GAME OVER - A.I. Designs CRAZY New ROCKET Engine
HTML-код
- Опубликовано: 30 апр 2023
- New alloys, additive manufacturing and AI have come up with a drastic new Aerospike rocket! Will this be the engine of the future?
Sources & Credits:
• How Relativity Autonom... (Relativity Space)
• Simulating Supersonic ... (RDE)
• Pangea Aerospace aeros... (Pangea Aerospace)
www.nasa.gov/centers/marshall...
www.nasa.gov/directorates/spa...
scitechdaily.com/rotating-det...
www.eos.info/en/presscenter/p...
www.fabbaloo.com/news/design-... 
Наука
The crucial part of this: It needs to be tested. Right now it's just a piece of art.
indeed
It's probably just bs. This channel makes a lot of clickbait titles like this it's just cheap advertisment
You should want this to fail. If you can have a computer design and build new functional products, the lives of men will suffer greatly. We are being phased out. Not sure why people aren't seeing this.
Apparently they are building one on small scale ,be interesting to see the results
@@jaba8625 or the lower barrier of entry of high quality and innovative designs will create a higher standard of living for more people across the world
Very organic, reminds me of different alien species in startrek that had organic ships.
Those organic designs in science fiction are based on insect hives like those of ants, termites, wasps, bees, etc.
That was the first thing I thought of too. Right down to the same Star Trek episode.
Reminds me of Aliens.
Definitely looks like an alien artifact.
H.R Geiger....
That AI designed engine looks like it would be absurdly difficult to do quality assurance testing on.
What makes you say that? Do you not think the rocket scientists who guided the AI in the design knew what they were doing?
@@Sajuukthey most assuredly didnt, ais are incredibly complex machines that are imcomprehensibly complex preforming billions of calculations a second.
@@Okratron-rr8we they make the parameters but they couldnt communicate how nor why the ai came up with said design. I dont doubt their qualification, however not even the leading scientist in ai models can do that so i doubt they can. People are jeeded to guide them and make sure it isnt being confidently incorrect but seeing into the processes that drive an ai to its conclusion isnt a possiblity yet
@@williamandrews6615 The concept of AI technology often seems like something out of science fiction, but in reality, we haven't quite reached that level yet. While it's true that reality can sometimes be stranger than fiction, I believe we are still in the early stages of fully comprehending AI. Personally, I have only engaged in minimal work with AI, particularly through Stable Diffusion and LLM's, which has already given me a modest understanding of their creation and function. don't forget that we're dealing with literal rocket science here lol.
@@Okratron-rr8we fair loll
I remember helping design that copper thrust engine about 10 years ago. It looks nice in this video but there was no mention that a big problem was having to cover this engine with a thermal restraint material and then attach it to a nozzle that is yet a different material. Once you do that where everything has variable thermal expansion you get all kinds of cracks and breaks. Oh, and 3D printing a copper alloy is not a reliable for smooth surfaces, causing lots of leaks. Still, good video, good job.
I think the A.I. used to model that last engine design was trained on the art of H R Giger!
That is interesting! I think there will be a point where they can just trial run/revamp until the CAD algorithms can produce a robust design.
Or Gaudi
It looks rather organic
Or Alien if you like , which was designed by Giger 👾
Hahaha I was about to make the same comment, it certainly looks like Gigers work.
Whether it is developing F1 suspension components, or Airbus A320 passenger cabin partitions, or drone frames, AI designs tend to look "organic".
There's a reason for that. Nature has had a few billion years to tinker with various structural designs, and more efficient designs tend to be more successful.
Yeah wow, it gives that mech organic scifi style of Giger a whole new dose of disbelief suspension
The AI Engine looks like it evolved in nature. love it!!
Makes sense... Both living things and this engine were created through numerous iterations and selection pressure. The latter just took place in an AI-driven design simulation.
"looks like" yet, just wait until it had time to simulate more than 4.5 billion years :)
Which is precisely what AI does. It's not thinking, it's evolving, everyone gets that wrong.
@@ErLeuchten it's different though, natural selection is a very slow process
@@ObeyNoLies The most powerful models are definitely doing some of both. Their thoughts just don't happen the same way as our thoughts and they are not, as far as we can tell, sapient beings.
