Nvidia's NEW 'AI NEURALANGELO Takes the Industry By STORM! (NOW ANNOUNCED!) 

this is exactly what google maps needs. imagine truly traveling through maps in 3d in the highest quality possible.

Not only do you not write your subtitles yourself, you don't even check the generated ones. Irritating mistakes in it.

"What you are seeing on your screen right now, is what we call - Maps." Every 3d artist: NORMAL MAPS, ITS GOD DAMN NORMAL MAPS, NOT JUST MAPS.

This is a technology that could be easily implemented in Google Street View.

"This year one of the biggest AI announcements" in an absolutely crazy year already

1:27 - No, its not extracted, it's guessed by the AI. If you have a scar on your back, and take a photo from the front, the scar won't appear on the 3D model. Also, photogrammetry already existed for 10+ years. This is basically that technology with guessing algorithm for obscured details.

It's like a mix of photogrammetry and AI. Love stuff like nerfs and this capturing all that material quality...so awesome! The detail is really good with this...especially things like capturing and reproducing reflections from different angles like around 5:08!

this pretty much means you'll be able to add any real location to any video game you want depending on how long it takes for these to work you may also be able to start recreating full 3d environments based on sonar, drone footage, or even just walk through footage someone could take use case: imagine using this to create a 3d environment you could explore on mars based on the rover footage, a space simulator based on data gathered from the hubble and james webb telescopes, or a 3d exploration of a town you'd like to visit and you want to make sure you know how to get somewhere by walking a digital representation of were various shops or other landmark locations would be

The applications for this are very interesting. Got a 3D printer? Grab an object you saw in a RU-vid review... BAM, you have the model for it. I'm sure that will require some refinement and conversion, but that will happen. In fact, I can see RU-vid being flooded with "parts videos". People search for a "6-port 3/8" OD air line manifold", and there's a video of one, which is all you need to print the object. And then likely a comments section filled with variations and modifications to suit any need. And you don't need to be a 3D designer to produce that content, just start making videos of common objects from several angles and you're good to go. Copyright will get interesting, lol. Or 3D body scanners. These are very useful for health and fitness, but presently are VERY expensive. With software like this, you don't need to spend $12,000, you could do it with a cell phone, or a camera built into a mirror. And with more AI to analyze that data, it could track body fat, suspicious skin blemishes that should be checked out, detect subtle things like enlarged lymph nodes, etc, long before you would be able to detect it with the naked eye. I bet LEGO will love this, lol... Scan an object with a phone video, and here are the exact blocks you need to order to build a lego model of it. Click here to buy now. Combine with a drone and property surveys get really cool. 3D realestate walkthroughs online will be common. Want to see what that couch looks like in your house? Well, you have a 3D model of that house to scale, so dropping in a virtual object can be done in seconds. The lack of 3d Assets is what has held VR back since it's inception, and simple stereophotos are not enough to create the immersion people are looking for, and many VR environments go with very simplified geometry, almost cartoonish looking. A 360° stereoscopic camera now gives you a live, real-time virtual presence. You're not just watching a 3D video, but are able to move around in the virtual space. For the outdoorsman, trail cameras, game cameras, and even fish-finders could be taken to a whole new level. Imagine having a detailed map of the bottom of a lake and being able to see an actual fish in full 3d, to scale, before even dropping a line in the water. The adult industry will have a lot of fun with this as well, I'm sure. No more yelling "Stupid Camera man... damn it lady, move your elbow", lol. Just rotate the scene for a better view. But the coolest thing isn't even the direct application of this technology, but rather it's use to generate additional training data for other AIs. Imagine a physics AI that is focused on how real objects interact, having access to full 3D models to build out simulated data and learn. If a picture is worth a thousand words, then a 2D video is 30,000 words per second, and a 3D model that can be viewed from any angle is exponentially more, and no harder to create than taking a photo or video is now. The sheer amount of useful data generated will grow exponentially with virtually no additional cost or effort. And training data has been one of the biggest AI bottlenecks so far. It's not even about what this tech can do, but the other technologies it can unlock down the road. Things we can't even imagine yet.

