Daylen's Journey

Where are you and where are you going?
jacob
Site Admin
Posts: 15906
Joined: Fri Jun 28, 2013 8:38 pm
Location: USA, Zone 5b, Koppen Dfa, Elev. 620ft, Walkscore 77
Contact:

Re: Daylen's Journey

Post by jacob »

daylen wrote:
Fri May 19, 2023 10:29 am
As for energy, rings are supported by swarms of fusion-drive spaceships.
7Wannabe5 wrote:
Sun May 21, 2023 9:08 am
Are you applying the principle of indifference to low energy future vs. high energy future in your model?
That doesn't seem to be the case.

On a semi-related note, I found the second half of Seveneves interesting because in the novel humanity never recovers from the integrated circuit peak of present times. While future humanity is space-faring in the year 3000, their computing ability is 1990ish. It IS interesting how technology went the way of exploring the nano-scale about 1-2 generations as opposed to continuing to push the giga-scale. In The Three Body Problem, a single proton-sized computer is sent to infiltrate Earth from a nearby star system. This kind of space travel is cheap as accelerating a proton to near light-speed is doable with present technology. Accelerating an entire human, not so much.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

From my point of view, there seems to be evidence to suggest that some of the middle is excluded to some extent and thus should be discounted in my baysian updating of future energy levels for say the next hundred years. Applying the principle of indifference to the possibilities or priors without evidence would give you a uniform or maximum entropy distribution across a range of energy levels in a hundred years. Physical evidence of atomic energy at the small scale and heat dissipation limits at the large scale bound that range. Finite oil and other resources on the planet necessary for the production of circuits and whatnot may bifurcate future energy paths due to inability to jump from wood and coal to say fusion or distributed wind and solar when lacking oil and other necessary resources.

This seems to be my current line of reasoning but more evidence to reverse that excluded middle would be interesting.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

Oh, and yeah I read The Logic of Science a couple years ago and it significantly updated my thinking process. Another application would be investing where indexing tends to approach a maximum entropy distribution of the securities. Market and player information can update that distribution to become lumpier.

I've tried to read The Three Body Problem a few times but just can't seem to get into fiction unless cinematic. I am probably missing out on all kinds of ideas, ha. Though, I do watch Isaac Arthur on YouTube and he mentions sci-fi stuff a lot.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

So, in some sense I am not applying the principle of indifference since I have no direct evidence I can point to that would justify excluding the middle. Though, this would be like limiting epistemic knowing to the computational paradigm where you have to run the universe to get the universe. On the other hand, the mystical paradigm or active inference at scale can in some since take shortcuts that over many simulations start to skew priors based on intuition or feeling. Similar to how if you have eliminated most of the possibilities, whatever remains, however improbable, must be the truth.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

jacob wrote:
Sun May 21, 2023 9:18 am
This kind of space travel is cheap as accelerating a proton to near light-speed is doable with present technology. Accelerating an entire human, not so much.
Right, so I imagine that it is still quite possible to crawlonize the solar system (first with silicon then potentially with carbon-based entities) to become a K2 civilization. Then, the transition to K3 would require a very slow crawlonization but still possible in proportion to how large the agents are and how susceptible they are to pressure differences.
7Wannabe5 wrote:
Sun May 21, 2023 9:08 am
Could there also be a moderate energy future given steady, rapid, but not too rapid, population decline?
With Volution, I am just considering the K2 case for now. Another possibility, of course, is that we manage to achieve something like stagnation on Earth for the next few hundred years or longer. This could involve the gradual ease off oil reliance and may actually be a preferred scenario (where the high-energy phase molds into a low-energy phase over hundreds or thousands of years).

One question I ponder a lot is whether allocating resources to a sol economy(*) can actually help us utilize planetary space better? The major factor of course being the energy required to transfer resources across steep gravitational wells. A space elevator would drastically help with this and similarly relies on nanotech like carbon nanotubes.

(*) That is, a presence in the solar system that may be entirely automated to beam down solar energy, provide communication networks, store resources at the edge of Earth's gravity well, and so forth.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

This is good, I am updating. :lol:

Okay, so the best case I can make for an eased-energy future is that AI becomes general enough to help us achieve the planetary coordination(*) necessary to avoid races to the bottom of the oil pit and the systemic shock that follows it diving us into a low-energy future.

(*) But then they may also be suited to crawlonize space, perhaps somewhat independently from organic life.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

A middle future where humans do not have very advanced computational capacities and still manage to ease off oil over hundreds of years seems somewhat improbable to me. Though, maybe it is possible given we approach something like collective enlightenment.

jacob
Site Admin
Posts: 15906
Joined: Fri Jun 28, 2013 8:38 pm
Location: USA, Zone 5b, Koppen Dfa, Elev. 620ft, Walkscore 77
Contact:

Re: Daylen's Journey

Post by jacob »

daylen wrote:
Sun May 21, 2023 11:02 am
Though, maybe it is possible given we approach something like collective enlightenment.
Fat chance. Somewhere in between filling TPS reports and pushing buttons is more like it...

