Video wird geladen...
Video konnte nicht geladen werden
building a force-based alife sim with fable 5.2, including hunting, avoidance, and blended gene inheritance. 🎶 μ-ziq & mrs jynx — secret garden
24,261 Aufrufe • vor 5 Tagen •via X (Twitter)
35 Kommentare

i want to play this like it's spore

@grgrdvrt Omggggg love this

@grgrdvrt thank you! feel like there’s endless room to explore these systems. check out @OA_paperclips and @arsiliath’s work for the real good stuff.

@grgrdvrt @OA_paperclips @arsiliath Oh yeah very familiar with @arsiliath I even took his shader course!

@grgrdvrt @OA_paperclips @arsiliath me too, such a legend! took the compute shader course over covid and opened my eyes to high-perf sims.

@OA_paperclips @codetaur @majormobius @abelian_soup

Do the hunters start to look different after a few hundred generations of blended genes?

it’s not super stable past a few generations (haven’t tuned the feed/decay much yet), but you can start to tell features in the offspring from what the parent ate (for better or for worse)

Fair, tuning feed and decay will probably eat your weekend

I love this so much

emergent behaviour from simple forces is fascinating

woah what inspired you to do this

i’ve loved biosim/alife/cellular automata work for a while now! but had never done a full force-based one before, and fable 5.2 was able to give me a base to build from.

nice! relief to know there are designers out there who are bio brained like this

Amazing

love the art style

A really nice video! Is there any further explanation somewhere about the details of how this simulation works?

that's iconic :)

Is there an incentive to consume energy?

yeah! species expend energy to move and hunt, and eventually die out. there’s no explicit “avoid death” directive, just to follow food source trails.

>> including hunting, avoidance, and blended gene inheritance. nobody: "love", "symbiosis", "arousal", "reverence" everybody: "All. Hail. The Killing Spree!"

So when do they turn to animals?

Thanks for sharing.

🤯

Brute force = hare Deterministic = tortoise The turtle always wins

Does blending the genes eventually make the population look alike, or do distinct groups stick around?

good q! i've put weight on the parent's genes to avoids homogeneity, the parents lose energy when creating a new species, and there's enough random variation in there to keep things interesting (until one super species wipes out everything). lots more to do there though!

Insanee

@fraserpricee be like this, instead of whatever crack you're on

@fraserpricee haha! fraser's work is insane, so glad he's doing it. always wondered what would happen if you gave individual cells "true" intelligence in a sim like this.

@nisapha777 ay real recognise real, this looks sick af 😎 🤝 😎 whats the mechanism? filament-based?

@nisapha777 hell yeah! yeah, partly filament-based. particles have spring forces in chains, with rules around breaking behaviour and energy consumption. cell "walls" are filaments connected to themselves, the "hairs" are filaments with an open end.

@nisapha777 very cool; your rendering mogs mine too 😂 are there mutation/selection mechanisms, or is it a fixed/preset population? good one to read from Sakana/ALife Institute if you havent, they also did a filament-based sim recently :) @ciaran_regan_

and your implementation eats mine alive! mutation is very basic rn — every species is spawned from a set of ~60 behavioural scalars, and i just track what the parent consumed then spawn a species with those scalars weighted together. love sakana (have been following @hardmaru's work for years) but somehow missed this! will read tonight, appreciate the link!

@nisapha777 @ciaran_regan_ @hardmaru honestly this is my first time playing with continuous alife systems, all my work before was discrete and grid-based. lenia always fascinated me but i never went too deep with it. have you been to any of the alife conferences over the years? always curious what they were like.
