I think the reason why you got bigger creatures in your 100h simulation but smaller ones in this is because while your 100h simulation was (relatively) very stable with lower competition and only a few species, this one had a very high competition, low resource, and instable environment. Since an instable environment favors creatures more able to exploit smaller, short-term advantages, this experiment favored smaller, faster reproducing creatures. Meanwhile in a stable ecosystem with little competition, creatures with longer-term advantages are favored.
@@saosaqii5807 we aren’t really big though, and definitely not compared to the competition around the time. It’s just the minute we found out about making spears as sharp as tiger claws and shields as tough as elephant hide(not sure if those are actually tough) that allowed us so much versatility the competition just became irrelevant
I'm laughing so hard that Fodder, all of Bibites, beat out the most intelligent, highly engineered Bibites, despite its creator just constantly dunking on it, thinking it's a worthless pile of Fodder.
Fodder needs to persist. The bottom of the food chain is obviously going to win out from an evolutionary perspective. I mean, how many mass extinctions has algae made it through?
Maybe the person who made Fodder knew it wasn't that bad but called it a "miserable little creature" because then it's funnier to watch carefully engineered bibites fall to Fodder. I sometimes name myself something like "HarmlessFly" in video games for that effect.
you know what would be cool? if some parts of the world would have different simulation rules. like friction or food efficency. making local pockets for more niches
Sounds really interesting, especially since it sounds like it would be relatively simple to code, and diversity has been lacking in this video and the previous 1000 hr simulation.
@@neumo5005 I actually think it should come first: it would allow to simulate very roughly what the plants would do before their implementation (which seems to be far in the future).
Yeah, I think the next big step for Bibits is varying environments to discover how bibits evolve with varying evolutionary pressures, food availability, etc. Something more complex than the "Islands" test run
Agreed! Life can vary so much in a small space between tide pool, beach, reef, floating plastic, side of boat, and deep sea. "Terrain" of simple rule changes or obstacles would create really cool niche islands and allow more diversity in one simulation. Great idea.
Adding a Bibite identifier string would fix the problem with tracking generations. The identifier would start with 2 or 3 characters and add a random character (a-z, A-Z, 0-9) with each mutation creating a unique string for each species and let you track the evolution easily. The number of characters will identify the number of total mutations and the number of identical characters at the beginning of the string will show at what point species diverged. For example, bibites a8j3ktV8p and a8j3ktc4q diverged after the t mutation.
It might be interesting to test some of the contestants that are out already against the winner, to see if the selection round had very different pressures to an actual tournament round.
Can't you do a smaller separate tournament for honourable mentions that didn't make it in the tournament? I believe that a lot of us want to see how more diverse bibites would compare against each other! Like some of the ones with very complex brains like Subrufa aversor and Einstein, and some very big ones like Cerelus viator and The Immortal. You could let people vote on the 70 (86 -16) remaining bibites that they would like to see and then choose the top 8 or even top 16 for a side tournament?
@@chrisbolland5634 So far. I think that when the food delivery system evolves the rest will too. I'd love to see different types of vegetable, some not edible by everyone, and so on.
The 16 semi-finalists are all small lizard analogues. They grow fast for a short time then stop. Most of them only grow in the presence of food. A distinctively reptile trait. And they are all solitary. Not one of them has a positively connected herding neuron.
could be that the simulated environments design is inherently conducive to such a creature (probably just one of many causes; a lack of environmental niches probably contributes to this, as well as feature costs being overtuned in certain cases)
what i would love to see being added is some kind of "warp speed mode", activating it would disable the visuals completely (or just lower the framerate to like 0.1 FPS) and run the simulation as fast as the hardware can handle until you deactivate it. or just allow the maximum speed to be a bit higher than the x11.18 it currently it, so you can speed up the entire evolution process
hmm i see, then i wonder what else could be done to help speeding up the simulation. maybe off-load some work onto the GPU? for example throwing individual neuron calcutations at individual CUDA cores so entire layers of the neural network could be done in parallel while the CPU works on other stuff
To be honest, it would have been more interesting to have just kept your original simulation running with the final sixteen contestants than to have done a ranked tournament.
1v1 doesn’t make scene here. Even putting each bibit up against every other one at a time makes no scene. Many battle royals in different starting conditions might be a better ranking. But looks like the simulation is just too slow on normal computers. Can this be sent to a render farm to run at faster speed. Can it run on GPU. Can it run on analog computers. I hope that maybe one day it can.
I think brain costs should be lowered. They should definitely have a cost, but the most interesting behavior is tied to the brain, and with how predation and difficult-to-obtain-but-high-energy food are not as present as would be hoped, there's really no current reason for creatures to evolve the complex-but-benign brain structures that eventually through luck become very clever behaviors; it's just more efficient to have as small a brain as is useful.