Outstanding video! But you can't compare Luma Aí with Neurangelo, because Luma AI is definitely using Instant Nerf open source by Nvidia too 👍

@TheAIGRID I see multiple times constructive feedback over different videos on the subtitling tool you use. Genrally people seem to find it annoying and it often contains mistakes, which distracts from the good content you produce. What is your take on that and is this something you are looking to improve? I can't speak for others, but for me it is a reason to hold off from subscribing to your channel. Curious to hear your take on this! Thanks in advance.

You've done a lot of research and I really really want to watch this channel. The consistent flashing of the two to three word captioning/subtitling distracts from the important information that you are providing. I plan to continue to watch, but maybe put more words on the screen at one time?

Essentially a great video again. Essentially I love your essential breakdown and analysis. Can't help myself, I essentially have to make a bit of fun or your essential choice of words

This is great! If they can generate normal maps with low poly, that would be killer

Friendly suggestion if I may: Watch your video at least once and look for misspellings if you are using AI to transcribe it. The mistakes are kinda devaluing the hours of work you probably put into making this.

As soon as this can be done within milliseconds, we will be able to walk around in vr movies. This will also nicely combine with AI video security.

How can one use it, or is it just research papers only?

This is awesome ! Ai 2d to 3d is the future

Very interesting video; would be even better if you proofread the text on screen.

❤sweet❤ Bet you still have to tough up the maps and textures.

Hey, dont take me wrong here, but I've been a dedicated viewer of your content for quite some time now and I absolutely adore it. However, I have a small suggestion regarding the titles of your videos. Most the time, when you give them grandiose names, it creates an expectation for something groundbreaking, but the content ends up following a familiar format. Cheers

Lol… ‘what are maps? Well, it’s simple, essentially when you want to use less ram or less compute that your computer has to do, you of course use normal maps.’ Amazing how little some RU-vidrs know about subjects.

WHEN do we get the tools?

_Dear Ai, help me build a time machine with household items..._ _...Soo I can go back when Ai doesn't exist lol_

Finally, still more

jaw dropping

Do the cameras used for this data have Lidar and Radar technology behind the imagery?

I was waiting for this

Did they say anything about releasing the source code?

is the code available for this ?

Still doesn’t justify its x200 PE

Not available on windows, just linux+ does not extrac texture /colors

Luna AI is only available on iOS. Let me know when it hits Android, or better yet PC.

When the tool is so good, that you can actually give it a crap name ^^

This is just basicaly photogrammetry superpowered by AI.

nerfs seem to work a hundred times better with shiny or transparent objects than this technology

New drinking game: Drink everytime he says essentialy or actualy

Watch the movie "TinTin". The rapid swirling camera perspective "tricks" the brain into perceiving a 3-dimensional moving scene.

Did you just say "softwares"?

Can’t wait till this goes to DID.

I don't like how you prepare titles and the video itself. The content is ok tough. Just don't use buzz titles like by storm and stop using text on black screen. Don't show the same scene over and over again but rather analyze what's good and bad about

Please dont get me wrong, first and fore most I appreciate your video and all the effort put into it but for future improvements, I would say, you dont need to "waste time" with explaining before every single new info why its there. As in, ah I added this because this is also interesting, specially if you dont mention why it is interesting... so to me when I click on your video, everything related to the topic is interesting and when I have to listen before each section about how another section is added is a little bit unnecessary. This is the second or third video I watch from you and in all of them you did the same.

😮😮😮

Oh goodie .. we can now amuse ourselves more .. the world SO needed this !!

I think Nvidia will change name to >>> SKYNET

impressive but it's not the real 3D model, there is no way from one 2D perspective to get it, the AI is guessing a model that looks good in the original point of view, when moving to other point of view the difference will show. if what you want is a game NPC is does not really matter, but if you want a true 3D model of your face, it will be wrong.

Rip 3D environment artist job

why are you talking about the same thing 5 times over? then a bit later you talk again about the very same stuff, longer videos to put more ads in?

*exhales into the microphone*.. how do you like it

*low quality

Check your mic settings, your voice is clipping. Very annoying, please use a limiter!

Second🎉

nvidia smoke and mirrors

First 🎉

Great, more distractions from our reality.........

so what.