7Wannabe5
Posts: 9369
Joined: Fri Oct 18, 2013 9:03 am

Re: Daylen's Journey

Post by 7Wannabe5 »

daylen wrote:From my point of view, there seems to be evidence to suggest that some of the middle is excluded to some extent and thus should be discounted in my baysian updating of future energy levels for say the next hundred years. Applying the principle of indifference to the possibilities or priors without evidence would give you a uniform or maximum entropy distribution across a range of energy levels in a hundred years.
Right. I agree that the bi-furcation between high energy future and low energy future is most probable looking out to 2100. What I really meant to ask was what probability are you currently assigning to the high energy path vs the low energy path, if not 50/50 (indifferent)?
Another possibility, of course, is that we manage to achieve something like stagnation on Earth for the next few hundred years or longer.
For better or worse, it seems like something like stagnation (at least localized) seems fairly likely until my estimated check-out date circa 2050-60, though I do believe that my as-of-yet-theoretical* grandchildren will be cursed/blessed with more interesting times. It's also the case that although I believe that low-energy future is most likely outcome at 2100, it might be in best alignment with self-interest to bet with those who are betting on high-energy future for another decade or two.

*this bifurcation of the future is also increasing the possiblity in the present that my grandchildren will remain theoretical, because even for individuals who are operating on unconscious level (as opposed to my adult children who are conscious), it is more "expensive" to prepare for both possibilities.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

@7w5

Around the end of HS when I had formed a more stable conceptualization of something like a future with technology and a future without, I was biased towards optimistically assigning the high-energy future a large weight. Over the college years I gained more evidence for a sharp decline into a low-energy future and eventually assigned more pessimistic weights. More recently, that pessimism has reversed a little bit. I'll assign [currently as an instantaneous prediction] a weight of 85/15 in favor of a low-energy future projected out to 2100.

Based upon @jacob's comment perhaps there is a middle case not being properly integrated into this estimate involving a future extending beyond 2100 that still leaks oil-based energy into the economy without a race to the bottom and sharp population decline into a low-energy future? I may be misinterpreting his comment, though.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

I guess it gets kinda hazy if you consider that the economy would become decoupled at various levels and what seems like an eased-energy future from one vantage point may look like an immediate low-energy transition from another. So, perhaps some players will hoard the remaining reserves and be cut off from further oil extraction while still managing to sustain that reserve for a small local population?

7Wannabe5
Posts: 9369
Joined: Fri Oct 18, 2013 9:03 am

Re: Daylen's Journey

Post by 7Wannabe5 »

Yeah, that's around where I would currently put the odds too.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

Metabolism, Temperature, and Scaling

All agents metabolize their arena. Some agents metabolize other agents. Primary metabolites contribute towards growth, development, and reproduction, while secondary metabolites tend to mediate antagonistic and mutualistic interactions between agents. Increasing temperature increases metabolic rate. Increasing agent size increases the efficiency of metabolic processing while also slowing it down and thus extending agent lifespan. These two power laws roughly cancel to homogenize the total number of heart beats over an agent lifetime across scale.

I first ran into these power laws a few years ago when "Scale" by Geoffrey West was popular here on the forum. More recently after chatting with @mountainFrugal I came across the Metabolic Theory of Ecology (MTE): https://en.wikipedia.org/wiki/Metabolic ... of_ecology and https://www.pnas.org/doi/10.1073/pnas.1423502112

Looks like there may be slight differences between plants, endotherms, and ectotherms at each scale but the 3/4 power in Kleiber's law seems quite universal. This makes for the perfect nexus of constrains when simulating a biotorus. The agent size and temperature influencing metabolic rate which in turn regulates food consumption or hunting/grazing cycles. Secondary metabolites further increasing species diversity or niche availability.

One of my original motivations for producing Volution was to be able to watch speciation and extinction occurring over great swaths of time. Smaller agents undergoing a more rapid evolution at a higher speciation rate. Though, the real challenge is reconciling a slow grind of watching an agent in real time with a fast grind of watching whole biotoruses explode in diversity and eventually stagnating or undergoing mass extinction. The first running as 1x speed and the latter at like 10^3-10^8x speed, or up to a million years per day. At some point you run into compute limits and must either slow down or approximate agent interactions. I figure this can be worked out as I go. :lol:

Mainly, the goal is to get something interesting to happen in terms of dynamical systems theory. It would be nice to be able to play with stability, phase spaces, limit cycles, and the like while also being able to analyze agent behaviors and perhaps even minds. Then perhaps once the sandbox is "fun enough" I will polish it a bit and sell on itch or steam for like $5.