The fact that AI can design something like that is both fascinating and terrifying at the same time
how is it terrifying
Because we will blow up when they try and launch this thing
@@superwhine Im glad u can make fun of it and feel better. Because even if this engine design is terrible, there will soon be one no human would have ever thought of that works better than anything we thought possible. And eventually after that some random AI will just do whatever someone made it prioritize, but it will do it so well it will be unstopable and evade every attempt to stop it before we even actually try, and eventually wipe out the human race casually because our attacks were slightly affecting its efficiancy by a few thousandths of a percent. But make jokes now. While you still can...
@@cHAOs9 Someone's been watching too many movies.
@somerandomguy4240 There's a reason people make movies like that.
The ONLY thing that enables humans to be at the top of the food chain, and kill all the other animals we want, is that we are smarter than them. We're going to design something smarter than us. And because we're so used to controlling everything we can't even comprehend how uncontrollable something smarter than us would be.
If you better understood how AI is made, you would understand how we dont design its thought process AT ALL. We design what its brain is made out of, show it a ton of examples of things, and then expect it to make decisions about new things. WE ARE NOT PROGRAMMING HOW IT THINKS. These are NOT the normal computer programs that you know. They are litterally virtual brains.
I know it sounds paranoid anytime anyone worries about a global disaster. But this is like building new viruses outdoors in time square. There are ZERO protections in place for this. It WILL go horribly wrong and the first time it does it will probably be too late.
Rocketdyne was using leftover J2 engine turbopumps on the Linear Aerospike. Lockheed Skunk Works was building the prototype at Rockwell North American's Palmdale Site 10, Building 704, East High Bay. I knew a lot of the people working the program since North American Aero shared the building. A couple of years later I worked for the Skunk Works and stayed employed with them for 20 years after having worked 28 years for Rockwell/Boeing North American.
I love your videos, keep up the good work.
this is the direction that I've seen we need to go for several years now. the combination of ML/AI quick iteration modeling for designs that we haven't thought of, in combination with additive 3D manufacturing. it's finally reaching the point of practical real-world commercial viability. i've recently seen the NASA development of the RDE with center toroidal aerospike using Nikola Tesla's design for a backflow preventing valve without moving parts, which I think you showed here. It seems the company you showcased with this new design is incorporating all of these ideas. the same tech I think is also going to be useful for developing the new materials that we need to build with.and this extends to everything we make, on earth and off.
Yes, what could go wrong with a technology like this, what we did not understand, in the hand of the greediest corporation on this planet? I mean the tech is good, and powerful, so lets use it to create more greed and train them to be greedy.
This engine looks cool tho.
I do not believe we understand fluid dynamics to the resolution needed for AI to design something like this and for it to actually work. If we push to throw AI into applications like this right away we may well set back AI use a hundred years because we will just think it's a failure, when in reality it's our own scientific understanding that's at it's limits.
I just hope we can understand it* and be able to study it* rather than accepting and moving on. I don't want to return to "magic"
@@TK-xn3uf our main issue with fluid dynamics is that we cannot resolve the equations that describe this phenomenon (specifically the Navier-Stokes equation), not that they are incomplete (at least not wildly). Having an AI that takes the equations and constraints as input and resolves everything would be exactly what we need
@@TK-xn3uf its called machine learning and neural networks
Curious to see the testing results... please do a follow up video on it
I've been following stuff like this and it's really interesting. AI always gets confused at the physical realm. Though their designs are amazing.
Titans Of CNC covered something like this a few months back with a titanium 3d printed part that was designed by AI. Def worth a look at industry development.
1:00 small correction - the main reason for multi-stage rockets is that having a heavy high thrust machine push a lighter low thrust machine into a suborbital trajectory that the lighter low thrust machine then pushes into orbit is WAY way more effective than trying to do it all with a single stage. SSTO is itself only BARELY possible from Earth and not worth it for 99.99% of all space missions. Engine efficiency is only a small part of why it's not worth it though, and depending on your needs, there's nothing stopping you from making an SSTO with conventional engines. You just need to account for the inefficiency of the system. Aerospikes already solve the engine bell problem though, if efficiency matters, and they're baked into nearly all SSTO proposals. But again, for nearly every type of mission, you're better off with just a multistage rocket.
AI engine designs might look cool but as every rocket company says over and over again: engine designs are a dime a dozen. They in and of themselves aren't the main concern. It's manufacture and execution that matters way more, as well as choosing designs that best suit your specific needs and company goals. So unless this was design with intention behind it to achieve a company's specific goals, it's going to remain a paper engine forever. Cute though.
