I bet you could replicate/enhance some of these results with varishore tpu, trying out the code (seems like it's an IceSL addition) to make semiflex tools without joints would be awesome.
I want to praise this channel in a more unique way than the usual. As a scientist I do know the struggle of keep up to the date, expecially when there is some revolution in other fields that do not belongs to you. This channel does in fact generate an easy glance to all that it is new in the form of a pill, there is no need to read for ton of papers, it is a selection that does basically highligth great potential tools in a time effective manner. I firmy believe that this channel is source of ideas for many scientific groups out there, and it will be more in the future. It is a tool that allows to be keep up to date in a very light way allowing to spend time and redirect efforts easily that one may think. I do belive that tools like this channel will improve by years the adoption of new methods spreading the news way faster than is usually done. Keep it going, it is an incredible work you do for the entire world, this channel is an optimization tool that will be overlooked, but it is useful for the entire humanity.
the video reminded me of the 'crumple zones' you see in cars where certain parts of the vehicle will actually crumple more to absorb impact while the more rigid roll cage stops the occupants from being crushed
What I REALLY love about the showcases in this video is that all these could be printed on a consumer $300 printer. As someone that's been doing 3D printing for years, I love seeing how many ways people are coming up with creative and useful ways to put these little robots with a hot glue gun to use. 🤩
It's even more cool that this solution was found using a handmade algorithm instead of machine learning. Just goes to just how far materials and topology has come in melding math and engineering
Look at the RepRap project lol. We've been doing exactly that for years. Prusa also has a 3d printer farm to make parts for their own printers. These Papers have been made for like 10 years.
@@VincentFischer depends if you allow other printing methods than FDM. You probably could print the nozzle but the surface finish is probably bad without manual finishing.
@CanadianOtaku printing stuff like motors, microprocessors and electrical components? In theory probably yes, but thats still a long long way and probably not possible with the current methods (FDM)
@@sanjaymatsuda4504 Eh, I wouldn't be so dismissive of this. When AI was first introduced the singularity was just speculation, and we couldn't really know for certain when it would occur. The pace of progression in AI since then is insane, we're achieving things never thought possible before, and we have a much better understanding of how to create and train it. AI can already write basic code, it's only a matter of time until it can write the code of a better AI. I wouldn't be surprised at all if we reached the singularity in the next 20 years or so.
@@commenturthegreat2915 pfft we're no closer to actual learning than we were 50 years ago. When we first made computers, people considered them "thinking machines". No, they just do arithmetic really fast. It turns out we can apply that to pattern matching too. That's all modern AI is. It's not intelligent, unless you think a calculator is intelligent. We still have no idea how to do anything like what a brain does.
@@gracefool This is a different topic - the singularity simply refers to a point where AI becomes capable of creating better AI, and we're definitely closer to having achieved this milestone. As for your point - I heavily disagree. There is a massive difference between writing conventional algorithms and training AI using machine learning. In the former you're writing direct instructions for the program to follow, meaning it would never improve without your direct intervention. In the latter however, you could write relatively simple instructions which enable the program to continuously improve without you having to change a single line of code. Neural networks and similar learning structures were meant to mimic the complexity of biological computers and enable them to improve over time through minor tweaking, just like our brain does. And lo and behold, it works. The actual functions and connections in machine learning structures may not be the same as our brain, but they definitely achieve the overall effect - learning. This isn't something that was possible before. AI these days can actually be innovative and break the expectations of the people who created it. This is where this starts to get more philosophical than factual - can this be considered intelligent? I find it strange that you did not actually explain what you would consider to have reached that line. I would argue that pattern recognition is basically the same thing that makes us intelligent, be it on a much bigger scale. After all, our brains also consist of networks and firing signals, and could be considered to be just very complex computers. There's nothing inherent that prevents electronic machines from performing the same tasks. If we were to build a computer that can perfectly mimic our behavior using pattern recognition would you still say it's not intelligent? Where's the line?