Looking to real life, most animals will reduce their brain function when they don’t have enough calories to spend on more complex neural activity. Maybe both introduce a way to toggle some neurons and make the cost based on how many active synapses the creature has? Would also have the bonus of tying the calorie cost of brain function to the actual processing cost :p
it actually isnt that bad. as long as the brain isnt stupidly big wasting so much energy it cant be helped, it should be smart enough to reserve energy if its brain is that big
what if someone engineered a giant bibite with anti-parasite herding behavior? like if a giant bibite detects another giant bibite with a parasite it will attempt to kill the parasite or something idk how exactly it would detect them 'n whatnot tho, but if possible it could have made them last at least a little longer
It would need to be able to remember its health so it could detect a loss of health. But I don't know if this implementation allows recursive connections. I didn't see any in the brain maps shown. It's slightly easier to code NEAT without recursion so I suspect they are disallowed.
Doing tournaments like this are quite interesting. May I suggest that you also add a completely randomly generated creature to the mix as well, for the same reason that a random move chess bot is added to a chess engine tourney? It'd provide a good baseline.
I really enjoyed seeing the different approaches to preparing bibite competitors, the breakdowns of their behavior and brains, and your encouraging reactions to people who didn't have faith in (or seemed contemptuous of) their species. Thank you for another great video! Eagerly anticipating the next video :0
I'm only part of the way through but I have to say I really hope someone made an ant bibite that uses the pheromones to make trails and follow them to food. IMO that's one of the most interesting directions I could see this project going towards.
This is the first I've seen of this program, I wonder if you can engineer for a species that's small and efficient but herds tightly to sweep the map of resources, with the ability to swarm and kill anything else that comes near and get that high value meat. Kind of like a piranha but engineered to be more systematic
@@sapphire--9375 everyone runs towards food when found with only the fittest one survive to sweep for the next Pellet 😅 Leaving behind a massive trail of blood and vast, empty spaces. True Tyranids
It was nice to participate on the project, sadly I only had 1 day to engineer my bibite so I couldnt expect much PD: My species was the Ankymilon, and a friend of mine created the Microfliatus. So sad they didnt appear in the vid
Loved how you explained the brains! I left this video suggestion on the subreddit quite a while ago when there was not a lot going on there^^. (Probably not the only one :) ) Growing faster when being full is a great strategy, its investing in the bibits future when it can afford it while not waiting energy while in a drought. Its hard to imagine why that wouldn't be great for every bibit.
Here are a few of my future ideas for the Bibites(I don't expect any of these to be added to the sim, don't even know if they're possible, but whatever): 1) Different kinds of plants. Some plants could be plain green circles, others could be a much darker color, some could have leaves, or be poisonous, etc. Maybe certain plants can only live in specific zones on the map. Also, Bibites that live around specific kinds of plants would adapt to only be able to eat those plants. 2) Habitats/biomes Throughout the map there could be different biomes. These biomes could be differentiated based on climate, like having a hot equator section in the middle of the map, and a cold tundra at the top of the map. Bibites and/or plants in that area would adapt to their climates. I.e., Bibites in the tundra would have blubber to keep them warm, and ones in the equator would have a slow metabolism since food would probably be pretty scarce. 3) Mass extinction events. Every once in a while(maybe every 2 hours), there's a small chance that a mass extinction event would take place(say....5% chance? Player could customize the chances and the time intervals). These events could cause things like rapid climate changes in certain areas, or toxic gasses appearing in the water(or whatever the Bibites float around in). Maybe meteors could occasionally strike and wipe out a portion of the map! Again, just a few ideas of mine that I think would be fun to have in the simulation. I look forward to the next Bibites updates, and your next video!
I don't know why RU-vid recommended this but I love it, and I want to see the Bibites fight to the death to see who is really the best at surviving. You got yourself a new sub my friend. Good luck with the tournament all contestants that made it to the top 16.
the fact that the winners were all very similar probably means it wasnt luck. itd make 0 sense even with being lucky to see very few of the winners from last time and new bibites winning in a repeat ffa, it seems like the ffa settings and the whole enviroment heavily supported small, fast reproducing herbivores
I just stumbled across this project not too long ago so first of all, wow, this is *VERY* impressive and extremely cool. That said, I think this would be even more interesting if future tournament rules didn't directly penalize brain size as heavily.
