3D DESIGN WITH TEXT-BASED AI

3D DESIGN WITH TEXT-BASED AI
6 Comments by: Bryan Cockfield
May 12, 2023

Generative AI is the new thing right now, proving to be a useful tool both for professional programmers, writers of high school essays and all kinds of other applications in between. It's also been shown to be effective in generating images, as the DALL-E program has demonstrated with its impressive image-creating abilities. It should surprise no one as this type of AI continues to make in-roads into other areas, this time with a program from OpenAI called Shap-E which can render 3D images.

Like most of OpenAI's offerings, this takes plain language as its input and can generate relatively simple 3D models with this text. The examples given by OpenAI include some bizarre models using text prompts such as a chair shaped like an avocado or an airplane that looks like a banana. It can generate textured meshes and neural radiance fields, both of which have various advantages when it comes to available computing power, training methods, and other considerations. The 3D models that it is able to generate have a Super Nintendo-style feel to them but we can only expect this technology to grow exponentially like other AI has been doing lately.

For those wondering about the name, it's apparently a play on the 2D rendering program DALL-E which is itself a combination of the names of the famous robot WALL-E and the famous artist Salvador Dali. The Shap-E program is available for anyone to use from this GitHub page. Even though this code comes from OpenAI themselves, plenty are speculating that the AI revolution to come will largely come from open-source sources rather than OpenAI or Google, something for which the future is somewhat hazy.

HarveyH55 wrote:

3D DESIGN WITH TEXT-BASED AI
6 Comments by: Bryan Cockfield
May 12, 2023

Generative AI is the new thing right now, proving to be a useful tool both for professional programmers, writers of high school essays and all kinds of other applications in between. It's also been shown to be effective in generating images, as the DALL-E program has demonstrated with its impressive image-creating abilities. It should surprise no one as this type of AI continues to make in-roads into other areas, this time with a program from OpenAI called Shap-E which can render 3D images.

IBdaMann wrote:

HarveyH55 wrote:

3D DESIGN WITH TEXT-BASED AI
6 Comments by: Bryan Cockfield
May 12, 2023

Generative AI is the new thing right now, proving to be a useful tool both for professional programmers, writers of high school essays and all kinds of other applications in between. It's also been shown to be effective in generating images, as the DALL-E program has demonstrated with its impressive image-creating abilities. It should surprise no one as this type of AI continues to make in-roads into other areas, this time with a program from OpenAI called Shap-E which can render 3D images.

Please remember that "dialog generation" is only a fraction of the generative AI. I think (just a belief of mine at this point) that the vast bulk of generative AI is data synthesis, generating artificial (synthetic) data that AI systems can use to test and to "practice" (i.e. learn) or for legions of non-AI applications that need data for testing and for fine-tuning.

What I find the most intriguing are the "adversarial" data generators. These involve two competing AI systems. System A is up at bat and generates realistic synthetic data. System B is in the field and fields both real data and synthetic data. When system B fields some data, it determines whether the data is real or is synthetic data provided by System A (getting System A called "out"). System A, of course, tries to get a hit by generating data that appears real enough to fool System B, landing System A on-base.

Actually, there is no baseball involved, but the two systems each continue to improve the other. The "competition" is the machine learning in this case.

HarveyH55 wrote:
Link

3D DESIGN WITH TEXT-BASED AI
6 Comments by: Bryan Cockfield
May 12, 2023

Generative AI is the new thing right now, proving to be a useful tool both for professional programmers, writers of high school essays and all kinds of other applications in between. It's also been shown to be effective in generating images, as the DALL-E program has demonstrated with its impressive image-creating abilities. It should surprise no one as this type of AI continues to make in-roads into other areas, this time with a program from OpenAI called Shap-E which can render 3D images.

Like most of OpenAI's offerings, this takes plain language as its input and can generate relatively simple 3D models with this text. The examples given by OpenAI include some bizarre models using text prompts such as a chair shaped like an avocado or an airplane that looks like a banana. It can generate textured meshes and neural radiance fields, both of which have various advantages when it comes to available computing power, training methods, and other considerations. The 3D models that it is able to generate have a Super Nintendo-style feel to them but we can only expect this technology to grow exponentially like other AI has been doing lately.

For those wondering about the name, it's apparently a play on the 2D rendering program DALL-E which is itself a combination of the names of the famous robot WALL-E and the famous artist Salvador Dali. The Shap-E program is available for anyone to use from this GitHub page. Even though this code comes from OpenAI themselves, plenty are speculating that the AI revolution to come will largely come from open-source sources rather than OpenAI or Google, something for which the future is somewhat hazy.

With all the fear-mongering, AI taking over the world, we're all going to die posts... Figured an article about where AI really is, and can do would be appropriate. At the link, there are images, though not impressive at all. Given time, the programs will get better refined, more useful. This is still pretty useful though, save a lot of the initial time spent designing basic shape and features for a human designer to start with, and add detail. 3d models are just computer graphics anymore. 3d printers can make them real in plastic. Though my experience with a $200 printer wasn't impressive. Some types of things look nice, hold up okay. But most things break easy, if there is mechanical stress involved. GOPRO camera mounts always fail... Takes quite a while to setup, level the bed, usual an hour to actually print something small. Doesn't turnout useful, you tweak settings, change the orientation on the bed, and try again. This would be a good application of AI, and expect to see it integrated in to 3D printing before long.

Plastic is the only material that can be used with 3D models, CNC milling can do metal, and a lot of other materials. CNC laser cutters/engravers...

AI is a tool, that's been in development for decades. Still has a long way to go, before threatening our existence. A lot of the hype/hysteria is generated from people wanting to attract funding for their projects. It's meant to show how far they have gotten, and just the best they have to offer. That's not the actual, usual results, just the most ideal they can generate. Doesn't mean it's repeatable, or stable. I do believe Open Source will produce most of the advances. Most of the work done freely, voluntarily, by anyone with and interest, and some programming skills. Employers have specific goals for projects, very little freedom to do anything else. Open Source means you can branch off the original project anytime you want, and develop your own ideas and goals. You still have access to everything previously done, any progress, or other branches. Might even attract a few people, who like the direction your branch is taking.