The AI design is absolutely stunning!
looks organic
Who cares what it looks like. Will it work?
Looks like a work of H R Giger
real life sci fi
Looks like something Engineers from Prometheus would design !
That AI generated engine design looks amazing....like something out of an Alien movie. One can only marvel at the calculations that went on to generate that design!
They say the calculations were out of this world
AI is an alien.
@@hakimmohamad6216 That's quite possible I guess! I wonder if you could ask AI to supply the technical notes underpinning the design?
Art work looks good doesn't it.
Wait until AI designs and fabricates engines with 3D printing, then we really will only be able to marvel because no human has seen the inside of one.. not sure if funny or scary.
I got to see this in person at IMTS last year. It was truly breathtaking to look at!
That design made the skin on my scalp crawl. That means it would probably work perfectly - it looks more like a living functioning organ that it does a rocket engine. I think that AI is going to quickly help us solve a myriad of engineering hurdles that have thus far stymied us all.
It has that potential .. but it will be restricted is my fear. I have no doubt we have the capability for unlimited free energy, for example, but that has been controlled as well since they locked up all of Nicolai Tesla's files over a hundred years ago. But my hope is AI can help us reach the stars possibly even before I die, and that it will greatly extend our lifespans. I believe Humanity in general is short-sighted when it comes to appreciating our potential, even in the short term future. AI changes everything. If there is a way to achieve FTL travel, for example, it can find a way. This is fascinating and terrifying at the same time. I often wonder how many times our species has gone through this process. I personally don't feel this is the first time. Our myths, ancient ruins etc. tell a long and glorious story. Truth is stranger than fiction. I'm an Agnostic, but as the Bible points out .. there is truly nothing new under the Sun, in a scientific/philosophical sense. We tend to over-exaggerate our significance as an infinitismally small and insignificant species in this massively large and ancient Universe, based on the numbers, in my humble opinion.
Remember today's, "Traditional" AI, can only take ALREADY EXISTING information and stitch it together, if it deems combinable from some points of view.....
All outcomes tested in an instant for the perfect design.
And then kill us all...
I'm afraid AI will solve our reason for being alive, at all. 😅
Everybody wants to do single stage to space, which is easy peasy, relatively. But to realize a singe stage to ORBIT you have to have not only a nozzle that is efficient enough throughout the pressure drop, but a VERY light stage, which is problematical on several levels. One advantage is that you can recover a single stage worth of boosters for reuse, which is less costly. I don't see this happening any time soon though.
That rotational detonation engine blows my mind frankly. You get a lot more energy taking the fuel to very high order burn, but the shock loads must be FIERCE, and I would think over the long term they would be very difficult to design for. It sure is an interesting idea, when and if (probably when) they get there. I would expect that maybe materials better suited to the task may have to be designed, no matter the implementation. We will see.
Aerospikes have been around for a long time. The problems with sticking something in the middle of a blow torch from hell is obvious, again, materials and VERY effective cooling techniques are needed. The problem seems that it might be a bit worse than keeping an external nozzle intact. At least there one face is allowed to augment cooling with air flow and/or radiation, ultimately, to space. I don't know how much that helps in the total equation, but I expect it is not completely without effect. Not a thermal guy, haven't run the numbers, but just a guess. Maybe interior nozzle flow swamps all of that, and just liquid cooling and material, ceramics I expect, are the only real drivers. Just a few brain farts off the top of the ole pin head, probably not worth reading.
AI would resolve the aerospike. AI combine with data analysis and future quantic computer would be revolutionary. Engineers are very close to create quantique computer on super small chip.
one engine to space inefficient, you get there and then what. all fuel consumed to get there. travel slower than the voyager i guess 😅
Fascinating design and probably can be used to gain a lot of insight. The biggest issue is manufacturing this. 3D printing just doesn't result in the same quality as machining or casting
This reminds me of Sunny Emmerich talking about generational adversarial design using competitive neural net simulations to design her LACE/Ramjet RLV over a decade ago. She quite literally predicted this.
It's wild how the boundaries between organic and mechanical are slowly blurring
I am pretty sure whatever boundaries you are referring to are a complete fabrication on the minds' part.
Couldn't agree more! Wow.
That is interesting, but I’m not sure this is a positive thing
I find an ethereal beauty in it.
If there are '4 replies', how come you're only showing me 3, RUclips? #censorship
Anyways...