Honestly the first thing that came to mind with this is ultra comfortable mattresses. But clearly there are a ton of other really really useful applications
When thinking about the future I was always just thinking of more/better computers everywhere, but now I can also dream about how comfy our chairs will be :3
6:19 Dr. Károly: And, not only the source code of this project is available, but it is available for you to try on one of the best websites on the entirety of the internet -- Me: Weights & Biases! Dr. Károly: Shadertoy! Me: ...
Doing this with metal 3D printing could probably generate _exceedingly_ safe cars by redirecting impact energy precisely. They might even spring back into shape after a collision!
This could be useful for airbag like things in cars where one side has to be rigid for montage and holding in one place and the other side right next to it has to be flexible and force absorbing.
It looks so beautifully organic, like finger prints or muscle structure or bone structure. I wonder if this can be combined with fluids that can control the compression actuation and actually create a new kind of machine? Using flow and pressure control for example
There are so many possibilities with this, you could have a long rod that when you compress both ends, it bends into a particular shape, like some text or the outline of some kind of object.
Truly, what a time to be alive. This will really open a lot of options for all kinds of useful items to be mass produced in ways that were not possible before.
Maybe this research could help mitigating injuries from car accidents by designing vehicles which dissipate the impact forces even better than current designs do.
5:00 before this part, I was already thinking how this reminded me of like, leather car seats of the 70s/80s or something, IDK why exactly, something about the texture as it crumples? But using this for the internal structure for optimal comfort is obviously a completely new and awesome technology. The closest before now is multimaterial cases such as different stiffnesses of foam or springs
This reminds me of how in F1, you're not allowed to use moving/active aerodynamics, so wings and stuff have to be rigid as per the rules. To test whether a wing is allowed or not, they apply pressure to it to see if it bends. One F1 team used a clever technique with carbon fibres to make a wing only bend in a specific way, and not in another way. It passed the test despite being active aero.
These would make excellent synthetic heart implants. Once we have the necessary materials. Although the support structure could be used in the growing of artificial organs.
Mechanical parts that bend are also known as compliant mechanisms and they have a bunch of advantages and there is a great video on them by veritasium, look it up if you like the topic. I think it's very interesting.
@4:08 : this could be a model for space time and General Relativity, to compete with the stretch fabric and balls that physics instructors currently use.
These spatial algorithms can be included as infill patterns in slicers like Cura or Prusa Slic3r . I am more impressed with the multi axis 3d printer in the beginning that printed without supports
I would love to see this incorporated in the manufacturing of cars. Could save more lives than the current collapsible systems by improving on them. Would likely also make the cars lighter and/or stronger. Just need metal additive manufacturing to catch up with production speeds.
I knew I was on to something printing all those models with only infill and no outer walls or top layers! They had lots of stringing though, probably need to level the bed. lol
Hopefully Cura Slicer can incorporate this as an experimental feature. I can see this having an immediate use for protecting sensitive equipment such as camera lenses or mobile computer equipment.
Loved the video (and all the rest of yours of course)!! I’m always interested in new things about 3D printing! I just need to find an career/engineering career in the field of 3D printing, 3D printer manufacturing!!
Hmm. I play hockey as a goalie and have been thinking I would like to have a neck guard that covers my whole neck but doesn't inhibit movement at all. Wonder if I could make a neck guard like that with this?
very impressive! if the flexibility could be combined with the strength of ABS (which i don't think they are using here), then that would be really wonderful. the colour printing is very nice and potentially very useful and the caustic model is beautiful.
So my dad and I are at a disagreement and I hope it can be answered, do these experiments use a special filament or material? Or is it the same materials you would use for a regular print such as PLA etc.? Thanks in advance Edit: okay so I looked through the paper again and found that it is TPE for anyone else wondering
So awesome to see you touching on 3D printing again. I love the intersection of algorithms and manufacturing!