Make symbiotic relationships viable please! I tried to make a cleaner bibit that ate parasites off of big creatures, but there was no way to stop the bigs from eating the cleaners
Attempts to summarize the finalists for future reference (under the cut to avoid spoilers): Apophis apocalypsis (Skarix): Herbivore, likes to herd while eating Beyblade (pod_Scho): Omnivore(?), cannibalistic tendencies, goes spinny, avoid each other Bibby (Cole H.): Herbivore?, Funny workaround bibbit, fast maturity, grabs food and leaves Darwins disaster (Rapha'e'l): Herbivore?, Cyclops, terrible field of vision, rotates to see, grabs stuff and grows, avoids others Fodder (Bert): Herbivore?, small, weak, inefficient, dumb, miserable, pitiful, fine with groups, likes to yeet pellets Luscus xHybridus (Kazak): Herbivore?, Zooms around, stops while eating Magnus terra (Serih O. L.): Herbivore?, fast reproducer, infinite growth(?), slows while eating, avoids others while eating Micantes oculi (Kavitaet): Herbivore?, like to dance near red pigment Minima hunorum (JerseyImperator): Omnivore, latch onto food, tendency to steal, seek out meat Multido insectum (Unfamiliarmaniac): Herbivore?, grabs stuff and grows, always moves forward, no herding Nubbi competitor (Sweev): Herbivore?, dynamic growth control Obscuris Irrelevantis (Nobody 103): Herbivore?, Ability: Defeatist Parva fragrum (Guillermo R.): Omnivore?, spam eggs, mature babies, always starving Parvum caerelum (Coqueiro L.): Herbivore?, grabs stuff and avoids others, grows faster when full, drops pellets at random Ramsey Jr. (Ramsey Rude): Herbivore, slows or stops when eating, large Skippy grabby (Crabford): Herbivore?, skips and grabs, lovable, demanding growth
I would highely recommend objects to have surface tension. It would cause some very complex situations and new kind of learning. I would love to see this implemented.
@@pietrom2642 Since everything moves like it's in water. Why not reinforce that idea with surface tension and having creatures trying to overcome surface tension or welcome it as it could be used to help feed and reduce energy.
I have always been obsessed with ecology, zoology, and evolution before i could read. I used to spend all my time as a child outside trying to record what each animal ate, and how they reacted to natural events, like how do ants stop their tunnels from flooding Now as a young adult i am trying to breed new varieties of plants to produce yeilds at different times, or to survive different growing conditions without compromising the yield Basically at all times in my life, atleast one of my favorite things has been evolution based, whether the selection was artificial, or natural. For years I've always wished someone would make a smart, evolution based computer simulator that everyone could use. It may not be as advanced as i would have hoped, but it is certainly a beautiful creation, and by far the best I've been able to find You sir have my respect and admiration. Your creation is the greatest thing my computer can do, and probably all it will be used for for the next month lol
I think this tournament shows that we are on the right track! High competition and low energy availability killed off large or big-brained species. This makes me wonder who would go into the brackets, if energy was widely available.
This is awesome, i got the game and played a little, its really complex and the tutorial does not help a lot XD. But its lots fun, can't wait to see it evolve in something even better!
21:02 maybe that’s because it was introduced to a parasite bibite that latches onto food then eats the bibite once it’s close enough, and constantly grabbing and ungrabbing food makes the parasite run off with the food and the bibite benefits from food and safety
I came back after a week to peek at the comment section. I wanna say that I disagree with those saying your voice-over is annoying. I may be biased, but I very much enjoyed it and hope to see something similar in future tournaments. Cheers💖
When imagining a fairly simple videogame environment, I can't help but think an engineered one would win. This just feels like the sort of thing where someone finds a novel strategy, and it's easier to turn up to the max than evolve there organically. Depending on how local maximums might appear, engineering also has more potential to surpass those.
I think I know how you can get predation. you need decomposers. maybe you need an intermediate step where there are some bibites that can eat deceased bibites, training them to go after dead ones, and when the dead ones go away, they might need to adapt to eat live ones or something like that.
are the finalists gonna be used as they evolved over the course of the tournament, or as they were submitted? I'd love to see a much longer run of this, starting with every contestant. It makes for a much more dynamic environment, more likely to be interesting. I'd *especially* love a repeat of this once sexual reproduction / crossover is back in: Imagine all the hybrid potential of a start as complex as this!
Interesting that all the finalists are quite small. Most Bibites I evolve end up with selection in favor of being pretty big. Sadly they also seem to evolve to drop their mutation rate which means it can reach zero, ending their evolutionary progress.
Oh, and I just thought something: making bibtes see eggs. unlike live bibtes, eggs can't run away from predators, making eggs an easier food source or carnivores. Probably not an original thought, just bringing it up...
I have a question, how does the plant life grow/spawn around? Perhaps feces could be used as a fertiliser which would spawn food around it. If you won’t make the children eat sh*t make the feces grow what they eat. Might just be me but i think that’s a solid idea.
On more notice then they get attacked they show no response at all so think you should ad them properties to have noticed been attacked and give chance to respond in that. And that should make many interesting interactions and turns in simulation. I love your project its awesome. 👍
Is it just me or does anyone else find it to be a near impossibility for a non-filterfeeder Bibbit to be successful? Ive ran 100's of hour of simulation with so many different settings. All i get is constant generation 1's dying and spawning in for all of time. Or i get a filter feeder. I'm really looking forward to some updates of The Bibites. Keep up the amazing work!
Maybe some day you could develop bibites that eats eggs and some protectiveness scale simultaneously. So some species puts tons of eggs and ignores them, and other that puts fewer but are killing protectors. A scale of time of "hatchness" would be good. Some species take a long time for the egg to hatch, others don't, depending on the complexity of the being.
