I think it is because it is perceived as being money for nothing.

Neither are roads, bridges, railways, airports and sports arenas, yet many are funded by public money because a large portion of the public thinks that they are important.
Essential for survival, no. Essential for the community to thrive, yes.

99% of the Amiga floppies I still have work fine, but back in the day, I had to throw away 99% of my floppies quite early because of read/write errors.

Indeed US-centric ... but the poll was not about if you got it or not, but how much bandwidth you would use if you weren't limited. Apparently 1Gb/s is an astronomic figure to an American ...

I recently got upgraded from 100Mbit/s to 1Gb/s, and I did not even ask for it - Internet is included in the rent in the apartment block where I live and there was no price increase. The portion is around $15/month or so.
Indeed I was surprised recently when I heard how much US:ians pay for their low-speed Internet. Apparently, it is because networks were deregulated in 1995 in a way that allowed the largest providers to consolidate into an oligopoly, quashing the competitors by buying them instead of competing with better prices and/or services. The Tek Explains it.

I think that Samsung and Google are doing it all wrong.
They are still making smartwatches be "companion devices" to smartphones, yet you still have to write custom code to run on the device.

I think that the best type of smartwatch would be one that would act as a dumb terminal to the phone. Let it act as a second screen to the phone with a few button/touch actions plus a few sensors that feed data in the other direction. That would satisfy the most common use cases where a smartwatch would be useful. The others could be hard-coded not as apps but as system features.
This would be best for the developer, as you would only have to develop one app - not two.
This would be best for the user, as the program code on the "watch" could be simple you would need only a microcontroller that runs at tens of megahertz, and you get long battery life approaching what you are used to get in a watch.
But of course, such a device would be too cheap to make and Samsung would not be able to sell it at a premium...

I would say that the Kinesis is in a class above any Microsoft keyboard in terms of ergonomics.

Not only are the hands slanted, but there is more separation between the hands.
It has low-force mechanical Cherry MX Brown key switches that are relatively smooth, where as the Microsoft keyboard's keys bind horribly on off-centre key presses. (Cherry MX key switches are all the rage among PC gamers right now... ;) )
The lack of numeric keypad is actually ergonomically better in that it allows you to keep your mouse closer to your centre.
The keyboard layout can be fully remapped (without drivers, stored in the keyboard) and it can record macros.

Drawbacks is that 1) you must touch-type properly using all fingers and that 2) the keyboard is quite high because of its curved key wells.
#1 isn't really a drawback in the long term though, as learning proper touch typing will make you a better typist overall. It is easier to learn touch-typing on a Kinesis than on a flat keyboard.

I know people who collect vintage computing equipment that would kill you if they had seen you do that. (figuratively, of course)

Nah. The plastic nubs holding the barrel plate to the steel backplane tend to break one by one, and then the barrel plate cracks at the hinges. (it was moulded flat)
Enthusiasts often "bolt mod" their Model M keyboards with missing nubs: cut off the remaining nubs, drill up through the barrel plate and install proper nuts and bolts of steel instead. (not necessarily all in that order)

Indeed. I remember laptops made for Microsoft's "Window for Pen Computing" which was a special version of Windows 3.11 ... and this was back in 1992.
There were different types of swivelling screens like on laptops made today for Windows 8.1. Lenovo Yoga and Dell XPS are not novel in the slightest.

Large systems are still being done in APL.

The Swedish medical journal system TakeCare is one example. It handles practically all journals in the greater Stockholm area. It has sure had its slew of security problems, although I don't think that those could be attributed so much to the language as to sloppy sysadmins.

But hey... a few years ago I thought that nobody would use Erlang for anything significant, until it became a popular language for web services.

... and that option could be broken out into
- Those who download only free apps
- Those who don't see any need for downloading any apps at all

The only option would be "some other wavelet-based image format".

JPEG-2000 is completely different to ordinary JPEG. It is crippled in that the encoding is quite complex, has a tonne of different ways it can be encoded and is therefore difficult to do at speed. The software decoders that are not dead-slow are proprietary.
You wouldn't really win anything with using JPEG-2000.

I bet I could lift her computer. :)

Back in the '90s, I carried my computer and 14" CRT monitor to and from LAN parties many times.

What do you mean "not handheld"? I carried them in my hands on the subway.

I was surprised when I was a kid back 25 years ago, that my dad could skim through text very fast.
He worked as a journalist, and as such he was used to skimming through a lot of text to find the good bits that he could use as leads and sources for his articles.

The difference to the Internet today, is just that more people are exposed to larger amounts of many different types of text, just like "text-workers" like my dad was back then.

I think the general rule should be that the patient should decide about his/her own health.
I have met and heard about people that have had various conditions that have opted out on diagnosis, because they want to opt out on a certain treatment. People who have had cancer multiple times and would rather die from cancer the next time than suffer through radiation therapy and chemo, or people who have had an implanted automatic defibrillator that has provided a very painful experience.

If any kind of medical test is done, be it genetic or otherwise, then the test results should automatically only be available to the doctor who had requested the test. Permission to /portions of/ the test results should be available to other doctors only if the patient gives explicit permission.