Hacker News Reader

dmarcoson World Labs: Generate 3D worlds from a single image

This indeed looks more like photogrammetry than a diffusion model predicting the next frame. There's 3D information extracted from the input image and likely additional generated poses that allow reconstructing the scene with gaussian splats. Not sure how much segmentation (understanding of each part of the scene) is going on. Probably not much if I have to guess.

dmarcoson World Labs: Generate 3D worlds from a single image

Yeah. That's the attitude! It's all about playing around the constraints. Tech has limitations? Yes, but also opens tons of new possibilities.

dmarcoson World Labs: Generate 3D worlds from a single image

if they fulfill the mission it will apply to domains other than games like movies, robotics, architecture...

dmarcoson World Labs: Generate 3D worlds from a single image

It's definitely a balancing act. World labs was stealth for a bit. Without a brand, stated mission, examples / demos of what you are capable of... is harder to hire, fund raise or get the attention and mind-share you need once you are ready to ship product.

The risk is setting expectations that can't be fulfilled.

I'm in the 3D space and I'm optimistic about World Labs.

dmarcoson World Labs: Generate 3D worlds from a single image

You can definitely mix gaussian splats and “traditional” meshes

Splats very new and still many things to figure out: relighting, animation, interactivity.

dmarcoson World Labs: Generate 3D worlds from a single image

Yeah parallax, reflexions, shadows are as important as stereo. We’ve been always sold that stereo = 3D but it’s just one among many cues that the brain relies on.

dmarcoson World Labs: Generate 3D worlds from a single image

To be fair they haven’t launched they are showing progress and laying out the vision.

What previous work are you referring to?

dmarcoson World Labs: Generate 3D worlds from a single image

In some angles you can see there’s some gaussian splat / point cloud representation underneath. There’s definitely a 3D representation. But yeah navigable volume is limited at the moment. It will improve

dmarcoson World Labs: Generate 3D worlds from a single image

Image generation has improved a lot in just 2 years: no more 7 fingers hands, text rendering, general image quality…

We’re just getting started with 3D and incentives for it to improve are strong

dmarcoson World Labs: Generate 3D worlds from a single image

Definitely if there’s a future for 3D and immersive video it depends on adding more cues other than just stereo. Lack of parallax one of main reasons causing discomfort for many.

dmarcoson World Labs: Generate 3D worlds from a single image

Good hack!

dmarcoson World Labs: Generate 3D worlds from a single image

Fair criticism. I’m also not a fan of hyperbole. Still find World Labs stuff super intriguing and I’m optimist about them to be able to fulfill the vision.

dmarcoson World Labs: Generate 3D worlds from a single image

In the looking ahead section of the post it says:

“We are hard at work improving the size and fidelity of our generated worlds”

I imagine the further you move from the input image, the more the model has to make up information and the harder to keep it consistent. Similar problem with video generation.

dmarcoson World Labs: Generate 3D worlds from a single image

Yeah A-Frame! It makes me happy to see my many years of maintaining it paying off

dmarcoson World Labs: Generate 3D worlds from a single image

I’m sure It’ll improve. I imagine, the further from the input image the more the model has to make up stuff. Gen-AI video models are limited to a few seconds. In 3D you’re constrained to a volume

dmarcoson World Labs: Generate 3D worlds from a single image

In some angles you can appreciate artifacts that resemble those of gaussian splats. I’d bet is a 3D representation but not your traditional mesh. Very cool

dmarcoson World Labs: Generate 3D worlds from a single image

Don’t know the exact details but I imagine the further from the original input image the more the system needs to make up stuff. Same why generative video models are limited to a few seconds. It will improve

dmarcoson Javier Milei: "My contempt for the state is infinite"

Is not Milei in Argentina the first self-identified libertarian president in history? Where has libertarianism failed?

dmarcoson Relativty: An open-source VR headset for $200

6DOf not only necessary for room scale. Lack of parallax of 3DOF a common cause of discomfort for many. I’ve been in the space for a decade and given hundreds of demos to people.

dmarcoson StabilityAI releases Stable Diffusion 3.5

The video above is not easy. AI is automation. A lot of what required a large team and budget before can be done by individuals or small teams. To do something good still requires skill, talent and dedication. The bar of what it's noteworthy raises. Creators that can harness the new tools in unique ways will create the new Seinfields and Pokemons. I'm excited for example that 3D animation movies a la Pixar won't be exclusively produced by multi-billion dollar companies. Small studios or even individuals can participate and the lower costs will allow for much more creative risk.

dmarcoson StabilityAI releases Stable Diffusion 3.5

Fair if you don't like it. I find it interesting, enjoyable and many other people too. The author would not have been able to create something like that without AI. We're in year two and stuff only gonna get better.

dmarcoson StabilityAI releases Stable Diffusion 3.5

To make the most of AI tools you need to understand how they work. Just prompting to me is just like using clipart or stock images. For interesting stuff people are doing search for ComfyUI tutorials on YouTube. It's a skill as like in any other digital art tool.

dmarcoson StabilityAI releases Stable Diffusion 3.5

“Awed” is a big word. I’ve been awed by art only a handful of times in my life. I start to see people using AI in ways I find enjoyable and interesting.

Below my fav video series that heavily uses AI (images, video, lip sync, sound). I enjoy it and I think it has artistic merit

https://x.com/dmarcos/status/1848835003079397646

It’s done by a single person. Individuals will be able to make stuff that needed teams and large budgets before. It will allow for much more creative risks.

dmarcoson StabilityAI releases Stable Diffusion 3.5

AI art is just art. Most is bad (same as without AI). AI is just another tool in the toolbox. Shared link below in a different comment. My fav video series that heavily uses AI (images, video, lip sync, sound). I enjoy it and I think it has artistic merit

https://x.com/dmarcos/status/1848835003079397646

Even with AI, one needs talent and skill to create something noteworthy. Just the standard of what’s compelling will change. I think people capable to adapt to the new tech will do fine creatively and financially.

dmarcoson StabilityAI releases Stable Diffusion 3.5

AI is just a tool in the box for humans to take advantage of. You can do bad and good stuff with it (most art is bad regardless of the tool). This is my fav video series that heavily uses AI tools (images, video, audio, lip sync):

https://x.com/neuralviz/status/1848176393282326595

I enjoy it and I think has tons of artistic merit

dmarcoson Famous AI Artist Says He's Losing Millions from People Stealing His Work

Look for ComfyUI tutorials on YouTube. Creating something unique is elaborate and requires developing skills no different than Photoshop or any other digital art tool.

dmarcoson Orion, our first true augmented reality glasses

Ctrl-labs bought Myo band IP from North (formerly Thalmic)

https://techcrunch.com/2019/06/27/ctrl-labs-scoops-up-myo-ar...

dmarcoson Orion, Our First True Augmented Reality Glasses

No true hologram afaik. They mentioned waveguides. Looks same tech lineage than hololens / magic leap.

I think Microsoft was first using holographic buzz word for these non-holographic displays

https://www.microsoft.com/en-us/hololens

“holographic device”