Author Topic: ai corona renderer  (Read 1945 times)

2024-02-06, 15:42:06

MARDUK

  • Active Users
  • **
  • Posts: 54
    • View Profile
I know that the topic has been talked about a lot everywhere
But it is necessary to confront this matter and not hide from it
3D designs eventually turn into images
Image generation is easy through artificial intelligence algorithms
There is no comparison between human creativity and the creativity of artificial intelligence, as it is undoubtedly superior because its creativity is based on mathematical logic, accuracy of calculations, and sound measurements without errors, and it will develop more and more.
I feel like we are in a state of extreme confusion and everything has to be AI driven
Even the Corona Render should be like that
Otherwise, it will be of no use soon and without exaggeration
I do not know what the plans of the Corona team are, but the least thing is to introduce artificial intelligence into the socket by generating a wide choice of internal and external art instead of going to sites here and there that operate with artificial intelligence algorithms. We upload the sketch and it produces a wonderful image for us and we design based on it.
Designing with these programs' only benefit now is taking shots from multiple angles and having freedom to modify
But soon artificial intelligence will become adjustable with quick button presses, so there is no need for 3D programs to create an interior or exterior design, for example.
Maybe for animation
Instead of waiting for this to happen, why doesn’t the Corona team amend all its policies and plans to suit this development and add artificial intelligence to the socket to generate images and facilitate the creativity process, so that the subscription process for a fee becomes logical in addition to the freedom to modify within the socket itself?
Just ideas I'd love to put forward

2024-02-06, 16:21:16
Reply #1

TomG

  • Administrator
  • Active Users
  • *****
  • Posts: 5575
    • View Profile
Chaos' plans for AI are outlined at https://www.chaos.com/next

AI is a long way away from delivering the high quality, highly controllable means to generate specific results and then make specific changes, which means there will be no replacing of 3D modeling software and existing render engines any time soon (next 2 years or more at least). I've been experimenting with AI a lot, and so far my conclusion is "it is highly freeing, and highly restrictive, both at the same time". Things that are trivial in 3D modeling and rendering are currently impossible in AI generated imagery.

Where AI can help is in enhancing existing renders, or speeding up workflows in assembling scenes etc, and this is where our focus will lie for putting AI to good use. As for things in 5 years time, well we'll see when we get there (or before we get there to be honest, as we're watching the direction of everything).
Tom Grimes | chaos-corona.com
Product Manager | contact us

2024-02-06, 16:28:52
Reply #2

James Vella

  • Active Users
  • **
  • Posts: 581
    • View Profile
my favorite thing about AI so far is:

user: Write me a maxscript that creates a cube:
AI: const geometry = new THREE.BoxGeometry();

user: thats javascript, i need it in maxscript
AI: Oh Im sorry, its box{}

user: we dont use curly braces in maxscript to create an object, we use ()
AI: Oh Im sorry, its box{}

user: really?....
« Last Edit: 2024-02-06, 16:35:44 by James Vella »

2024-02-06, 16:46:26
Reply #3

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 12845
  • Marcin
    • View Profile
This post contains several errors and incorrect assumptions. Firstly, it oversimplifies the process of image generation through artificial intelligence algorithms, which are not always flawless and can produce errors. Additionally, the claim that AI creativity is undoubtedly superior to human creativity is unfounded, as human creativity encompasses aspects beyond mathematical logic and calculations. Furthermore, suggesting that all design processes should be AI-driven disregards the complexity and nuance involved in creative endeavors. Finally, the idea that the Corona Render team should entirely shift their policies to accommodate AI-generated images overlooks the diversity of user preferences and needs.

;)
Marcin Miodek | chaos-corona.com
3D Support Team Lead - Corona | contact us

2024-02-06, 17:54:52
Reply #4

MARDUK

  • Active Users
  • **
  • Posts: 54
    • View Profile
Chaos' plans for AI are outlined at https://www.chaos.com/next

AI is a long way away from delivering the high quality, highly controllable means to generate specific results and then make specific changes, which means there will be no replacing of 3D modeling software and existing render engines any time soon (next 2 years or more at least). I've been experimenting with AI a lot, and so far my conclusion is "it is highly freeing, and highly restrictive, both at the same time". Things that are trivial in 3D modeling and rendering are currently impossible in AI generated imagery.

Where AI can help is in enhancing existing renders, or speeding up workflows in assembling scenes etc, and this is where our focus will lie for putting AI to good use. As for things in 5 years time, well we'll see when we get there (or before we get there to be honest, as we're watching the direction of everything).



Thank you for your reply
I wanted to inquire about the direction of those responsible for these 3D programs and their providers
Because this field took a lot from our lives in terms of learning and trying to hone the skill, and we are surprised by the presence of artificial intelligence, which destroyed all this long journey on which we worked for many years to try to understand scientific matters related to colors, art, learning software tools...etc.
It was frustrating, it was impossible to escape the fact that AI was superior, and every programmer knew that it relied on logic that was stronger than human logic.
Ultimately, humans created programming to do logical and complex operations for them
We cannot understand all this advanced complexity now
And of course with AI eventually it will take all our jobs, and that's a good thing because it will give us peace of mind.
We'll just do yoga
But from here until then it will be confusing and dangerous, especially for those working in this field

2024-02-06, 17:56:11
Reply #5

MARDUK

  • Active Users
  • **
  • Posts: 54
    • View Profile
my favorite thing about AI so far is:

user: Write me a maxscript that creates a cube:
AI: const geometry = new THREE.BoxGeometry();

user: thats javascript, i need it in maxscript
AI: Oh Im sorry, its box{}

user: we dont use curly braces in maxscript to create an object, we use ()
AI: Oh Im sorry, its box{}

user: really?....


hahaha
But he will get better soon, my brother, and with words and sounds he will make any model you want

2024-02-06, 18:06:24
Reply #6

MARDUK

  • Active Users
  • **
  • Posts: 54
    • View Profile
This post contains several errors and incorrect assumptions. Firstly, it oversimplifies the process of image generation through artificial intelligence algorithms, which are not always flawless and can produce errors. Additionally, the claim that AI creativity is undoubtedly superior to human creativity is unfounded, as human creativity encompasses aspects beyond mathematical logic and calculations. Furthermore, suggesting that all design processes should be AI-driven disregards the complexity and nuance involved in creative endeavors. Finally, the idea that the Corona Render team should entirely shift their policies to accommodate AI-generated images overlooks the diversity of user preferences and needs.

;)


Thank you for the response, brother
« Last Edit: 2024-02-07, 00:41:33 by MARDUK »

2024-02-25, 21:41:20
Reply #7

sebastian___

  • Active Users
  • **
  • Posts: 197
    • View Profile
I had pretty good results using AI to write max scripts. And of course much more complex ones than just making me a cube. But you have to try multiple LLM, like bing chat, maybe switching between algorithms and so on.

The development of a script would be much improved if a LLM would be able to test itself what he writes, instead of back and forth: here's the code, and now give me back the debug error and now here's the next version, and so on.

As for the rendering, I think it would be great to have an AI renderer, or "interpreter" directly in max viewport or render buffer, instead of taking an image from Max and moving that image on all kinds of websites.
As far as I know there's no such initiative for max, but there are a few for blender and maybe c4d.