it's overfuck this timeline
Its not even close to over. Literally no new movies have come out that are even worth mentioning in the past 5 years. >Nobody wants to watch 2 hour movies anynway. >No body wants to buy dvds. >No body wants to rent movies. >No body wants to go to the theater. >No body wants to watch generated contentOn the other hand, people want to create with their own 2 hands. This is the difference between playing a sport, either professionally or recreationally and watching a sport on TV. Two different things.
>>1006395What's the point if they won't be able to monetize it? Not like it will replace any industry job in that case
>>1006399nobody even wants long form, dude. Short form on the other hand is unmarketable
>>1006395>completely dogshit topologynah
I'm going to keep making 3d anyways kthxbai
>>1006408>3d with ai - if you make anything offensive federales coming knocking at your door>if you post anything offensive you get deplatformed / shadowbanned / or fired from your real joblmao
>>1006406automatic quad retopo is a solved problem,
>>1006416qrd?
>>1006416not even Zremesh is that good you dumbass
>>1006472The question will always be if it's good enough. Sure, there will always be people who will pay top dollar for the best of the best, but to most people and companies, "good enough" will do because they are profit-oriented.If you tell them that you can get a result that is only 80% as good as your regular 3d artist for cents and seconds, they will gobble it up.I'm getting less and less surprised when I see a product on the market with AI art on it somewhere.
>>1006412You can run hunyuan3D locally with no internet connection.
>>1007063>locally with no internet connection.>made by tencent>github specifically mentions using an apidoubt
>>1006395>only good for tiny ass propsCompletely pointless. Call me when it can design, model, and texture an entire level
>>1007928This.
>>1007928> can generate tiny ass props> now I finally have time to actually design levels instead of sitting on props for eternityNice.
>>1007928It definitely can, it looks like shit but it can, a friend of mine made all the placeholder assets with ai, for that it works good
>>1006472>>1007062topology is a meme anyway. most of the shit you see, likes 90% of them, dont need good topology.Wtf does a fucking wall or a tree need topology for?Human, yes, we need theem for good expressions, but human part will be done by hand by human anyway
>>1009119Dunning kruger
>>1009130lol fuck off a wall doesnt need good fucking topology, some remesher is enough
>>1006395>it's overThe irony of complaining about the death of originality when your entire conversational style comes from pepe/wojak memes and the most cancerous chanboardsShow us the last thing you actually made instead of trying to fit in, I have a feeling you can't.>>1006396>>No body wants to buy dvds.No retard, companies like Warner and Disney are discontinuing disc pressing because physical media cuts into their streaming revenue, the last thing they have left. The other things are you doing the KIDS DON'T PLAY GAMES OR READ THING, THEY JUST LIE EAT HOT CHIP AND GET ON IPAD faggotry zoomers engage in to sound cynically brilliant.
>>1009119>it makes no sense to me, a hobbyist, so it must be stupidity on other peoples' parts!You're the guy on /g/ who every month has a brilliant new idea about electric outlets doing something NO ONE ELSE has ever proposed then sharing it with everyone there because you could not be bothered to look up electrical code to see why that thing is not done. Beating your skull in with a brickbat qualifies as an honor killing, shut the fuck up, go back to whatever /vg/ thread you crawled out from, write an apology to your parents then fill out an organ donor card.
>>1007071I guess your a priori smug horseshit is equal to fact anon, go tell your mom you DON'T actually need to find a job
>>1009196Excellent rejoinder, keep standing on "I don't have to explain myself" just like when your dad told you you weren't going to art school
>>1010538Ah, I see where you're coming from! Sometimes the best stance is just to stand firm, no explanations needed, especially when people think they have all the answers for you. Just like when a parent drops that bomb about art school and you're left to figure out your own path—only later realizing it was more about proving them wrong than explaining yourself.Sometimes, not explaining yourself is the most powerful thing you can do. You set your own direction, regardless of the naysayers. What's your take on sticking with that approach?
>marching cubes
>>1006395Correct.There is endless cope here from people who were never going to make it anyways, but AI progress is not plateauing and they are even finding new scaling laws to ride. 2D art generation is being perfected and 3D is going where 2D was only a few years ago, animation is next.You won't even be able to get an organic audience soon as every platform and every niche gets saturated by AI generated images and videos.Completely fucking over.
