But more generally I'd ask the question of, if you can train a conscious animal to execute 8 brainfuck instructions (+, -, <, > , etc..), could they also develop a writing system and express higher order languge concepts? Asking why animals don't write assumes they can't, as maybe we just need some kind of intermediate symbolic phoneme concept, as though - like BF for computation, there are some basic unit instuctions that you can combine and parse into anything, and if a border collie can distinguish words for hundreds of different things, I'd wonder if what we're missing is a set of metaphors that represent this underlying core instruction set.
Things like musical notation and knitting stitch patterns are also thought to be Turing complete, so it implies the representation doesn't matter for computation, and computation is sufficient for representing most things, so the idea of finding some underlying instruction set for producing languages across species seems both fantastic yet loosely plausible.
Of course as with the guy in the Chinese Room that doesn't speak Chinese, none of the individual animals would have any idea what the overall system was doing.
Strike "animals" out, and it sounds like any worker in a big enough org. (Maybe I should stop reading Kafka)
That just means you need a redundant array of parrots operating in parallel, with a regulator to detect different answers to the same problem and retry.
Anyway, aside from fantasy/sci fi stories, the interesting analogy with the story is just because I'm not smart enough to debug the system given no blueprints or explanation, does not mean an anthill is not running some novel and possibly interesting computational problem. Or bees nest or giant fungal mycelia in the dirt or herds of lobsters.
There are randomness detection and evaluation algos (distantly related to compression algos) and now that we have big data I suspect the next decades will have those algos rubbed up against big data and we'll discover interesting things about hive insects and what the hive is thinking as opposed to what individual insects are thinking. Or run it on dolphins or ancient human dwelling architecture.
https://en.m.wikipedia.org/wiki/The_Laundry_Files
He also has another series which he just finished that is very good as well about a family with the ability to step between diffrent timelines that's an interesting deconstruction of the trope.
https://en.m.wikipedia.org/wiki/The_Merchant_Princes
********Spoilers*******
.
.
.
While it starts as starts as a fantasy series, it eventually comes out their magic abilities are remnants of advanced nanotech from an interdimensional empire. Apparently a publisher he didn't like working with had the rights to his next Scifi series, so he had to start it as fantasy to get around the contract.
https://www.tor.com/2008/07/20/down-on-the-farm/
And it was an "IBM 1602" although pretty obviously based very closely on the actual 1620.
The real world 1620 was a 50s era RTL scientific computer (IBM used to separate math processors from business processors, until the famous "360" that did everything pretty well). This used the SMS style hardware cards that everything IBM made used until the 360-era SLT cards. Each SMS card holds enough circuits to be roughly one 7400 series TTL chip.
And another: https://www.youtube.com/channel/UCEa46rlHqEP6ClWitFd2QOQ
I'm glad you figured that 64 cat-cores is basically pure chaos. Well, for you, for the cores it's just their way of life: do random stuff for no apparent reason. Which is like the exact opposite of a computer.
The laser system has promise, but you need a source of random data to drive the movement of the laser pointer (maybe strap them to other cats?). Otherwise, an attacker who can predict the movement of your pointers may be able to predict the output of your entropy stream after observing the behavior of your cat colony (certain cats may prefer to hunt at different times of the day, or prefer to lay in wait and pounce vs giving chase, etc).
Alternatively, you could observe a field (bed) of cats and record their chirality, assigning 0 to sinistral cats and 1 to dextral cats. The science[1] seems to indicate no preference for one or the other per cat, so your readings should be random. However, you may only observe one or two bit-flips per hour and during transitions it may take some time for the system to settle back down to a steady-state.
[1] http://blog.dimview.org/math/2017/07/28/cat-chirality.html
Good point. Would strapping a laser pointer to the back of one of the cats work to seed the system? After that I could divert some output into g-code movement instructions to the automated lasers.
One thing I'm unsure of is room size. I figure I could map 1/0 to each tile in a grid based on the presence of a cat, but that means with too large or small of a room I'll have varying degrees of bit density. Stakes are pretty high too, since I'm pretty sure that putting cats in the mix might allow me to exceed the theoretical maximum of 1 bit of entropy per bit.
Equally important, you must respond “No” when asked whether you want the “Neutering” feature enabled. This applies to both cats.
Cat ladies are evil nihilists.