It was so refreshing for my brain too, loved it! Thank you very much for the kinds words and for dropping by!
I bet you could replicate/enhance some of these results with varishore tpu, trying out the code (seems like it's an IceSL addition) to make semiflex tools without joints would be awesome.
Oh, he got his hands on this stuff himself? Awesome! I thought it was example footage from the authors.
My sentiments entirely. I cannot wait to see what utter madness awaits us in the 3D printing world!
"touching on" -- I see what you did there
I want to praise this channel in a more unique way than the usual. As a scientist I do know the struggle of keep up to the date, expecially when there is some revolution in other fields that do not belongs to you. This channel does in fact generate an easy glance to all that it is new in the form of a pill, there is no need to read for ton of papers, it is a selection that does basically highligth great potential tools in a time effective manner.
I firmy believe that this channel is source of ideas for many scientific groups out there, and it will be more in the future. It is a tool that allows to be keep up to date in a very light way allowing to spend time and redirect efforts easily that one may think.
I do belive that tools like this channel will improve by years the adoption of new methods spreading the news way faster than is usually done. Keep it going, it is an incredible work you do for the entire world, this channel is an optimization tool that will be overlooked, but it is useful for the entire humanity.
You are too kind. Thank you so much for the heartwarming message! 🙏
@@TwoMinutePapers And you're too great! Immense respect, please keep doing what you do.
As a researcher in Biology I wish there was such a good channel in my field.
I'm a Ph.D. student working in AI + CG, and I agree wholeheartedly!
We
Imagine 3d printed protective gear made so the expected area of collision is rigid and tough but the insides are soft and breathable
the video reminded me of the 'crumple zones' you see in cars where certain parts of the vehicle will actually crumple more to absorb impact while the more rigid roll cage stops the occupants from being crushed
You can already do that with two materials - rigid material on the outside, soft on the inside. It's not the best use case for this paper
@@jokinglimitreached1503 but it would be substantially lighter I don't know if you play sports but most protective gear is quite heavy
@@gameodemosthene9252 Is the 3D printed material light? If it's also strong, yes, I would agree 3D gear would be nice
@@jokinglimitreached1503 the futures gonna be crazy man. Imagine durable flexible and protective
KHAKIS
"there were probably a lot of injuries... So clearly, this is very cool!"
-- 2 min papers without context
Wow! This paper really encompasses the phrase: "what a time to be alive!"
So many possibilities!
What I REALLY love about the showcases in this video is that all these could be printed on a consumer $300 printer.
As someone that's been doing 3D printing for years, I love seeing how many ways people are coming up with creative and useful ways to put these little robots with a hot glue gun to use. 🤩
It's even more cool that this solution was found using a handmade algorithm instead of machine learning. Just goes to just how far materials and topology has come in melding math and engineering
Just two more papers down the line, we can 3d print a 3d printer and return the original!
Look at the RepRap project lol. We've been doing exactly that for years. Prusa also has a 3d printer farm to make parts for their own printers. These Papers have been made for like 10 years.
its the goal of the reprap project
you can do that already (almost). ofcorse it can't work for stuff like the nozzle or electrical parts
@@VincentFischer depends if you allow other printing methods than FDM. You probably could print the nozzle but the surface finish is probably bad without manual finishing.
@CanadianOtaku printing stuff like motors, microprocessors and electrical components? In theory probably yes, but thats still a long long way and probably not possible with the current methods (FDM)
The singularity is coming just a few papers down the line 🦾
The singularity has been coming Very Soon™ roughly since 1955.
@@sanjaymatsuda4504 Eh, I wouldn't be so dismissive of this. When AI was first introduced the singularity was just speculation, and we couldn't really know for certain when it would occur. The pace of progression in AI since then is insane, we're achieving things never thought possible before, and we have a much better understanding of how to create and train it. AI can already write basic code, it's only a matter of time until it can write the code of a better AI. I wouldn't be surprised at all if we reached the singularity in the next 20 years or so.
