📣 New research from GenAI at Meta, introducing Meta 3D Gen: A new system for end-to-end generation of 3D assets from text in <1min.
Meta 3D Gen is a new combined AI system that can generate high-quality 3D assets, with both high-resolution textures and material maps end-to-end, producing results that are superior to existing solutions — at 3-10x the speed of existing work in this space.
Details in the technical report ➡️ https://t.co/lrkShaRCju
Robots on solar farms! ☀️
These Luminous robots assist in constructing solar farms, minimizing the need for manual handling of heavy parts.
To reduce transport damage, module manufacturers are now using a "vertically stacked" standard, but this makes automation more complex.
But these robots can still handle them! 🦾
Since large-format modules are heavy and long, robotic arms must handle the weight, balance the load, and move safely while avoiding collisions and external forces like wind.
Using monkey brain data to operate robotic arms! 🙉
Project Jenkins explores brain-computer interfaces by decoding neural activity into robotic movement and generating synthetic brain data from motion patterns.
Using neural recordings from a macaque monkey named Jenkins, researchers developed models to decode brain signals into robotic arm movements and encode movement data into simulated neural activity.
The project utilizes robotic arms and an interactive web console for real-time brain data generation from joystick inputs.
The research paves the way for adaptive brain-computer interfaces by modeling brain activity more flexibly.
Congrats to authors @yanovskyd, and @anzahorodnii! 🦾
Robotics and CV! 👓
Computer vision gives the robots a sense of vision and additional intelligence.
Pralines are packed by a delta robot using computer vision.
Packing chocolate pralines into a box with the delta robot. Unlike most robots, this one has a fixed TCP, and the base moves in response to vision data.
To adjust the base, the vision system determines exactly where the praline is located on the conveyor belt. 🍫
Using this application, the system knows where each praline should be placed based on its orientation.
A true engineering showcase!
I just gave a GDC talk about what's going on with the Quest Ecosystem and where we think it's going next. If you couldn't make it, don't worry! It's back! In blog form. https://t.co/m9Skdpnyf3
The first humanoid robot priced under $20k was built in China: for just $14k, it can do VR teleoperation, imitation learning, and connect with DeepSeek and VLM for specific planning tasks. You can imagine it at an assembly line workbench.
We're closer than ever to a future where we remap the earth 24/7 with a plurality of sensors -- phones, drones, cars, planes, satellites, and oh my!
This physical 3D model of the world serves as the connective tissue between the world of bits & atoms.
It's Tesla cars today, and AR glasses tomorrow. It's social media in our pockets, and ambient computing in our homes.
The end result is a 1:1 digital twin of our planet and all activity in it.
A perfect fusion of SIGINT + GEOINT to use intelligence parlance.
Wow. Recreating the Shawshank Redemption prison in 3D from a single video, in real time (!)
Just read the MASt3R-SLAM paper and it's pretty neat. These folks basically built a real-time dense SLAM system on top of MASt3R, which is a transformer-based neural network that can do 3d reconstruction and localization from uncalibrated image pairs.
The cool part is they don't need a fixed camera model -- it just works with arbitrary cameras -- think different focal lengths, sensor sizes, even handling zooming in video (FMV drone video anyone?!). If you've done photogrammetry or played with NeRFs you know that is a HUGE deal.
They've solved some tricky problems like efficient point matching and tracking, plus they've figured out how to fuse point clouds and handle loop closures in real-time.
Their system runs at about 15 FPS on a 4090 and produces both camera poses and dense geometry. When they know the camera calibration, they get SOTA results across several benchmarks, but even without calibration, they still perform well.
What's interesting is the approach -- most recent SLAM work has built on DROID-SLAM's architecture, but these folks went a different direction by leveraging a strong 3D reconstruction prior. Seems to give them more coherent geometry, which makes sense since that's what MASt3R was designed for.
For anyone who cares about monocular SLAM and 3D reconstruction, this feels like a significant step toward plug-and-play dense SLAM without calibration headaches -- perfect for drones, robots, AR/VR -- the works!
LV1 MAPP is heading to the Moon! A versatile space mobility platform, MAPP will be the 1st commercial rover on the Moon. Equipped with innovative technologies, MAPP is ready to accomplish mission objectives. Learn more: https://t.co/uXoJFz70eN #TheNextLeap
Last week we launched a $50M Meta Horizon Worlds Creator Fund alongside early access to advanced creation tools to help creators unlock their creativity and get rewarded. ✨ What excites you most about building for Horizon Worlds and joining the Meta Horizon Creator Program?
Several people have confirmed that Valve is aiming to release new standalone, wireless VR headset (codename Deckard) by the end of 2025. The current price for the full bundle is set to be $1200. Including some "in-house" games (or demos) that are already done. Valve want to give the user the best possible experience without cutting any costs. Even at the current price, it will be sold at a loss. A few months ago, we saw leaked models of controllers (codename Roy) in the SteamVR update. It will be using the same SteamOS from Steam Deck, but adapted for virtual reality. One of the core features is the ability to play flat-screen game that are already playable on Steam Deck, but in VR on a big screen without a PC. The first behind closed doors presentations could start soon.
Always envisioned that a compact and versatile javascript 3d engine was going to be a part of the puzzle in the future.
Didn't expect that the AI was going to be the actual user though.