A ring in a K2 civilization once built is very likely to be populated from the overflow of an existing population (like earth, mars, or other rings). This kick-starts the evolutionary process a bit based on a coalescent view back in time. The time between the emergence of alleles in a population ecology increases almost exponentially looking back to the LUCA (last universal common ancestor). So, migrating a mature ecosystem with agents scaling across many sizes from ring to ring catalyzes the introduction of novel alleles into the future.

Even if our future is unlikely to lead to a K2 civilization that can afford ring construction projects, the project of building a simplified model of agent populations may provide perspective on the rather more complex case of Earth's biosphere. Perhaps a biosphere can be approximated as some linear combination of various biotoruses that fix altitude and biome. Each ring could be thought of as like a terrarium and/or aquarium that maintains homeostasis.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

Agent Movement

Movement can be framed as a control problem where a steering force is applied to correct the current velocity towards the desired velocity. Steering force vector = desired velocity vector - current velocity vector.

The desired vector can be calculated in a variety of ways. Each way being bundled up into a behavior leaf that can combine with any number of other leaves to smoothen movement transitions.

seek(target) desires a vector pointing towards the target (target position - agent position).

arrive(target) desires a vector that seeks the target until within a certain range of the target, then proceeds slowly and precisely to target.

pursue(target) desires a vector that anticipates where the target will be in the near future.

flee(target) desires a negative seeking vector.

wander() desires a vector that adds a more stable and lengthy primary vector to a less stable and shorter secondary vector. Like a random walk but with more purposeful movement in the primary direction.

flock() desires a vector based upon cohesion, separation, and alignment with other agents in a group.

Some other movements include circling around a target, spiraling away or towards a target, and migration based upon a vector field controlled by the arena.

These movements make up the motor of agents which combines with sensors for sensory-motor control.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

Institutions

Perhaps the arena is smart enough to detect when a particular institution is desired by the agent population and temporarily erect a structure with origami. Or maybe the agents access fab labs to construct and deconstruct structures.

Either way each polygon from triangles to hexagons represents an institution.

Triangles are spaceships. They may be privately owned by agents or publicly owned by the arena. They allow agents to participate in multi-ring economies.

Squares are forts, houses, or businesses that are privately owned. They can be used for defence, storage, or as a base of operations.

Pentagons are publicly owned buildings that facilitate democracy. They could be as simple as poll booths or as complicated as congress hearings. Here the rules of the ring are decided upon.

Hexagons are publicly owned gaming and sporting events that test agent abilities in controlled environments. Some events may give away prizes which may influence agent fitness. The outcomes can be betted on, unless betting is outlawed for the ring.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

Hex Games

The hex games take place in hexagons composed of tiles that are generated by nesting four different hex types: open, closed, mazed, and noised. Hexed tiles can be wrapped around tiles indefinitely. The first layer requiring six hexes to wrap a single hex. The second requiring 12, then 18, and so forth (6*layer).

An open hex is comprised of tiles without walls. A closed hex has a "fence" at the edge and is open inside. Mazed hexes are generated using randomized depth-first search. Noised hexes are generated using simplex noise with mazed hills.

There is a rather large space of possible hex arenas these four generator functions can produce. For instance, an open hex of size 3 could be nested in a mazed hex of size 5 inside a closed hex of size 7. This arena could work well with a game like king of the hill where the objective is to maximize time spent in the open center.

Current games I am considering are hex golf, hex cheese(*), hex tag, hex ball (like soccer or football), hex survival (battle royale), hex hill (king of the hill), and a few more I can't recall at the moment.

(*) Like mouses in a maze gathering cheese.

The games are meant to be quite simple since they are mainly for artificial agents to play. Though, perhaps another project can be spun off that allows players to play against agents they have volved with Volution.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

Agents and Vegents

Volution is concerned with the eukaryotic domain of the phylogenetic tree. Agents are of the animal kingdom, and Vegents are of the plant kingdom. Below kingdoms are phyla that roughly distinguish between various body plans. Agent and vegent bodies are approximated as spheres that appear as circles from any angle. Body plans or phyla are thus implicit in the behaviors of the gents, though some genes like "can_walk" or "can_fly" may give explicit hints. Each gent having a randome, genome, metome, phenome, functome, and connectome that correspond to different sections of their script.

randome - probability distributions and sampling (random number generator)
genome - inheritable and mutable variables (variables)
metome - metabolic processing based upon mass, external temperature, and internal temperature (process)
phenome - behavior tree (physics process)
functome - functions used to construct behaviors (functions)
connectome - signaling to self or environment (signals)