Or it happened a long time ago and we are living in one of the simulations for a paper.
@@commenturthegreat2915 pfft we're no closer to actual learning than we were 50 years ago. When we first made computers, people considered them "thinking machines". No, they just do arithmetic really fast. It turns out we can apply that to pattern matching too. That's all modern AI is. It's not intelligent, unless you think a calculator is intelligent. We still have no idea how to do anything like what a brain does.
@@gracefool This is a different topic - the singularity simply refers to a point where AI becomes capable of creating better AI, and we're definitely closer to having achieved this milestone.
As for your point - I heavily disagree. There is a massive difference between writing conventional algorithms and training AI using machine learning. In the former you're writing direct instructions for the program to follow, meaning it would never improve without your direct intervention. In the latter however, you could write relatively simple instructions which enable the program to continuously improve without you having to change a single line of code. Neural networks and similar learning structures were meant to mimic the complexity of biological computers and enable them to improve over time through minor tweaking, just like our brain does. And lo and behold, it works. The actual functions and connections in machine learning structures may not be the same as our brain, but they definitely achieve the overall effect - learning. This isn't something that was possible before. AI these days can actually be innovative and break the expectations of the people who created it. This is where this starts to get more philosophical than factual - can this be considered intelligent? I find it strange that you did not actually explain what you would consider to have reached that line. I would argue that pattern recognition is basically the same thing that makes us intelligent, be it on a much bigger scale. After all, our brains also consist of networks and firing signals, and could be considered to be just very complex computers. There's nothing inherent that prevents electronic machines from performing the same tasks. If we were to build a computer that can perfectly mimic our behavior using pattern recognition would you still say it's not intelligent? Where's the line?
You are an awesome Human. Thank you for sharing the knowledge as it develops!
You are very kind. Thank you so much! 🙏
That knee is sooo futuristic, imagine a leg with that fiber with metal bones and motors. A true humanoid robot, specially lightweight
I propose carbon fiber instead of metal.
@@davidwuhrer6704 awesome, even better
Honestly the first thing that came to mind with this is ultra comfortable mattresses. But clearly there are a ton of other really really useful applications
hmm, those already exist in a way, the Purple mattress.
@@applepie9806 I know, that partially is why I thought that: This probably can even improve on them
When thinking about the future I was always just thinking of more/better computers everywhere, but now I can also dream about how comfy our chairs will be :3
The perfect sofa, when?
A perfect end to a friday!
What a time to be alive!
This is amazing, I am about to go study at a Higher Vocational Education school to become a 3D print technician :)
So this video is pretty awesome ;)
5:52 wow what a optical effect!
I love how two minute papers does more than just ai algorithms. Anything cool is always welcomed.
6:19 Dr. Károly: And, not only the source code of this project is available, but it is available for you to try on one of the best websites on the entirety of the internet --
Me: Weights & Biases!
Dr. Károly: Shadertoy!
Me: ...
2:00 so hydro dipping something that has been a thing for quite some time
Doing this with metal 3D printing could probably generate _exceedingly_ safe cars by redirecting impact energy precisely. They might even spring back into shape after a collision!
This is one of the best papers I've seen you make a video about. Seems like it'd blow the field of metamaterials wide open!
i could totally see how this could be used to design better crumple points in cars to protect vehicle occupants in collisions
Just imagine the combined findings/techniques from compliant mechanism this and origami
Veritasium has a video about compliant mechanisms which have expanded the field of origami.
@@davidwuhrer6704 Yes that's how I came to know about it too. Fascinating times
Nice, this can sure prove useful in the future! Good that researchers are trying to be imaginative!
This could be useful for airbag like things in cars where one side has to be rigid for montage and holding in one place and the other side right next to it has to be flexible and force absorbing.