Agree. Understand that the flagellum has been touted by proponents of Irreducible Complexity as a microscopic electric field motor. The boundaries were never empirically distinct, they were only distinct in our thinking.
This is ultimately what ai should be used for! Not social engineering, hardcore physical engineering. 👏🏼🙏🏼🙌
AI understanding human psychology and behavioristics is, in my view, the greatest danger. In the Terminator movies, Skynet understood technology and tactics brilliantly, but not human creativity, adaptivity, and resiliency.
@@HuntingTarg You already have that, they're usually called targeted advertisement. It used to work via cookies and tracing internet usage but nowadays it's an AI that analyses your behaviour around some website or more commonly google's or meta's ecosystem. They can sometimes know your needs better than you do.
And our economy should be geared towards real engineering instead of social engineering
Both use it for both
@@prime12602 social engineering is good when getting people to see things from a bigger perspective. Closing peoples perspectives just to keep them blind to reality is in fact helping nobody. Coexisting with nature in order to live longer healthier more sustainable lives would be an ETHICAL form of social engineering. Making them fear going outside to achieve similar results would be UNETHICAL. You will have more willing volunteers the willing slaves. 👏🏼🤦🏼♂️
It would be truly fascinating to see this concept being tested if it ever gets to the testing phase. AI has already designed some remarkably advanced and game changing designs.
Very cool design! But I wonder how it would hold up structurally overtime. Hopefully the nice design lowers the stress?
Small correction. Rocket staging is never about nozzle efficiency, it is about dropping some of the mass as it is no-longer needed. Yes upper stages have vacuum optimized engines because they can, but that isn't the reason for staging.
He did say “partially” and he would be correct.
Calculation of payload (staged or otherwise) requires calculation of engine performance…requires determining quantities such as pressure, temperature, density and gas velocity in the combustion chamber, at the throat, and at the nozzle exit.
@@y0uCantHandle I would say no, there is not a single rocket in existence that was staged rather than single stage due to nozzle performance considerations. Yes, maybe single stage to orbit would be marginally easier if nozzles could be optimized for any air pressure, but I doubt that we would see more single stage to orbit attempts. Aerospikes if ever successful would be great but I very much doubt they would have any effect on staging. They would simply add extra performance and possibly make return vehicles more efficient (Starships second stage for example needs a mix of nozzle sizes because it needs to land)
Staging is the main reason we don't use aerospikes though. Just 2 different nozzles are more than 95% effective. No need to even look for anything better. Unless you go single stage.
@@TimothyWhiteheadzm you are punching a strawman. That is not what he said at all. He said it is a partial consideration, and he is correct. It is definitely a consideration when determining how many stages are required.
@@TimothyWhiteheadzm To flip what you said… there isn’t a multi-stage orbital rocket in existence that ignores nozzle performance.
Keep in mind no single stage to orbit rockets have ever been flown. There are some concepts though that can use the same engine for all altitudes, which is a problem for traditional engines with a bell-shaped nozzle. Depending on the atmospheric pressure, different bell shapes are required. Engines designed to operate in a vacuum have large bells, allowing the exhaust gasses to expand to near vacuum pressures, thereby raising efficiency. Due to an effect known as Flow separation, using a vacuum bell in atmosphere would have disastrous consequences for the engine. Engines designed to fire in atmosphere therefore have to shorten the nozzle, only expanding the gasses to atmospheric pressure.
The efficiency losses due to the smaller bell are usually mitigated via staging, as upper stage engines such as the Rocketdyne J-2 do not have to fire until atmospheric pressure is negligible, and can therefore use the larger bell.
Be interesting to see where A.I. leads. For myself fluidflow simulation we use at work has been flaky and next to useless for decades . Only now - and by using supercomuter tie in does it produce realistic answers for simple situations. So I would like to know how this accomplishes something as complex as complete rocket engine in minutes. Where did it get computing power from that we don't have? (lab facility where I work is a world hub for supercomputers).
Same situation with self drive cars. Even 10 years ago they could produce some impressive looking road skills - at least by first impressions - and might not even crash after a few hours of driving - depending on traffic conditions etc. A decade on - they all looks marginally more impressive - but it is still likely to crash and injure/ kill you by the end of a day's driving. After some initial *wow* factor results they plod on and develop slowly - but in reality we are not that much closer to self drive cars than we were 10 years back.