In Godot, agents are scenes that extend RigidBody2D and vegents are scenes that extend StaticBody2D. I am still working all this out but each agent might have a RayCast2D node called Vision, an Area2D node called Olfaction, and an Area2D node called Audition. Vegents having an Area2D called Vegetation that can be consumed by herbivore/omnivore agents. All gents going through a lifelong metabolic process of growth that scales mass and size. I haven't figured out the exact equations but each gent could have an age that is relative to their reference frame (e.g. a ring at non-relativistic speeds) in addition to a metabolic age that is dependent on their metabolic activity and ticks down to metabolic death. This is a little tricky and non-linear as exercise generally increases lifespan. Exercise could be thought to increase metabolic activity locally and decrease metabolic activity globally.

Ideally, the gents should be very independent but still reliant on their arena. The arena being a ring that has a biotorus and a magnetotorus. The biotorus having a lithotorus, a hydrotorus, and an atmotorus. The lithotorus, hydrotorus, and atmotorus each being an Area2D that calculates information like temperature, humidity, pressure, and so forth as well as handling weather and routing gent interactions.

The magnetotorus is responsible for perspective taking on the ring. This involving managing the camera which can be in a detached, third-person point of view or attached to a player (CharacterBody2D) that is controlled by the player/user. Perspective taking is strictly 2D, but the bodies are capable of switching collision layers/masks in a way that imitates 3D. That is, an agent of the atmotorus, being capable of flight, generally hangs out in collision layers above the lithotorus and hydrotorus.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

Windowed versus Lit Rings

To support a biotorus, rings must either filter out bad radiation from the sun or imitate good radiation from the sun. The latter seems much easier engineering wise. The former would have the inner half of the biotorus windowed to let light in while also reflecting high-energy gamma waves and such. I don't know how this could be done without highly advanced mirroring materials that curve light around the protective magnetotorus. The much simpler option is to close the ring in a dense magnetotorus and just emulate the outside universe with an LED screen on the inner half of the biototus. This could easily be done with just solar power as with most of the vital ring functions, but I am presuming fusion power is readily available anyway. The closed option also avoids catastrophic failure of the window resulting in the complete devastation of the biotorus.

daylen
Posts: 2528
Joined: Wed Dec 16, 2015 4:17 am
Location: Lawrence, KS

Re: Daylen's Journey

Post by daylen »

Ring Cycles

I am imagining a standardized ring that has a central river (i.e. the hydrotorus) spanning over half the land (i.e. the lithotorus). If the ring is spinning around every minute or so and acting as a giant centrifuge, then would there be relative spins between the lithotorus, hydrotorus, and atmotorus? The lithotorus would be fixed to the spin of the magnetotorus. Though the hydrotorus and atmotorus could experience a delta in pressure/friction that leads to the inner water/air moving slower than the outer air/water near the edges of the lithotorus?

Basically, I am curious to what extent a spinning torus filled with liquid can experience turbulent flows? Would laminar flow be the norm unless there is some major temperature or pressure delta across the ring?

Many agents will continuously migrate around the ring forming their own cycles. The ring could emulate Earth's day/night cycle just by shifting around when the sun is visible on the screen, allowing gents to follow a diurnal cycle. Perhaps some agents could match up their sleep schedules to their migration schedule to always sleep in dark or always awake during the day. Some agents may just go with the flow, literally, by flowing with the central river.

If migration lines are longitudinal then predation lines are latitudinal. Agents awaiting behind vegents to attack migrating herds of agents. I wonder what other kinds of ecological cycles could emerge from relatively simple behaviors. The arc of creating an ecological simulation is to very gradually build up complicated situations from simple compounding behaviors until eventually the simulation becomes complex. At which point you become more like a navigator of the ship that is the simulation. Hence why I spend 100x more time thinking about the architecture than actually coding. :lol:

User avatar
mountainFrugal
Posts: 1125
Joined: Fri May 07, 2021 2:26 pm

Re: Daylen's Journey

Post by mountainFrugal »

I had a few follow-up thoughts to our conversation:

1) When building these models you go really deep for each layer or node, come up with a nice summary abstraction, and then move away from the details. How does it work to revisit the ideas again? It could be that as you learn more in another area it could inform an area that you already went deep on... so how do you go back, if at all, and revisit the older material/learning/modeling?

2) After seeing your conceptual diagram research notes stack (not in detail), do you have any examples already written out on the forum of how you are condensing a field of information into an abstraction? Would you be willing to make a simpler example of the process with the caveat that simplifying may not be how it is really done in your research process.

Post Reply