Future grandmas gonna 3d lrint sweaters, what a time to be alive!
It looks so beautifully organic, like finger prints or muscle structure or bone structure. I wonder if this can be combined with fluids that can control the compression actuation and actually create a new kind of machine? Using flow and pressure control for example
Lovs to hear your idea in more detail, Juliana Payson
Sometimes you see a thing and you just know that this is the future. Soft, printable robots everywhere.
This is good for coverings on mechanisms with joints that also have hard shells attached to them.
There are so many possibilities with this, you could have a long rod that when you compress both ends, it bends into a particular shape, like some text or the outline of some kind of object.
the field of meta-materials are unbelievable.
Truly, what a time to be alive. This will really open a lot of options for all kinds of useful items to be mass produced in ways that were not possible before.
This is kind of like compliant mechanisms. I love it
Maybe with this they can make robots that don't look like such losers
Maybe this research could help mitigating injuries from car accidents by designing vehicles which dissipate the impact forces even better than current designs do.
I am in love with that hydrographic method
I've never seen a machine doing hydrographics, but it seems perfect for colouring parts!
5:00 before this part, I was already thinking how this reminded me of like, leather car seats of the 70s/80s or something, IDK why exactly, something about the texture as it crumples? But using this for the internal structure for optimal comfort is obviously a completely new and awesome technology. The closest before now is multimaterial cases such as different stiffnesses of foam or springs
As an industrial designer I can see a couple fantastic use cases for this and how it could develop further. Actually intrigued.
WOULD LOVE TO HEAR YOU ELABORATE, TIRO
This is amazing! There are so many applications for this.
Those patterns look a lot like those produced by reaction-diffusion conditions!
This reminds me of how in F1, you're not allowed to use moving/active aerodynamics, so wings and stuff have to be rigid as per the rules. To test whether a wing is allowed or not, they apply pressure to it to see if it bends. One F1 team used a clever technique with carbon fibres to make a wing only bend in a specific way, and not in another way. It passed the test despite being active aero.
i am more of a formula e kind of guy
but its really cool
Assuming I can combine conductive materials this is going to be so useful for my work, amazing work!
The possibilities seem endless.
This can revolutionize compliant mechanisms...
stuff like this makes me more interested in what the future robots will look like
The mattress industry will go cray cray
YES MORE ADDITIVE CONTENT PLEASE!
You always find all the coolest papers.
Auxetics are so cool! I believe they will be everywhere soon and transform our world in ways we cannot yet begin to imagine.
Karoly Zsonai your accent is perfect!
shadertoy is where i get alot of my procedural abstract clips for videos. its great!
If one were to add ceramic to the surface, one could hypothetically 3d print conformal body armor that wouldn't restrict movement at all.
“Research is not only fun and joy.” Agreed. I cried a little when they drowned the cat.
Today our simulation is designing new materials! This is the realest paper yet!
This is going to be a huge gamechanger
I love the background boom effect at the starting of every video
The pliers blew my mind
These would make excellent synthetic heart implants. Once we have the necessary materials. Although the support structure could be used in the growing of artificial organs.
Mechanical parts that bend are also known as compliant mechanisms and they have a bunch of advantages and there is a great video on them by veritasium, look it up if you like the topic. I think it's very interesting.
"The structures showcased here can be used in soft robotics, soft mechanisms, prosthetics..."
Me, with bad knee caps "YEAH! SCIENCE!!"
@4:08 : this could be a model for space time and General Relativity, to compete with the stretch fabric and balls that physics instructors currently use.
I wonder if that kind of knowledge can be used to lessen the computing stress of 3D muscle simulations for physics interactions with game characters.
Blown away by this
Hmm. wonder if this would be useful in tires and suspension systems. How about all terrain tires?
These spatial algorithms can be included as infill patterns in slicers like Cura or Prusa Slic3r . I am more impressed with the multi axis 3d printer in the beginning that printed without supports
This infiller is available in the latest beta of IceSL.