As you see I'm a skeptic. I think AI has all the prospects of occassionally producing something genuinely useful an dinteresting and of being effective mimic of apparent intelligence and creatiivity - without being fundamentally either .
@Karl with a K I think that any A.I. of general utilty intelligence should be able to navigate the roads long before it is declared of superhuman intelligence in the more challenging intellectual environments. Even in the 1960s they had belief human A.I just round corner. In the 1980s/90s there was big buzz on neural networks being next big thing. In the end working 40 years in high end CAD sims I've seen change but only tiny fraction of what I supposed. Given my past experience I'm not believing that *the* big AI breakthrough is imminent.
Humans: Aliens must be super advanced and smart to design their spaceships.
Aliens: Siri, make me a spaceship.
Hey AI, design an engine.
AI: reads Lovecraft.
AI: prints the above.
I've been thinking about this. A.I. and 3d printing will allow us to create some truly amazing technologies.
Yeah. If I remember correctly, Formula 1 engine blocks are already manufactured in this manner, or at least the mold is 3D printed and then the engine block is cast into that mold for better reliability. Those engines are working in extreme condiitons and I think rocket engineers should borrow some homework. :)
A.I is already designing many parts for plains and cars
By the time this will happen, elon musk will already have bases on mars bro.... My guess 50-100years till this will really going to change rocket industry.
But it it heavy, it didnt even reach the space.
@@t0lex14 I don’t think you appreciate how fast technologies and new methods of production will advance once artificial intelligent and quantum computing gets established. Artificial intelligence will beget the next generation artificial intelligence very quickly and the rate of progress will be exponential. 20 years will see enormous changes to our world. The trouble is it will be disastrous for our freedom. Those who control ai will hold all the power and the rest (99%) will be totally disenfranchised. Far from being bright, the future looks very dark to me. I have no faith in the humanity of those at the top to show anything other than total disregard for anyone else’s rights.
I think the engine is awesome! But I think AI is scary as Hell, and should only be used with an incredible overabundance of precaution, not to mention controls out the wazoo.
That ship has sailed.⛵️
"AI should only be used" which part of AI? lol.
Simulation will make things safer
yall watch too much scifi
Bad things will need to happen first and then controls will only happen in response.
Oh, a channel that doesn't treat its viewers like the type of people who debate if Marvel movies are realistic. Subscribed.
Would be nice to hear a more deep analysis of this pattern.
The f 117 was also designed by computer algorithms and suddenly it had some issues regarding aerodynamics, maybe this piece of alloys would have some issues on saving fuel, I dont know really.
This is just the next step in engineering. If I remember correctly Mazda already used AI Design for some cars so it’s just a matter of time before CADGPT becomes a thing.
more like an optimising function added to cad, but i think it'll take on the order of a decade before a polished program like that gets made
thats why their cars are crap. Except the Miata, which is old school.
"hasn't yet been tested." Test it. That was easy wasn't it. Looks remarkable. It would indeed be unimaginable that a human could design that.
You're not serious 😒 Even A.I is human designed, lot of technology is not even showcased in the open to people. What you usually see and is been presented as new is usually quite old and only reason why they show it is cuz they got something better.
LOL have you ever met a concept artist? Bro...
Wish to see the testing as soon as possible
Art: my dad was an actual computer and rocket scientist. I have a pair of cufflinks that were made from gold discs that were a part of a very early 1950’s computer system. I worked on the design and construction of wafer fab facilities. At one point I became fascinated by an integrated circuit design. I talked them into blowing up the image a thousand times. The resulting image looked like a wild design for a new city. We messed with the colors a bit and had a very large image plotted. I mounted the print on a rigid back and then carefully gold leafed all the parts that were actually gold. Then, using my surfboard fetish, glassed the surface. It is 36” x 48” and a real conversation piece. Science and engineering become art.
I certainly see the new organic nozzle design as a base for a glass table. You may be looking at $50,000 for one with documented provenance.
Remember how we all felt when we saw Atlas the Boston Dynamics robot? Let me summarize: Awe and Fear. Well, that's the same feeling I get when I look at this engine.
As everyone else has said, to me it looks alien and organic. Reminds me of something from Halo or war of the worlds.
Organic but not necessarily alien, a lot of organs and even plants would look very strange in a layered cross-section, as this engine is presented. There are a lot of surfaces shown that aren't really there, they only exist because of the cut, really everything is intact in 360 degrees as walls rather than the folds of the presentation effect. It's like cutting an onion into a flower.
