Rendered at 03:52:02 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
andai 23 hours ago [-]
This is neat. It doesn't seem to conserve volume. If you spread the blob "inward", all the fluid disappears and you get a tiny thin spot.
That's a general problem I often run into with simulations. When you get it in the opposite direction it can have some very amusing results (e.g. energy in a physical system continues to increase).
apetresc 16 hours ago [-]
That's just the cream being absorbed into the finger/surface ;)
lbriner 11 hours ago [-]
She mentions this in the section "This Is Not a Physically Accurate Fluid Simulation"
charlesrice 13 hours ago [-]
If you linearly squish all three colors into the center to try to mix them, you end up with so little paint that you can't even paint outside the initial circles and mixing region, even if you do the circular motions that are supposed to make more paint.
Tadpole9181 22 hours ago [-]
You can also make infinite paint by spinning a small circle, then cycling between larger and small circles! It seems it also converges to secondary colors over time if any other primary color is added. Very cool!
They explain why this happens in the article.
22 hours ago [-]
etatester 22 hours ago [-]
The conclusion made me sad:
> Publishing this may mean the method is no longer something special.
augment_me 21 hours ago [-]
If you step one step outside your ego for a moment, and see the world from a place of connectedness and unity rather than individual achievement and putting your name on things, you won't have to be sad about a statement like this.
crashabr 17 hours ago [-]
What is connected when my LLM, extrapolating from a vague prompt about a cool liquid effect, takes this method and reproduces it without acknowledging the original author? Certainly not me and the author.
Using connectedness and unity to defend the fragmentation of the social bond is ironic, to say the least.
augment_me 12 hours ago [-]
The information you produced (probably aided by your LLM), was used by another human (and probably their LLM).
Historically, like little children who find it hard to find intrinsic self-encouragement and value, many people needed to have a pat on their back for their work "hey good job buddy you made it first".
If you still need this, its okay, we all grow up some day and start making things for the sakes of it (giving back to the community, etc), but this is not something inherent to the human but is rather a learned ego-seeking.
xpct 21 hours ago [-]
Next time someone wants to make a spreading liquid effect, their LLM is more likely to know how to make it. They won't have to read this work when looking for references, won't get to know the person who thought of this and shared it with the world.
Surely it's at least understandable why the author could be feeling a bit sad about it?
nomel 21 hours ago [-]
If I made something for fun, for the public, for free, and it exploded into common use, I would be super happy that I made something useful. But, I also always license any software I write (outside of work) as public domain like.
strogonoff 19 hours ago [-]
We are social creatures, and wanting to impress fellow beings is part of the reason why we do original things.
When you made something public domain pre-LLMs, you generally could expect people to give kudos to you.
With LLMs, if you make something public and an LLM gobbles it up, not only you get zero credit but you also can’t convincingly claim to have come up with it (since anyone can just assume you used an LLM), and neither can anyone else (even if they have come up with it fully independently).
If you don’t seek any recognition for your original work, then that’s fine, but I suspect you might be in a minority.
tag-ppi 12 hours ago [-]
I guess nobody is going to create or get recognition from anything new starting from today until the end of the world?
augment_me 9 hours ago [-]
I mean if information is decentralized, this is not far from the truth. Historically attribution was given because you could synthesize or create information in a way others could not, and it was useful to be able to trace things back to the source.
Now everyone can create everything, and the source is shared. You are adding to the pile of human knowledge and it gets processesed, that's it.
If you step aside and think about it from a holistic perspective, imagine a research project to find a cure for cancer, or the best compression algorithm, or whatever.
Who will find the solution faster:
1) Entities work in isolation, produce their results, gets attribution, and then it's integrated into other work.
2) Entities work in parallel, dumping their findings and work into a big pile that is continuously parsed and built upon.
The second approach does not discriminate partial or bad results, it does not place value on a specific contribution. It might be that two entities did partial solutions that only make sense together, and this is able to be synthesized on the fly.
roncesvalles 16 hours ago [-]
Except there are people out there who will rip it and present it as their own. Egolessness works only if everyone is egoless.
doublerabbit 17 hours ago [-]
It's okay to cherish a knack you've come across and it is sad to hand it off. It is special when you produce something traditional.
Show some empathy, the world is far from connected and united.
alala-lala-lon 15 hours ago [-]
[dead]
lichenwarp 14 hours ago [-]
It would be funny if the one that works only with touch gestures left fake fingerprints.
socalgal2 20 hours ago [-]
what is the point of three.js in this demo? It feels like showing me a CSS effect on a DIV and saying "built with React". It's a nice demo. But it's not doing anything that needs a giant library to do.
skrebbel 20 hours ago [-]
Might be as simple as “author always uses threejs for stuff like this and couldnt be bothered figuring out how to put a shader on a texture without it”
brandonpelfrey 22 hours ago [-]
This seems useful (after modifications) for building a calligraphy application.
gosen 15 hours ago [-]
that's so satisfying, except the moments when I tried to spread it on static images :D
Conscat 23 hours ago [-]
I wish foundation really did spread this smoothly.
wickedsight 18 hours ago [-]
The color mixing brings back memories of mixing a bunch of paint colors in high school. I guess that's an indication of how well it was executed.
accumulator 23 hours ago [-]
This is magic
oxedom 11 hours ago [-]
Love it!
soft-boy-xyz 22 hours ago [-]
I'm impressed.
zodwick 16 hours ago [-]
Very cool work ! I was unreasonably disappointed by the element at the top ( header ) not being interactive thgh :_
That's a general problem I often run into with simulations. When you get it in the opposite direction it can have some very amusing results (e.g. energy in a physical system continues to increase).
They explain why this happens in the article.
> Publishing this may mean the method is no longer something special.
Using connectedness and unity to defend the fragmentation of the social bond is ironic, to say the least.
Historically, like little children who find it hard to find intrinsic self-encouragement and value, many people needed to have a pat on their back for their work "hey good job buddy you made it first".
If you still need this, its okay, we all grow up some day and start making things for the sakes of it (giving back to the community, etc), but this is not something inherent to the human but is rather a learned ego-seeking.
Surely it's at least understandable why the author could be feeling a bit sad about it?
When you made something public domain pre-LLMs, you generally could expect people to give kudos to you.
With LLMs, if you make something public and an LLM gobbles it up, not only you get zero credit but you also can’t convincingly claim to have come up with it (since anyone can just assume you used an LLM), and neither can anyone else (even if they have come up with it fully independently).
If you don’t seek any recognition for your original work, then that’s fine, but I suspect you might be in a minority.
Now everyone can create everything, and the source is shared. You are adding to the pile of human knowledge and it gets processesed, that's it.
If you step aside and think about it from a holistic perspective, imagine a research project to find a cure for cancer, or the best compression algorithm, or whatever.
Who will find the solution faster:
1) Entities work in isolation, produce their results, gets attribution, and then it's integrated into other work.
2) Entities work in parallel, dumping their findings and work into a big pile that is continuously parsed and built upon.
The second approach does not discriminate partial or bad results, it does not place value on a specific contribution. It might be that two entities did partial solutions that only make sense together, and this is able to be synthesized on the fly.
Show some empathy, the world is far from connected and united.