased arrays. One way or another, the Chinese government (and everybody else) can endure this problem; they already are, and have been for decades.

Technically, yes, although they try to hide it on the e-ink models. It's a pretty limited thing, much more akin to a portable media player than to a smartphone, but it's there. They don't release an SDK for it anymore, though (there was one, and some really old e-ink Kindles still let you load third-party apps/games, but that was dropped years ago).

Kindle Fires (the ones with color LCD or LED screens) are just Android tablets running a customized version of AOSP that Amazon calls "Fire OS". The vast majority of the AOSP functionality is there, and you can install APKs (from a store or sideloaded) to your heart's content.

8.0 had major flaws in its general design (Why can't "Metro" apps be windowed? Whose idea was it to segregate the Start search results?!?) and it was obviously a stupid UI for a desktop or even a laptop. The interface worked pretty well on a tablet, though. I mean, you wouldn't want an interface that works like an Android (or iOS) device on a desktop or laptop, but that doesn't mean they have bad interfaces in their intended niches.

More to the point, the Win8 "desktop" interface and the Windows Phone 8.x / Windows 10 Mobile interface are pretty different. Both use "tiles", sure, but beyond that there's quite a lot of difference. No "Charms" bar like Win8, stuff was generally arranged for one-handed operation, and in general it's actually very similar to Android. There are differences that mean some learning curve going between the two OSes, but there's almost no differences in UI capabilities (or at least, there weren't back when W10M was still getting regular feature updates).

Also, the dev tools were nice. It was easy to make apps re-flow nicely depending on screen size and orientation, etc. (in fact, they did by default; it took some effort to make them not). Sideloading is actually easier on W10M than it's ever been on Android. The capabilities available to third-party devs were unfortunately limited on WP8.x, but were expanded quite a bit with W10M (although many require special permission from MSFT for you to use them in an app published on the store, which was - and probably still is, for Win10's store - obnoxious to obtain).

The browser works fine for most things. The main thing I miss are the special-purpose integrations that used to work and don't anymore. It's going to be a problem for me when WhatsApp stops working, though; my parents moved over to it almost entirely a few years ago.

I'm not aware of any RTS bots in the 90s or even 00s that could play with anything like human player limitations (no resource advantage, no "maphack" information advantage) and routinely beat 99th percentile human players. Limiting the AI to under 300 APM and forcing it to scout for information like any other player makes for a much trickier challenge.

DotA and similar games are certainly impressive achievements for an AI, but they're much less strategic than Starcraft. Much more of the essential individual skill in a MOBA comes down to response time and situational awareness, tactical aspects that software excels at. The team coordination aspect is probably harder for an AI, assuming they don't have one program running the whole team or at least using faster-than-humanly-possible communication with the other "players", but it's not *that* hard to get the basic heuristics right for that.

The main strategic aspects of MOBAs are in hero selection and a relatively small decision tree (push / defend / farm / roam / retreat / bossfight, with a handful of location options for each) that needs to be evaluated frequently but is usually easy to decide. Item builds definitely also matter strategically, but there's a lot less flex there; there aren't that many items in the game, even fewer are both relevant for any given hero and any given set of already-owned items at a particular point in the game, and variations from a default item build are mostly reactive (responding to what the opponent builds, and what your allies have been able to build). The rest is tactics.

Again, I'm not saying this isn't an impressive achievement for an AI. It absolutely is, especially if it had to figure out those item builds and team compositions and strategic decisions without experienced player input to start from. However, the strategic depth is much lower than Starcraft, where you not only have far more units than in any MOBA spread out over a larger portion of the map (any spot of which could come under attack suddenly), but also have much less knowledge of the map and of what the enemy is building/researching so countering is a matter of scouting and guesswork much more so than in a MOBA.

why don't you (the US, not you personally) move the date to a weekend, so most people won't have to skip work for the time it takes to vote? That's how it's done over here in the first world.

Because the election day is set by historical precedents that date well back to before universal suffrage, when if you had to "skip work" you probably weren't eligible anyhow. It hasn't been changed because enough of our politicians like it that way (both the "tradition" aspect and the fact that it depresses turnout among people for whom a few hours out of a workday is a hardship). The USA adopted popular elections early on, from a modern history standpoint, but we didn't get the implementation quite right. A handful of US states and local governments do make election day a civic holiday, or at least mandate that employees be allowed time to vote without retaliation or loss of compensation, but federal legislation to make election day a non-working day has never made it through both chambers of congress.

People who have to work anyway, or know in advance they're travelling, can apply for postal vote.

Many (most?) US states allow vote-by-mail for some situations. A few states now use postal voting for every ballot, with "polling places" essentially nonexistent. Early voting (typically by mail) is allowed in most states, at least under certain circumstances (such as travel) as well.

See https://en.wikipedia.org/wiki/...

Sad. There was none of that in SA 20 years ago. Not saying SA (or the world in general) wasn't a pretty different place 20 years ago, but back then, you paid the price listed, and there was never even a suggestion you should tip.

I suspect American tourists who don't know any better have been expecting they should pay more than the stated price and introduced this stupidity worldwide, and now service staff try to milk tourists for all they can get. I wonder if they expect locals to tip, yet? Or if their wages have been drastically cut because "you'll make up the difference in tips" yet, the way it's done in the USA.