Interesting to see you cook with the coal bed set lower with vents wide open. With my cooks I tend to bring the bed up for a quicker cook. I will try your technique
Amazing concepts. I'd love to see these tested and wish them success. Meanwhile, does anyone else think it should power a spacecraft from "Alien?" :)
The important thing is the engine looks freakin AWESOME. ❤
You know they threw some HR Geiger into the algorithm. They must have.
would like to see this thing working, good in theory
This is the future of manufacturing and it's gonna change a lot. The same thing has occurred in gas turbines, where these 3D printed combustion machines have all of their liquid channels and simple "parts" printed all at once in one body. I think the gas turbine went from like 600 parts to 85 or something like that.
I wanna see the test run tho, looks wild, maybe Ai will sort out more advanced space travel
Imagine the capabilities we can have once this is mature, it would be endless.
maybe one day we reach moon again...just mybe but hope is there.
we will get to a point were we will say build the US enterprise and it will do it. automatic fabrication plants will contruct everything.
I love how new manufacturing methods allow for drastically new designs. I remember reading about the FEM(finite element method, its the basis for these "organic designs") 10 years ago, but at the time I still didn't understand all the spatial limitations of traditional manufacturing methods(CNC, casting, etc). 3D printing is a wonderful idea and I can only hope it achieves the strength of other methods while continuing to fall in cost. A new era of design is upon us.
The next manufacturing revolution will be using nanobots or something alike, in some hundreds of years, but this is a great middle step!
Not hundreds of years, more like in the next 5-10 years from now. A.I. is going to accelerate technology at a uncomputable rate. A.I. will take over the world in the next 20 years. Read your Bible, get saved. Jesus is the only way to Heaven. It's real, the end times are here.
The AI engine design is so organic and beautiful. It reminds me of some commercial jet designs AI created that had a very organic quality. Humans think in such gridded, boxy ways. Nature and AI understands how energy flows along elegant curves.
Make it design a camera that can record live feeds of a single rocket leaving earth for the first time in history....
Very cool. Algorithmic design seems to have lots of promise. Probably lots of problems with manufacturing and testing, especially at scale. Great video!
I think I'd want to see the test results before judging the design. If the AI design was weighted towards producing an alien looking design then that's just the sort of thing you'd get. If it was weighted towards producing an efficient design but still came up with this then that's quite interesting. If the engine actually works then that's news worthy!
The test is the thing; how it looks will be reevaluated in light of the test results.
It’s not “Alien” it’s “organic” which makes sense, organic designs are extremely efficient. Mother Nature produces nearly countless iterations of a design over extremely long periods of time to come up with something optimal. AI does this but much much faster.
@@ecbrown6151 It's organic if it goes through those millions of years of evolution where less efficient designs are weeded out. Here we don't know what the testing algorithms were, it could just be the prettiest design that still works, it might be the looking for the most alien design that still works. One thing is fairly certain though, if the testing algorithm was going for pure efficiency then you wouldn't have the weird "alien" designs on the outer casing.
Out of curiosity, does the engine actually work? If it doesn't then the genetic algorithm should really of killed it off immediately. If it does work, is it more efficient?
@@mikebikekite1 Not true, organic means living, life like or resembling a natural object. Like an organic house etc.. where the designs are smooth flowing
@@hihosh1 Yep, I can accept that but, at the end of the day, you're just arguing semantics. The video talks about a great new rocket engine that's AI designed and look's biological. My main point is the AI design will reflect whatever design parameters are put into the AI. The fact there's no real testing mentioned or comparison with existing rocket designs means the engine is just being evaluated on looks. So if the AI is set to reward designs that look like they came from the film Alien then that's what the end result will look like.
I’ve had my own arrow spike engines since the early 90s, STO would certainly be a game changer.
2:19 Love the Skunk Works emblem on the tail fins.
I'm excited to see how the AI design, tests could be awesome, or it could be really restrictive.
If Aliens ever encounter an Earth probe, we want them to marvel at the organic looking rocket engines, wondering if we simply grow them.
Topological optimisation really does wonders. I am curious how the future designs will evolve around AI and geometrical perfection. Not to mention the production methods that will emerge to put these designs into use, simple 3D printing will no longer cut it.
