Amir Rubin

@carltongibson Not reduce(lambda a,b: a^b, bools)

@Poyonoz @Scobleizer Exactly, I don’t see this as LiDAR market shrinking, more like the utility and use cases for low cost camera-only solutions will grow faster and outpace LiDAR adoption

@AmirRubin @Scobleizer But I guess that's a low enough amount of uses to not matter for most products right?

Everyone who bet on LIDAR is wrong. Everyone who did will have to admit their billions spent is lost and will need to start over. I feel bad but this is what is about to happen.

@Poyonoz @Scobleizer … but End2end deep learning reconstructions will enable cameras to perform well, even in low light, for a much larger emerging class of use cases (ie construction monitoring, facility documentation, digital twin creation, infrastructure inspection)

@Poyonoz @Scobleizer Vegetation is a use case where LiDAR will always be better , especially to penetrate the canopy and capture the ground 😎🌲 Pitch dark environments like underground mining operations same story, the active LiDAR is required… Same for bathymetric and boundary surveys…

@vikasreddy @barmijo @Scobleizer Andrej Karpathy addressed that in his CVPR talk - the latency from LiDAR is slower than camera+IMU input which messes up their models and ground truth labeling m.youtube.com/watch?v=g6bOwQ…

@barmijo @Scobleizer That sounds more like a limitation of their software / ML systems than an actual fundamental limitation. Unless the input is totally random noise, logically any additional information/input should only help (or at the very least not hurt).

@quantombone @spillai @m_mokros @GiseopK @HanmeKim The results from iPad LiDAR are really impressive! I’m very curious about the multicam system

@spillai @m_mokros @GiseopK @HanmeKim @AmirRubin check this out

📢New paper out!!! Is it possible to use the #Apple #LiDAR in #iPad or #iPhone12 Pro to acquire point clouds of #forest sites and measure the DBH accurately? Is our multi-camera #prototype reliable to generate data in mobile mode? Check it out here: sciencedirect.com/science/articl…

3D Reconstruction from public webcams arxiv.org/abs/2108.09476 Hey @AmirRubin, check out this 3D #computervision project which uses SuperGlue, a deep feature-based matcher—extremely useful for creating a robust #digitaltwin of outdoor spaces. Thanks @ducha_aiki for sharing!

@quantombone @ducha_aiki I love this! SuperGlue and deep front ends are game changers for 3D reconstruction. The robustness of learned features+homography is bringing digital twin creation into the hands of non-expert users 🤓 I mean, I’m not talking metaverse scale just yet but who knows? 😎😎

@jackmorrison @quantombone Details coming soon 😎 And next time I'm back in CO let's hit the slopes 🏂!!

@AmirRubin @quantombone Looking forward to your next big thing. Come back and visit us in CO sometime!

I had a great chat this past weekend with @AmirRubin about the intersection of 3D #computervision and #digitaltwins. There are many intriguing ideas in this space, especially for me, who spent five years working on SLaM for #MixedReality.

@quantombone Anyone interested in chatting more please reach out/DM me!

@quantombone We are at the very beginning of the impact 3D mapping+ SLAM will have on the future; Metaverse concepts will transform how we play and interact, but Digital Twins are the foundation for transforming industry!

@jackmorrison @quantombone Ha you are too kind! And huge congrats again on the Scythe launch!!

@quantombone @AmirRubin @AmirRubin is the man!

@lifemath_ 1…e5 !

I would totally believe #NikhilKamath has not used any unfair practices to build up his portfolio #sarcasm #Cheaters

@AjdDavison I’ve read hypotheses that certain molecules in bird eyes become quantum entangled with earth’s magnetic fields allowing them to navigate and stabilize relative to the magnetic field - very strange!

Can anyone from explain to me how a (kestrel?) is able to achieve this stabilisation accuracy? It is high; ground not that textured; effectively monocular vision; why is rot/trans ambiguity not higher? Of course it has inertial and other cues too but still just surprised/amazed.

@vikasreddy Congrats Vikas! The demo is great - this is an exciting idea that tons of creators are going to love!

I'm excited to introduce the new company I started: LightTwist, a virtual video studio on the iPhone for creators and creatives! blog.lighttwist.io/p/introducing-…

@quantombone The land use of our neighborhoods and cities are dominated by cars, which waste precious space when parked, and are dangerous to people when driving - We need a new infrastructure that allows neighborhoods to reclaim streets for families!

What would a city/town/village without cars feel like? Would a car-free lifestyle enhance our lives? Would our communities be strengthened? #daydreaming

@ducha_aiki @quantombone This is interesting- there still needs to be a reliable and effective infrastructure that doesn’t rely on cars!

@quantombone Ukraine. The place is not terrible, it is nice. But without a car it would be much harder. Even with a boom of delivery services, public transport is not good.

@WTEric @FBRealityLabs Congrats Eric!

AR/VR Family - excited to announce that I’m joining @FBRealityLabs as a Business Development Manager, working on Enterprise Platform Partnerships (Oculus & Portal). #remotework #WorkFromHome #ar #vr #xr

@edersantana @quantombone Ha! It is time to make NeRF fields ‘op’ for robots (is that a good LoL joke?) We’re bring those nerf guns back😎

@quantombone I wonder what my friend @AmirRubin would say about NeRF. Specially given that we used to have actual nerf guns back in Paracosm xD

With this new wave of NeRF-related research, I’m starting to daydream about what this means for real-time mapping and autonomous navigation! 3D #computervision has always been around, but NeRF is making 3D cool again. Time for a startup involving NeRF-like methods for robots?