Tauchain in the Algoverse

karov (56)in #tauchain • 7 years ago (edited)

Lets build an universe [1], [2]. I realize this blog post is the most 'psychedelic' up to now and for long time to go, but some 'poetry' never really hurts ...

We discussed already the worldmaker effectoring [3].

It is quite ancient but also exponentially growing business ... in all possible forms of science [4], faith [5] and art [6]. This modeling [7] usually serves to play out what's possible and what's impossible. Gedanken eksperiment [8], yeah, but isn't all thought [9] merely algorithmic [10] and mere action [11] ?

Usually the posited universes are made of variations and combinations of substance/matter, structure/form and action/process rules. Though, the algorithmic component is always the essential ingredient. Yes, the Laws of Physics are full-fledged, literal algo [12], too. I have those conjectures that it is impossible to think out, make or discover (which is one and a same thing) a lifeless universe [13] and that substance-structure-action are inexctricable, but these are separate topic for some other times to address [14].

Lets put together ours toy-universe [15] out of only pure algorithm. I've never seen such a construct, although the Orbis Tertius [14] is enormous and I bet this vision have occurred gazillions of time in zillions of minds.

It is like an ocean. The primary coin-toss algo which outputs 0s and 1s [16] makes the water. We don't know (yet) if there are even deeper and more fundamental numerical bases [17] for running algos. Most probably the answer is yes by analogy with the Dirac Sea [18] - the deeps to be made of simpler and weaker algos. The most elementary coin-toss thing makes out the ... probabilistics, perhaps the primordial form of logic. The laws of physics (and of machine learning [19] and of darwinian evo algo [20] ...) tell the rule-set how to stitch together lotsa coin-toss outputs. A hint on inspiration for that - David Deutsch's Constructor theories [21]. The laws of physics as entropy [22] limitation of the allowed elementary algo cumulative output. For information being a verb, not a noun - isn't it? Very interested philosophic perspective on algorithm as randomness constrictor [23] raise up...

So, if the Algoverse ocean water is made of elementary coin-toss molecules, being ''liquid'' is just another phase or aggregate state [24].

There is deep duality [25] between probabilistics and logic. Just like the zoo of dualities discovered in accelerating pace by the mathematical physics in the last decades [26] Probability/statistics we make now by logic [27], the reverse ... - well, nobody yet cracked it. Even Kolmogorov. But I bet we will. Most probably the breakthrough will be Ohad Asor name-labeled... To find the know-how to do it the other way round, do logic with probability/statistics. The statistical algorithmic - not the SAT [28], brute force, alchemist [29] way as with NN/ML [19], [30] and other known beasts. This will be nothing less but full merger of maths/logic/philosophy/thought... and physics. Literally!

Excuse me for the haiku [31] simplification. It is deliberate due to realization of my grok constraints. :) Regard it as sharing a poetic impression.

Is there deeper and weaker algo than the digital - the radix-2, deterministic, unitary one? Intuition says ''yes, of course!'' Like with these radix-1 Half-coins [32] of negative and other non-unitary probabilities ... which take two tosses to yield a bit... and there must be transfinity [33] of lower ones, also transfinities of higher and sideways ones ... which is almost as counter-intuitive as Dirac's bottomless night of negative energy [18], but I bet also as much useful. (Lets not even touch numeral bases of Pi, i, e ... etc.), and lets stick to strictly binary 'water' for our oceanic toy-universe for the sake of sanity.

The next important notion of the Algoverse ocenic model is the Algorightmic strength [34] - the weakest algo would be that which takes infinity of tosses to get a full bit. The strongest?

Algorithmic ephimeralization [35] - essentially to do more with less. Or faster - Speed Prior [36] ... which is just another way to say 'more'.

Some algos are too strong - QM, M-Theory - they return way too much bits per 'toss'. Their vcdim [37] converges to infinity. Exponential walls [38] in all directions. Not exactly what Freeman Dyson had in mind [39]... In our ''mockup'' they could be depicted as too hot. Changing the phase of the elementary algo 'water'. Like.:

but because we are all for peaceful use of algorithmic energy - we reject those up here, too - together with the non-unitary statistics down there.

Last piece of the picture - the Algoverse ocean is habitable and inhabited!

By higher algos as life-forms, stronger - but not so strong to turn the 'water' into roaring steam or plasma.

Examples: Calculi [40], geometries, algebras ... software [41]. The genetic inter-algo connection should be that calculus came from Leibniz and Newton and numerous unknown others ... heads, but it is the blind watchmaker [42] of evolution which put those heads together ... (I disagree with Dawkins only on that evolution and design are both algorithms, alternatives but not opposition).

Thus entropically [43] and combinatorially [44] algos kinda-sorta come from one another - the stronger from the weaker.

The stronger are the life-forms living in that ocean. Cause randomness [45] permeates everything, isn't it?