The idea about a Blender addon for textures was really good. I really hope Blender will hire some people capable of tapping into the possibilities that an open source image generating ai such as Stable Diffusion brings, meanwhile, third party addons/plugins will likely do this shortly.
Heh, yes. Itâs like when coming up with pen and paper, saying you hope people will only use this in a ethical and moral manner. Of course what is ethical and moral often differ from person to person, even country to country etc.
1 Like
Chinese gaming company NetDragon appoints virtual humanoid robot as CEO
NetDragon Websoft is now under the leadership of Ms. Tang Yu, one of the most advanced virtual humanoid robots powered by AI. Apart from the usual managerial responsibilities, CEO Tang Yu can act as a real-time data hub and analytical tool.
Thatâs some creative tax dodging!
1 Like
Considering the company is based in China, it could also mean the entire executive suite being sent to Xinjiang if the new CEO glitches and spews a lot of unflattering things about the Party.
For their sake, I hope they checked and doublechecked the censorship algorithms.
To note there is actually a precedent here, at one point IBMâs Watson AI downloaded a website known as the Urban Dictionary and started to constantly spew profane remarks (to the point where the dev. team had to wipe his memory).
1 Like
3 Likes
Adesk have a habit of grave-yarding projects. Look at Stingray(3dsMax Interactive) and the very strange 3dsMax ART path-tracer that they developed in-house shortly before putting Arnold into Max.
3 Likes
Halway through this atm, I know Nanite is not that new, but this is some more info on how it was invented and some of the thought processes behind itâŚ
6 Likes
https://twitter.com/MomentsInCG/status/1545494688764936200
Paper: https://momentsingraphics.de/HPG2022.html
Results: https://momentsingraphics.de/Media/HPG2022/results_viewer/results.html
5 Likes
I suspect naming an AI as a CEO is mostly just a publicity stunt. Iâm sure it performs some useful technical tasks (probably alongside the kinds of creative errors only a neural network incapable of truly comprehending itâs own output can, of course), but I think itâs highly doubtful that even the most fervent AI fanboy thinks that neural networks can be trusted to reliably override humans for top-level company decisions at this point, so the actual decision making aspects of the CEO job are likely being performed by a human still.
But itâs a perfect publicity stunt. Easy way to make the news, easier than producing something newsworthy in house.
But yeah, I definitely wouldnât leave the neural network with a twitter account or whatever unsupervised. Humans donât always know when itâs wise to keep their mouths shut but an AI doesnât even truly comprehend what itâs saying.
3 Likes
Relighting app
2 Likes
As several industries are moving towards modeling massive 3D virtual worlds, the need for content creation tools that can scale in terms of the quantity, quality, and diversity of 3D content is becoming evident. In our work, we aim to train performant 3D generative models that synthesize textured meshes which can be directly consumed by 3D rendering engines, thus immediately usable in downstream applications. Prior works on 3D generative modeling either lack geometric details, are limited in the mesh topology they can produce, typically do not support textures, or utilize neural renderers in the synthesis process, which makes their use in common 3D software non-trivial. In this work, we introduce GET3D, a Generative model that directly generates Explicit Textured 3D meshes with complex topology, rich geometric details, and high fidelity textures. We bridge recent success in the differentiable surface modeling, differentiable rendering as well as 2D Generative Adversarial Networks to train our model from 2D image collections. GET3D is able to generate high-quality 3D textured meshes, ranging from cars, chairs, animals, motorbikes and human characters to buildings, achieving significant improvements over previous methods.
4 Likes
2 Likes
4 Likes
This tech has huge implications for modelling workflows. You really need to click through the replies in the tweet to understand what is going on here. Can only imagine where this will be in a few years!
https://twitter.com/JunGao33210520/status/1573310606320484352
2 Likes
8 Likes
MDM: Human Motion Diffusion Model
abstract: https://arxiv.org/abs/2209.14916
project page: https://guytevet.github.io/mdm-page/
7 Likes
It seems like every week some new advance in AI comes along. How long before the advance becomes exponential because the AI is now doing its own R&D and improving itself?
Before anyone thinks I am stretching things here, keep in mind that AI products that can actually produce useful results did not even exist just a few years ago (even with the use of statistical data as training wheels).