[!warning] Dumping random YouTube videos constantly is not a quality contribution.
If we wanted random YouTube videos, we could go find them ourselves. Please put effort into contributions if you’re going to contribute. If you have something to say about a video, say it
One has to appreciate my initial concerns a few months ago (which were delimited by mods which you all know) before the Anthropic episode. I am grateful the Blender Foundation and the Godot Foundation understand the absurd anti-artistic nonsense the creative world is forced to deal with. THIS the above…
above Joseph, obviously.
Look above… look at this utter anti-blender trash!
I just watched Higgsfield’s Oneric https://youtu.be/aAg9iDh9_BQ?si=NhMbn9rILVz6lS4U
While it still looks a bit sloppy in terms on the way the characters speak, it’s getting so damn close to realism.
What made me pretty depressed are the comments. overwhelmingly in favour of it.
As someone who always wanted to go into filmmaking but never did, it just feels like in the future the only way to compete will be to use AI since that’s where most content will be made. There is going to be so much AI slop, even if it is at a higher quality, it still feels like cheating to me
Why is the goal always to “compete”? Why not have the goal of making a good movie that tells a good story? AI hasn’t changed the odds of becoming a “successful” indie filmmaker, which are about 1 in 1000000000. Go to any Starbucks in LA and you’ll find 20 “failed” filmmakers. Forget being “successful”. Make good stuff how you want to make it
The KEY problem with every AI generated content is - and always will be - the total lack of originality and creativity: LLMs copy and paste (slightly different version of) already existing things, things created (and STOLEN by LLMs - the models, their training.. ) by real people.
So it will become more realistic and less sloppy .. but.. LLMs can’t create new contents, ideas, concepts: they replicate and mutate ad infinitum the things they ate.
Anyway.. I think the REAL problem in starting to do real filmmaking from scratch (like quite anything else in life) is always the same: money. Costs. Making a (real) movie costs a lot of money.
Anyway.. Hollywood wasn’t very well even before the AI hallucination.. tons of crappy movies abounded already.. AI only makes the crap being cheaper..
Just saw this yesterday..
“but ooooh using 3D software takes effoooooort… do you expect me to actually use my creativity and brain to make something good instead of following the current media trends and become brainlessly popular online?”
<rant>
quoting myself here, because now 4 years later now we have the image to 3d model generation technology, but instead of loving it i hate it especially if you have to fix the model (given by a client for example) which is a waste of time to fix depending how broken it is because the topology output may be bad like lots of non-manifold topology, holes and so on that is not easy to cleanup or time consuming to do, tho i don’t mean whether is quads or triangles topology but depends on the purpose as you may require to clean/polish it for 3d printing then triangles topology would be fine, but for rigging/animation then it means to retopologize and optimize all that shitty ai model into something usable…
Other than that also getting ads every day in youtube/social media of the popular ai mesh generators e.g Meshy, Tripo, etc without even haven’t used any of them LMAO…
</rant>
Now the other reality is that these generators seem that will get better over time as long as more training data is fed into it, and that’s sad/depressing as it kills the art of the modelling, sculpting, etc process for any model for any future wanna be 3d artists…
Also, most of the 3d mesh generators if you can see are from asian companies, not a coincidence, right? Asians just as with technology in general, are trying to dominate the 3d market with slop and less possibility of jobs or commissions for 3d artists, because you can make it for cheap and without no human labor cost?
BTW guys what are your thoughts on the recent Artstation and Sketchfab acquisition? no AI “artworks” anymore perhaps, or you feel like nothing is gonna happen regarding removing these slop content. I will hold my breath about that…
When you have your own unique voice/style well articulated, you will succeed. Finding a way.
Ca$h is key, but only to a certain extent. Blair Witch Project was shot for under 50k. Daniel Myrick success is due his somewhat unique idea for a film.
Tons of accolades, 250 million in ca$h later … his own unique vision is proof positive.
No way AI could ever be prompted to deliver such a powerful theme … that OI (original intelligence), comes from a human being’s very own human vision.
Compelling visuals will always be achieved with AI, that’s just human want. AI, is a tool to get shit done.