The combination of AI and 3D printing it just amazing and It’s so pleasing to the eyes because tap’s to the design master Nature itself… Every design has this beautiful organic feeling…
Crazy how organic the design looks. Our creator would make some sick rockets!
Check out the history (some being very hard to track down) of Viktor Schauberger! 👀
I have to agree with Elon Musk who is afraid that we will might devolve technologically before we become a space fairing species, especially if AI takes over our daily basic functions. People will forget how to build things.
I had been thinking why they not use A.I. for things like this... like new medicines, treatments, battery chemistry etc. Also: Isaac Asimov has a novel where an A.I. designs a working FTL drive :)
I would like to get ahold of an A.I. to ask it about making an FTL.
@@tonymc9102 I bet that putting a combo of a powerful language model and an engineering design AI on the job would get us somewhere faster than the pace we're going. The math checks out for the Alcubierre drive, the question is if you can actually turn it off again before it gets stuck in FTL or overheats catastrophically. There is also the not-so-minor question of where and how, precisely, one might be able to acquire and contain hundreds of kilograms of antimatter without bankrupting the entire planet and/or vaporizing in a fireball that makes the Tsar Bomba look like a firecracker, but if they can help crack fusion reactors and conventional rocketry to bootstrap orbital manufacturing infrastructure... This would not be a trivial task, even for an actual superintelligence, but once the groundwork is laid, it is not impossible.
Are you talking about hyperganic? I'm using it right now and it's awesome. Now only waiting for a combination of both GPT and hyperganic. It's been out a year now
the desing is slic and organic at the same time. Nice.
Fascinating video! AI will design us a Warp Drive. We need more and more AI. AI is The Way. No safeguards, no slowdowns. No Fear. Feed the Beast and let it Grow !!
actually theres already a concept for a warp drive. its called alcubierre drive.
I see you subscribe to Roko's basilisk
god forbid a AI is ever hooked up to a 3D printer directly without supervision
Dr. Evil thinks he is a great supervisor.
And there are many evil doctors among us.
And they all have access to AI bots, 3D printers and chemicals... Scary!
It's probably being done as we speak.
@@filonin2 It was nice to meet you :I Farewell
any chance AI could solve the heat distribution problem with the Aerospike engine?
Would be really nice to have a functioning single-stage-to-orbit ship!
"Or maybe A.I. will destroy us in the end." I'm fine with that outcome as well 😂
All that "AI made something" is such a click bait wich Im kinda tired already. Its not even an AI in any way. It is a software programm with good calculation power. It is a tool. At this point you can consider a wrench as an AI. They give it an input date and required output data. I wonder if any time soon I'll buy a tooth brash and the marking on it will also say that it is somewhat AI.
That’s about as alien looking as it gets. Crazy. Not sure how I feel about AI yet but this is cool
I got to physically touch the thing in the thumbnail. It was printed by EOS and they had it on display in Chicago at IMTS last year. Pretty interesting stuff.
How are they going to smooth the prints? The imperfections of 3d printing are still kinda high and those tiny fluctuations on the surface I would imagine would add turbulence to the flows.
I’ve been interested in science my entire life and during my education I often heard the saying that modern science knowledge would double every 10 years. That was mid seventies.
Here we are 50 years later and many advances that I envisioned have been fulfilled, particularly in Computer science, machining, health, and materials.
I think we are at the point that we have all the components to accomplish significant advancement in technology. In the past the big thinkers were limited by the materials and techniques to achieve significant changes or advancement.
I think we are in need of novel ideas and big thinkers, using modern technology to achieve great new things and real advancement.
Maybe it’s time to revisit some failed technology and design that was limited because of older technology.
I enjoy your video’s and content.
This makes those old Sci-Fi movies we all grew up with turn on a light that says to me "time travel is real"
"[t]his engine was designed beyond human imagination"
H. R. Giger: Hold my beer.
looking forward to testing results of the AI designs.
Ai designs of everything for optimum efficiency is exactly what I’ve been waiting to see
RDE are insane the super tiny ones in that lab put off enough thrust to flex the door on the other side of the room
impressive, i imagine a desing for water motor (hidrogen) really efficient and ecologic for the cars...
How are they going to mill and polish the internal surfaces to actually get to a testing phase is my question. Those inner surfaces are so rough.
What was the algorithm that created the hyperganic nozzle?
It looks nice I'll give it that but in order to step further out from the cradle then reliability and fixability are going to be important.