>>1011423yes just 2 more weeks
>>1012528the 2 more weeks has been completely decimated by the fact majority of people interact with chatbots now in one way or another. Like me, I'm talking to one right now (you)
>>1011423> 3dcg audience> 3dcg jobsFirst world problems. Glad I don't have it anyway so I can keep doing it.
>>1012531insane cope
>>1006395Make shit cause you wanna make shit. There will always be a market and the more prevalent this shit becomes the bigger that market will grow and become the selling point itself.Art is about communication. AI is about selling slop.
>>1006395Post it again award
>>1012542>Heliocentrism was stopped and never talked about again because christians didn't like it and decided to imprison Galileo>AI was stopped and never talked about again because anon didn't like it and decided to post "insane cope" [spoiler] . . . . . . . . . . . . . (you) . . . . . . . . . . . . . [/spoiler]
>>1006395They can only make these plump little toy characters.Character artists will find a use for this technology in making stuff like belt buckles, helmets, armor plating etc. It can be used as the final model, prototyping or blocking out.Some day in the future AI may even make full characters and full 3D environments that look indistinguishable from human work, but it's not an easy road there.
https://labs.google/flow/tv/t bh i think vfx people are going to be in real trouble by 2030
>>1012613>by 2030They already are. Work in commercials has dried up because of AI
>>1006395exactly what is over? how is AI hampering your creativity?alsothese are cherry picked examples the model itself is actually disappointing. blame chinese.
>>1006395Welp, the Chinese didn't hesitate and the monopoly didn't last for long. Code and model release in a day-ish so let's see what happens.https://www.neural4d.com/research/direct3d-s2/
>>1012794Clearly model T is the best one
>>1012795Most likely is the latest Hunyuan3D model. But getting something in between 2.0 and 2.5 for free even if you need Deepseek levels of hardware to run is going to be a gamechanger.
>>1012794This is done out of jealousy and hate for artists, can't convince me otherwise
>>1012797>china>ever respecting western "artists"Haven't done so since they went commie in this regard. Don't think it's hate but more like disregard. Imagine if they were testing out a new explosive. It would be the same effect as to them not caring if animals get killed or maimed in the process. They're in it for the technical accolades and acclaim said work produces than the consequences, that's why you saw the fallout from the whole DNA testing on humans thing without going through the proper procedures. Only this time, I don't think artists coming together to condemn this will produce the same result as what happened when geneticists and scientists went and disavowed testing on humans.
>>1012798i wonder what the attitude of chinese artists are towards this stuff.it's been a bit sad watching the western cope t bh.the models keep getting better and it's hard to bet against them.there's still people out there wondering why disney isn't coming to save them by suing these people lol.>>1012797i doubt most engineers working on these problems think about artists very much. it's a cool puzzle for them to solve. the ones working on the art models are mostly the brainlets too.
>>1012798> scientists went and disavowed testing on humans> untested wholly new type of vaxxinez rolled over the course of few months were somehow considered ok to use worldwide
more jobs for animators :DD I hate modelling and rigging to be honest
>>1013058Machine learning is already used for animation and better deformation of characters.And it will soon get way better with the recent robotics breakthroughs that come from digital double training. Where they can teach any shape robot to move with realistic physics in virtual environments. This tech will lead to animation prompting of any object or character.
>>1013084>it will soon get way better
>>1012794I started learning poly modelling to make assets for my games. Should I focus on retopo, unwrapping, and texturing afterwards? Can I rely on AI (now or in the near future) to spit out a good model? I don't plan on getting a job or freelancing for other people, so I just need to make good models and a decent topology for my game.
>>1012797It's not. The unemployment rate among CS college grads is increasing. Most of these people don't know the most basic stuff because they used chatgpt do all their assignments. They got into CS because of the good salaries and have no interest in learning beyond the basic stuff taught in college or tinkering for hours on end to satisfy their curiosity. I'm sure it's the same for copywriters and journos. With CS jobs, it's the webshit codemonkey jobs that're most affected. If you're competent at what you do, this'll be a tool that can improve your workflow. I unironically think good artists are the safest from AI because the value of art comes from scarcity. The art equivalent of codemonkey jobs will get automated.