But the OI ones that get remembered and get talked about, come from that place of our own shared unique human experience.
Insane dichotomy … unique yet common.
Insanely compelling. I’m there …
I personally think that AI will accelerate OI. Not as a push back, but rather because of an inspiration. And certainly not to thin the herd.
Funny, I see the same problem in 95% of all hobby movie projects. Maybe the 5% of humans that make very original stuff could theoretically do the same with AI?
The appropriate name for what is sold as AI is LLM (language learning machine).
Creativity and innovation comes from breaking schemes and rules.. and a LLM can’t (and never will be able to) create something that is not part of its training and model. Every result in video and image creation through AI has always something.. “off” .. that doesn’t make you think you’re seeing something original and really innovative, but a pejorative version of something you’ve already seen (and it is just that).
Anyway.. I see more and more enthusiastic people that seems to love that slop everywhere.. and I think the apex of evolution for humanity is already past behind the current times.. we are now (happily) descending down the slop(e).
This video is a great summary of the research done by Emily Bender, Alex Hanna, Cory Doctorow and Gary Marcus. It is also worthy to check Timnit Gebru´s work.
Lot´s of great insights on the LLM technology and a great start to not get (too) lost on the hype.
Kinda like human ideas are linked to our past experiences. LLM’s and genAI’s recombine btw, they don’t regurgitate. Just like us.
If anything, thinking about AI might teach us something about the nature of our own ideas, self-consciousness and ability to predict the next word or event.
This is sad
, I knew Cara sooner or later would get scraped just for existing in the internet…
This right here is why people hate AI. I hope Andrew P is happy
Comparing the complexity of a human brain and its capabilities with those crappy and so limited LLMs algorithms is .. so silly and dumb.
I don’t know why all AI fanatics don’t seem to understand anything about the “tools” they love and adore.. anyway LLMs can only recombine things they stol… uhm.. “learned” .. with others things they stol.. “learned”.
There’s 0% creation, 0% awareness… those algorithms are so dumb they even don’t understand a crap in what they do (that’s why is so difficult to “control” them).
So trying to compare the things those LLMs can do to the creative process of a real human being… I’m so disconcerted that I have no words left.
I’d like to keep it civil and avoid insults like calling each other dumb, is that okay with you?
I think we can agree that a large portion of human creative output is derivative or just imitation too. The remaining 10% is the Nina Simone’s, the Rodins, the (your favorite author)
A lot of what AI spits out at this point is laughable too. I rendered a video the other day with a man sitting at a bar, his back to us. Veo3 made his torso spin 180 degrees to look at the camera.
But I’ve also had discussions with LLM’s about a new business I started. Based on my detailed briefing, it came with good ideas for web content and even typography and logo proposals. I could trace back those ideas to what it had learned from humanity, but its output was way better than mine.
In 100 disciplines, a medium-to-good output will outperform a human who’s only good acts few things, we can’t be experts at everything.
By comparing generative AI to (any great painter), you cover only a small portion of the creative spectrum. I’m doing B2B animations where I sometimes input the client brand guide into a genAI system to make non-Hero illustrations to fit the budget. Those are very, very good.
Research confirms this: genAI compares to a creative medior profile, never the best humans in their field. That’s enough to support my creative workflows.
the fact that generative AI and LLMs have gotten such a grip on the 3d modeling community, esp in the “virtual world” gaming spaces has killed any desire to finish my projects. It is genuinely distressing to see whole swathes of people plugging code into Claude, putting textures and models through AI to “fix” because the idea of needing to manually edit or flat out create something is for cavemen and losers apparently.
I don’t want to attend your event if there’s an AI poster! I don’t want to buy your things if it’s made with AI! God forbid I point out that beyond the lack of creativity, the use of AI is causing blackouts and drought where the data centers are built, that ai psychosis is killing people who can’t tell that the yes-man text generator isn’t actually a real person. Then, the shills will pour out of the woodwork saying “You shouldn’t talk bad about AI, you can’t bring real world issues into this!”
Generative AI and LLMs are useless garbage that have only hindered our ability to think and create, and when the bubble pops I will cheer and scream for joy.