This would be awesome for soft robotics, prosthetics limbs and protective gear
I wish I had found this channel earlier. I love it so many kewl Ideas.
Those flexible object clips makes me want to eat a cake. Idk why.
Very interesting field of science. I think those techniques could also be used for miniaturisation, like nano-machines
thanks for making this video. i started to wonder what i can use this for.. now its something i didnt know i needed. thanks again.
So amazing ! This paper come from my city ! Impressive
I am thinking about a big flat hamburger, my guess is there is already a 2 minute papers video on that?
I would love to see this incorporated in the manufacturing of cars. Could save more lives than the current collapsible systems by improving on them. Would likely also make the cars lighter and/or stronger.
Just need metal additive manufacturing to catch up with production speeds.
this sound like the basis for an amazing fidget tool
this would be great for stop-motion animation!
This just triggered my creativity area holy shit I'm thinking of so many things we could do with this rn
Noa, can you share some examples?
@@ArchKomposer im sorry but i have no idea what I was thinking of 11 months ago
What a time to be alive!
There are some awesome puzzles to be made here. :)
I can't wait to have furniture that is perfectly cubic yet conforms to my body shape as soon as I sit down.
This is actually super usefull in soft robotics !
More people should know and use shadertoy! ;)
HAHAHAHA YES THIS UNCLE DOES 3D PRINTING STYLE VIDS!!!! TYSM
I knew I was on to something printing all those models with only infill and no outer walls or top layers!
They had lots of stringing though, probably need to level the bed. lol
Ow wow Some new news about 3D printing... This is amazing and Exciting
Can't wait to 3D print me some couches!
Wow! It must've taken a long time to render the hand simulations and the fiber structures simultaneously
I hope you are joking.
@@arunavaghatak6281 Yes lol
Loved it!!! More like this too
Hopefully Cura Slicer can incorporate this as an experimental feature. I can see this having an immediate use for protecting sensitive equipment such as camera lenses or mobile computer equipment.
It is available in the latest beta of IceSL
5:00 i also wish i could be rigid but still create a comfy sensation
Omg it's on shadertoy!
Loved the video (and all the rest of yours of course)!! I’m always interested in new things about 3D printing! I just need to find an career/engineering career in the field of 3D printing, 3D printer manufacturing!!
Hmm. I play hockey as a goalie and have been thinking I would like to have a neck guard that covers my whole neck but doesn't inhibit movement at all.
Wonder if I could make a neck guard like that with this?
very impressive! if the flexibility could be combined with the strength of ABS (which i don't think they are using here), then that would be really wonderful. the colour printing is very nice and potentially very useful and the caustic model is beautiful.
I believe that they are using a material like TPU, flexible and pretty strong.
Dual material is definitely a thing. Hopefully they'll try this technique with mixed materials soon! Imagine all the possibilities! :)
@@mosquitobrainbutcool Thermoplastic polyurethane looks interesting. not sure how my old wanhao i3 will cope with it, but i shall check.
@@BRUXXUS dual (and more) materials is exciting and as it has been around for some time now, hopefully it is reasonable reliable.
Has anyone got their hands on an example model? I really wanted to try it out at home.
Wondering the same. 🤔
These patterns remind me of fingerprints. I wonder if this is why our fingerprints look the way they do!
Those structures remind me of finger prints. I wonder if there is a connection?
A realy high quality product (a vr hip tracker) was made with 3d printing
Could you please make a video about 3D bio-printing :)
Now to make everything even more disposable.
Fantastic video! But, does anyone know how to actually use the demo, I'm confused
So my dad and I are at a disagreement and I hope it can be answered, do these experiments use a special filament or material? Or is it the same materials you would use for a regular print such as PLA etc.? Thanks in advance
Edit: okay so I looked through the paper again and found that it is TPE for anyone else wondering