Interesting how organic it looks. Whats more interesting is how many sci fi fantasy worlds depict and predict futuristic and alien designs to also be very organic appearing.
Good algo, very close to one of my old designs ❤️
It’s all about how well it works out in the real world. There have been numerous radical designs like the flying wing that seemed so “intuitive” and “groundbreaking” in their design stages only to ultimately prove to be impractical. Or Fuller domes and other predicted “game changing” innovations that ended up only showing promise in specific use cases.
~10 years ago I remember there was a wind turbine generated through simulated evolution on computers in a similar manner. The prototype exceeded efficiency over conventional turbines when tested, but I don't think the idea ever really took off. The overall process was similar to this rocket, but just with slower computers.
when is that cooper alloy grcop 42 gonna be available for engines.
My company provides the HIPs for different aerospace companies researching these rockets. Fascinating field.
This thing having kind of natural looking lines and shapes reminds me of that car that is also designed by AI and also features parts and shapes that look like stuff you would come across in nature since, appearantly, nature just knows how to design stuff well, I mean it had hundreds of millions of years to basically do trial and error so yeah it's propably even smarter than most of us think
AI design as stunning but like you said it also has to be designed for manufacturability, costs and service not just performance. On the RDE... I know of a couple testbed engines and like aerospikes they crop back up every so often. I am not aware of any that successfully transitioned off the testbed are you? Interestingly they are back under development again in DoD.
Lets see it run.
WOW!!! I have dreamt up all sorts of configurations in my head and the idea resembles that AI-designed piece at the end of the video. The idea, which I am unable to put my finger on properly or even understand, has to do with channeling the gases at different temperatures though differente configurations that would maximize the exchange of energy by modifying the pressure of the gases as they travel in one direction (toward the exit say) through a maze that builds up the pressure as it travels, sort of like a capacitor, or like a nuclear bomb that does not explode but where the would-be explosion is used to feed a longer running expansion of the gases in a controled fashion. I don't know if I am making sense, but that AI-designed thing looks like what I aiamgine it would be, so that essentially, one would get huge forces coming out of a very small generator, sort of like the speakers on smart phones, where if you put your finger on the exit hole for the sound, you will feel the forrce and the corresponding pain even though the generator is miniature (same with sith sand-blasters, water-pressure cutters, and so on.
You seem to have thought about this more persistently and deeply than most. You seem to be saying that an optimal design would concentrate pressure and flow at the combustion outlet(s) to be just below the mechanical and thermal tolerances of the alloy, so that maximum performance is realized in the design.
What another comment has helped me see is that Machine Learning & Optimization (which isn't true AI) can solve problems just like this, where there are many complexities and interrelations that lead to designs that are intricate and elegant - things that are both art and engineering.
Good video about Rooockets, are you by chance related to the Dry Eyes guy or the principal from Farris Bulers Day off?
We had a good run
I wish you had mentioned what company was working with the AI design motor so we could follow up with it in the future
The organic approach is definitely the way forward, it's the most natural way fluids and gasses are steered. Humans are too structural in everything we design, so I do believe that with the help of AI we can overcome that step.
That's things giving me 100% xenomorph conatations
Looks Alien like some type of flower bulb -- very organic. Extreeeeemly interested in the testing results.
I'm having my A.I. cereal and A I. coffee while watching this.
Oh... I have to feed my A.I. cat.
It’s crazy how organic it looks
I think the emergence and our eventual reliance on AI to solve many of our most complex problems is rooted in the inability of 99.9% of people to consider multiple different overlapping constraints and choosing an optimized solution.. Most people in most industries have severe difficulty considering and balancing more than say 3 constraints at once to arrive at an optimized solution. Take for example giving someone driving directions, most people will get confused if the directions involve more than 3 steps, 4 at best. Have multiple steps and optional directions they could take and everyone gets lost.
Would be interesting if part of the optimization process for generating this design involved accurate simulation of its manufacture including likely defects, that way it could balance theoretical efficiency against a more realistic measure of reliability (and potentially cost to manufacture).. i dont see any way of manufacturing this that isnt some form of 3d printing. And maybe thats okay.. but even if it is, some 3d prints are more durable and error prone than others.
It looks like it has come from a alien space ship, very cool...! I hope it works amazingly!
Looking forward for a future of everyday objects to look cool and less boring 👍😊
what happened to using the propulsion system we used in Flight of the Navigator? Is this no longer a viable option