Bold move in the current climate.

Where? Never heard of it.

Exabytes (millions of terabytes) must take multiple days if not weeks to copy securely even across 16 Gbps fibre channel links? Petabyte scale storage arrays (holding thousands of terabytes) are perhaps more commonplace nowadays but will still take many hours to copy across even a decent fibre link (i.e. 100 Gbps = 12.5 GB/sec) as most older source arrays are an order of magnitude slower than that anyhow, closer to 1.25 GB/sec on average, or roughly 108 TB per day (which'd take DECADES to transfer an exabyte at that piddly rate!)

Probably both. Some services actually come with preconfigured Android boxes locked down with their own apps based on custom skinned versions of the legit Kodi software. I was shocked to discover a less tech savvy family member in Europe had signed up for 120 Euros every 6 months for a box that came with a huge static movie and TV seasons streaming library plus every single live satellite television channel, including movies, sports, even the pay-per-view and most in 720p. Apparently the guy who sold the service just swapped the Android box out when the service was renewed.

Bad for who? If you're a government that can afford to buy (or otherwise facilitate mass production and distribution of) whatever drugs you urgently need from whatever supplier is preferred (for various reasons), then an existing patent-free product is perhaps more desirable to consider before breaking the established rules of global big pharma? This may only apply in some countries though.

Don't forget about injecting detergent. Surely a combination of all three will be very, very effective. Such a stable genius approach. Right I'll get my coat, I'm off out for covfefe.

You sir, have won the internet today. Have a cookie ðY

In terms of hard limits perhaps 256 characters is a more sensible choice for lines without a join or break character, so that any legacy 8-bit code (popular for old game ROMs, etc.) referenced by other code (i.e. for static analysis or similar) can use byte value references for code column/row position? Source code is just data, size limits are handy rules for avoiding buffer over-runs, use after free bugs, etc. when using external tools for parsing that data but only if the rules are explicit. Better/safer programming languages could have some protections in place to mitigate both unintentional parsing errors and intentional compiler/interpreter exploits?

I totally agree, fuzzing is a great experimental test approach but only if its really appropriate for the system under test. One challenge with system context, is learning to build a fuzz test framework to exploit whatever software APIs or other data device/network/bus protocols are accessible in new and interesting ways by exploring and exploiting various data inputs so that they hopefully return unexpected output results, usually where the input space is large or dynamic enough to potentially miss something like a weird control character or escape sequence that causes most sanitising code to fall over or otherwise do something else, other than just reject the fuzzy test input data as a bad input.

Also look for the original Apollo 11 Guidance Computer (AGC) source code for the command and lunar modules.

I vaguely remember it on my Commodore Plus/4 which was a slightly less beige and more angular 16 but not as good as the 64. Set me on the path to the Amiga though and for that I'll forever have happy childhood memories.

Finding all the bugs. Ha!

Nice, STONITH for timepieces. Most won't understand how much your quorum and fencing analogy makes complete sense. Do your highly accurate time servers also link to atomic clocks and occasionally argue over leap seconds when NTP packets drop, causing them to remove any errant nodes from the pool that became out of sync?

Rather than artificial intelligence, perhaps start with basic logic for artificial life, then expand the complexity for those models that succeed to a point where interesting patterns emerge. Conway's game has very basic rules that can very quickly become extremely complex and interesting but why has nobody revisited that work with more modern systems to simulate a more realistic, small (but genuinely intelligent) set of lifeforms say perhaps just converge a handful of multicell scale neural ecosystems, see how they play out given freedom of code construction and data I/O choices and a reward mechanism (perhaps additional processing or storage capacity) for generating (or individually regenerating) any useful code.

The UK originally had a similar 'herd immunity' outlook, until a detailed study from Wuhan and Lombardy showed how serious the situation really will be without intervention. Millions dead globally, hundreds of thousands nationally. Mortality rate in elderly and vulnerable is extreme, perhaps over ten percent, combined with a high infection rate is definitely not something to be ignored.