Beating Darwin: Accelerated Evolution in Thrive 

For those of you interested in the results of my further experimentation, I've gone ahead with making something I can put on steam, and now it's got a page! store.steampowered.com/app/2603740/Atoms_and_Time/

The greatest thing that improved my gameplay in thrive was learning to ignore auto evo when the time came to gamble for better adaptations.

I know I'd play and buy a fully fleshed out evolution game- the market is definitely there I think, remember how insane the hype for Spore was and even though it barely delivered on its promised features it still sold like hot cakes and was still dearly beloved by gamers. (myself included)

I loved the little animations in this- they had so much personality!

This is a great video! I am inspired to learn more about evolution now, thanks for that!

I did not appreciate the mother roast good sir :

Awesome video, great production value and very interesting topic! Keep up the good work.

This might actually work decently well to simulate evolution since it tends to be exploitation of new niches that drives evolution. Heck, predation (A killing/eating B, including herbivory) would just involve creating a new Miche tree with "Eats/Kills B" as the root node each time a new species evolves (and just deleting the tree when that species goes extinct). The hardest part would be figuring out how to dynamically assign or update a Miche tree with new nodes (or prune no longer valid ones). This is a really cool idea!

On Natural Selection, scientists have produced studies that suggest mutations are not just random but biased. Theories of what it's based on vary

I made these usernames before I thought I would want an online presence, and it's too late to go back now: raggaaa.itch.io/

Wow man, loved the video, put toghether very well. I hope this project continues to improve :)

I ran a text-based evolution game that tried to simulate generational niches in different environments. I did it by mutating the existing species to create randomized offshoot species, and then checking how well each living species eats every other species, assigning them a score based on effectiveness in every environment/scenario where they could eat them. This gave each species a list of other species that predate on them in those scenarios - lets say species A and species B both have the ability to predate species C. If A outcompetes B in all scenarios where the species can be eaten, it will be the only predator of C. But if species B is better than A in one of those scenarios, it eats a percentage of species C that the one scenario allows, and both species coexist as predators of species C. The remaining species become the new generation and this whole process is repeated again, applying to all new and parent organisms including plant life, detritovores, herbivores, carnivores etc. I found this system to be quite effective because every new mutation provides a new food source that every other organism can compete against each other to access, so the ecosystem constantly shifts around as species become specialized for different niches. As species gained unique traits and lifestyles, some species would specialize for one food source, whilst others became more omnivorous and generalist in their diets. The main flaw of the system is that it required human oversight to adjust the scenarios based on the environment and how its shaped by the ecosystem, since it was a text game lacking any concrete game mechanics that could simulate those changes made by the species present. I think in thrive, there could be enough of a simulation to represent the altered environment and have that affect species success rate in various situations, but I'm not sure how the system would account for those changes autonomously. Maybe some sort of rudimentary neural network that separates in-game species interactions into distinct categories to compare success, maybe? Or maybe simulated in-game areas populated by species, representing each scenario, where the species performance is tested. Neither seems all that feasible - Using existing world locations for example would not provide the intricacy of measuring two species on how they hunt aquatic prey at night vs day for example, or from above water vs inside of water. The scenarios would become more and more intricate as things such as species size, lifestyle, migratory patterns, etc diverge, introducing complexity as to what counts as a unique scenario to test. For example with large organisms vs small organisms consuming the same food source in different ways due to their body shapes and lifestyles, or consuming colony organisms from inside vs outside of the nest. Nevertheless, i thought I'd share this info because there are some similarities between my system and the system that you have developed, so it could perhaps provide some inspiration for possible additions to thrive in the future.

13:42 Mutations in Darwinian evolution aren't completely random. Variation isn't solely introduced via point mutations. Even prokaryotes heavily skew where variation gets introduced, with some parts of the genome highly conserved, as changing them would likely lead to a catastrophic failure of the entire organism. They even use horizontal gene transfer to introduce new genes or alternative alleles of existing ones via conjugation (from the neighbours), transformation (from their neighbours' corpses) and transduction (going viral). Let alone highly complex multicellular eukaryotes, with such classics as sexual reproduction with 2 chromatids (haploid) or more containing duplicates of the genome with slight aberrations. Or crossing over to introduce additional variation, but again limited to remixing small parts of your genome between your two chromosomes from your own parents, before passing the recombinant DNA on as your half of the parenting part. Most mutations that would immediately be fatal are avoided, instead of relying on pure randomness and selection then sorting things out. Breeding is done by humans, which are part of the same ecosystem they're trying to change and their evolution influenced by it, so it's not fully artificial selection in the sense a god or computer programm might be able to employ. If we breed cows to produce milk for us, we also adapt to digest milk well into adulthood. So both completely random introductions of variation, as well as entirely artificial selection processes are severely limited in the real world. Doesn't mean your auto-evo (I assume from 'automated' not 'auto' as in 'self'?) is Dawinian evolution despite both not containing completely random mutations, of course. Though it might be: If only entire populations can mutate and evolve, not individuals, one might simply consider each such population as a single multicellular organism, with each individual representing one cell (albeit an undifferentiated cell, as all individuals in each population are the same, if I understand this correctly). Cells also reproduce when part of a multicellular organism, of course, just without introducing mutations (that might end up as cancer) - so that fits the analogy, I think.

It was awesome working with you! The video turned out really nice! Also it’s now cannon you’re a sassy disturbance. Sorry- 🤷

Root systems preforming different co dependent functions for the plants.

"speaking as god" you say that which such confidence. Love that for you.

Wow that was cool. Good luck on the project 👍

1:54-2:12 Don’t quote me on this, but I wonder if it has something to do with the water and nutrient composition of the soil the roots of each plant species grow towards, and specializing to send roots to different spots in the same patch of soil with slightly different compositions. Again, just speculation.

9:16 "Mutational Precision" would be a compound rather then a phrase in my opinion, so yes noun is appropriate.

Honestly seeing the beautiful world god made for us (by evolution BAPTISTS), and seeing the complexity of it all, it truly deepens my appreciation of biology.

8:32 I thought it was for them to damage any prey they couldn't engulf on the way out

The next milestone is making it 3d!

Have you tried making major changes that marginally increase the accuracy of the simulation, but take forever to implement and are orders of magnitude more computationally expensive?

Miches is a very cool concept

Great video!!!

I am going to keep up with this also if it works on Chromebook’s I’ll probably try it out so good luck and keep it up

You used me as an example!! 4:11, this means i can die knowing i was in a video

gah i hope i am not to late, this tech demo sounds interesting i will give it a spin.

is there an autoevology discord or discord about these ponderings? i’ve thought a bunch about this for a game i’d like to someday make (less about evolution during gameplay and more about creating a realistic biosphere) so while it’s not tested in practice, it’s something i’m pretty interested in, along with simulating approximate plate tectonics (again for making realistic worlds)

👍

cool video

2:04r the secret being they don't occupy different niches and are in direct competition with each other. ala attenburoughs life of plants.

Thrive

is it possible for a species to change how it acts in its environment but not genetically, like if its lower leaves are eaten alot it stops growing them but it still has those genes so its children will still grow low leafs, this general nature that listens to environment heuristics would allow species to better during change as they could maximize multiple niches as the same species and also niches of combinations that dont exist yet

One comment on your editing: please don't go from a back screen straight to a white screen. I got flashbanged 4-5 times.

Make the real game. Just the evolution aspect of the game wil be interesting for Shure.

nah id adapt -mahoraga

👍