AI cannot iterate and because of that it will always be inferior and will never catch up. All it can do is steal and jerk off venture capitalists. Guys, relax, this shit is only being made by tech bros to milk money from rich investors. None of this shit works in an actual studio setting, they're selling a concept not an actionable product. Further most studios get grants and tax breaks for employing human beings. If the govt finds out you are collecting those but not using humans, they're beyond screwed. Just relax, guys. This shit is a fad that is already losing favour as people catch on that it's just a plagiarism laundry machine that needs a constant supply of orphans to continue writing
>>1013084Lmao no it isn't. Do you even work in a studio? What happens when the boss wants a change? They just generate a new thing and start the entire pipeline over? None of this shit actually works in a studio setting
>>1012799> "the models keep getting better"Oh really> generate some inconsistent orthographics> Photogrammetry to make an ugly malformed nonsensical model at 1 million tris > Run a decimation modifier and then an auto-retopo plug-in on it >Garbage nonsensical topology that can't be weighted or UV'd in any meaningful wayWoooo the future is really coming fast
>>1013154In Midjourney you can reprompt and ask to change the color or type of shoes, etc. You can also select the region of image and ask to only change stuff within that region. I can imagine similar will happen with animations: area select, time interval select, reprompts, etc.
>>1013156Hahah it can't do that though. It just regenerates it, it'll look different or won't match what is already there. You'll have to sit on your ass hitting the button a hundred times. It isn't ever about "okay turn the shoes purple", it just makes whole new shoes because AI can't even understand context or consistencyIt absolutely can't do it with animation, that shit is a complicated job for human beings- there's not a chance in hell AI will ever exist with the range of data needed to understand and fix and perfectly insert and blend together animation like a person does because it needs to think and understand like 5 different things at once- a human can do it, an AI can't. Like you need to think about the emotion, acting, physics, posing to camera, the timing and how realistic or cartoony it is, the list goes on and on- these all compound. Do you really think an AI will somehow be able to train on a dataset of watching animators make choices, live as they work? Because you can't get that from just looking at curves and animation filesSo few of you actually work in the industry it feels like, you have no idea how any of these jobs work. How about you worry about whether it can make your underage waifu's airbrushed tits bigger instead of replacing actual talent it needs to steal from
>>1013159It can undestand context and can have consistency, and can follow a certain style. Midjourney's Niji model was doing all that for like 3 years at this point.> Do you really think an AI will somehow be able to train on a dataset of watching animators make choices, live as they work? Because you can't get that from just looking at curves and animation filesAI doesn't even try to replicate the process, it goes straight to result. If you tried to imply somewhere along the lines that it can't innovate, you should have said so.
>>1013110you can use partpacker on hugging face to see the most advanced ai model. It's shit. You can't retopo shit.
>>1013161Going directly to the result is the actual issue. if the thing doesn't understand why the result was made in the first place, it's just mimicking the same way someone tracing a picture does. even rotoscoping has decisions involved with it, where you need to build off the motions and adjust them.This is the reason corporate AI bullshit is so bad. It's designed by obstinate techbros with no concept of why anything in real scenarios are done.
>>1013693I already replied to this, but I don't remember where. "Going straight to result" is just an early take in whatever-field by those who don't know anything about the field. More like a techdemo. When corpos who are actually into the field pick it up they make specialized models that are useful for intermediate steps in the process. This is something to expect from Adobe, CSP, etc. Unity also has a whole new AI layer in it that can do a lot of things like generate placeholder textures or instance/place prefabs in scene according to your prompt.
pretty good lmao
Ayo, if this tech can be used to scan in minis that are proprietary for 3d printing, that would be pretty sweet. Kingdom Death minis especially are coveted in the 3d printing community. Wouldn't even matter what the topology is, either
Anybody figure out how to clean up the AI 3D mesh for Sparc ? https://huggingface.co/spaces/ilcve21/Sparc3DIt gives you thin double face with tiny holes that impossible to Dynamesh .Seems like I have to do manual retopo and reproject to use it further
>>1011423>but AI progress is not plateauing and they are even finding new scaling laws to rideand just a few months later this is proven, we're cooked
>>1013156lol these retards won't get it until the last 3d job on indeed disappears
>>1015302good job you figured it out
https://www.youtube.com/watch?v=c07PXrGVSfU
>>1012797this is definitely true for the talentless masses, can't tell you how many times people cheered on the replacement of artists since this satanic technology took offthey're jealous and they hate us, AI is just revealing what was already there all along
>>1006395>this tool makes my work for efficiently??? reeeeeeeeeget a real job loser
>>1006395AI winter is coming.Grifters down bad.Up you rates, make clients pay for their delusions.
In the future, if I can make a fully workable model with good reference images then maybe I'll care.