It over! AI has won. We are about to become obsolete

At 3 minutes into the video, you see the AI can make great topology of any complex shape it is show a photo of. Making an artists job obsolete.
This is far beyond photo scanning. And im not sure we will have jobs in the future…

3 Likes

for sure those tools will help us - but replace us ? Not sure.

CG is an ever-evolving field.

As an analogy:
Robots in manufacturing just enabled better faster consistent work.
Some old job types went away but new one came.

In engineering I can model a structure by hand or let the AI solve it.
But the design I still need to continue refining!

4 Likes

Believe me, AI won’t replace us as said by @cekuhnen they will help us because at the end of the day, it is about imagination which comes from our feelings which are shaped through our experience, and in the near future, no robot will have that.

2 Likes

Which video did you actually watch? Must have been sth. different from the one you linked, I guess. Triangle-soup is a farcry from ‘great topology’.

It seems an interesting paper and approach, but in many ways, I don’t see how the process proposed (and the results obtained) is really all that fundamentally different from good ol’ photogrammetry, which you can use right now, in the form of both a number of open source solutions (e.g. Meshroom) as well as commercial closed source programs (e.g. Agisoft Photoscan).

What’s so revolutionary about ‘calculate a 3D-model from a bunch of photos of a real object’?

Maybe this gives comparable results from much fewer photos than photogrammetry, if so, that’s cool, but ‘great topology’?
I think you’ve gotten carried away a little…

greetings, Kologe

Imagine a very greedy corporation wanting to maximize its profits regardless of the social consequences of the community. So they will want machines doing the work of expensive paid humans, and no regulation to stop them from doing it. Nothing new…
So yes, eventually (maybe not soon) the work of 3D human modelers will be reduced.

(And please don’t be too literal. What I mean for example is that if you can easily get a good base mesh from a photo, maybe the company now needs only a single human modeler instead of 10 human workers)

AI can’t go out in the field and scan things with an iPhone or hand scanner, then turn on the computer and hit the “go” button on the keyboard (yet). In all seriousness, this is a tool that will make a human’s job easier, just like any other tool.

Are you forgetting that all meshes are triangulated when they hit any game engine. Video game companies at least will be totally satisfied with those meshes, and im sure they will deform just fine under animation.

I understand your optimism. But i can easily imagine a greedy company that write a simple script to automatically feed this AI photos from the internet and has it churn out these meshes, which they will then probably hire one human to do quality control, and then they will sell the meshes onto the public.
Our jobs as modelers will be made nearly obsolete except for edge cases. What next, the AI coming up with realistic materials?? oh they already have that, watch the video. And the AI will just continue to grow in versatility.
Im not so optimistic as you all are.

Automation doesn’t replace human labor, but it does displace it.

There are still people working in manufacturing in the US, but waaaay fewer with the invest in automation.

For example, check out this 6 minute video on plywood production:

Notice how many people are working in the factory back in the 50’s and how very few of them there are working nowadays.

People really, genuinely, literally lose jobs to automation. sure, they go find jobs elsewhere, but would you rather be making a family wage and pension working in the mill, or making burritos at chipotle?

4 Likes

I understand your argument, but i have to disagree with using the non-AI based automaton as an analogy. I think AI is different and unprecedented. We have to see the big picture, that these AIs are creeping into every single facet of work or production as we head into the future. It wont be limited or stopped by lack of creativity or imagination. We wont all just be able to move to new easier jobs because of it. I see it taking over and humans being left as useless eaters… yikes…

will many modelers be out of work? probably.
artists? probably not anytime soon… :slight_smile:

(there is a very inflationary use of the term “artist” in the 3d-animation world…)

It is exactly the same as any automation. Narrowly specialised tool made by humans. A lot of math in the paper shows that it is not something neural network came by its own.
BTW, if you study the paper itself, all this algoritm does is to reverse engineer a mesh from rendered! images. Not even photographs. Bust of Nefertiti, for example, said to take 25 renders as an input. Next to no practical use, purely academic interest. Even if it would work on photographs, it will be just a shitty equivalent of photogrammetry. So please stop helping clickbait channels spread misinformation. Society’s obsession with “ai” is concerning, but in most cases, especially in art field. results are meh.

1 Like

Topic moved from #general-forums:blender-and-cg-discussions to #general-forums:off-topic-chat

Guess it’s a good thing I moved on to mostly coding and shaders from modeling.

Now with a tool like this, should be easier to get the art for my game.

Also, you don’t even need to take pictures, AI can already generate side and front profiles, whatever you want, and can even procedurally generate monsters and races.
Neural networks are incredibly good, and will only get better with time.

It’s the singularity, folks.

Most versions of Utopia involve not having to work.

1 Like

None of it is going to help much to pluck the optimised NPR gaming assets I need out of my rather vivid, underground comic influenced imagination, is it?