Synthetic Data | World Models | Radiance Fields | Computer Vision | GenAI | Chief Evangelist at @LightwheelAI
Salem, OR
Joined February 2009
- Tweets11.8K
- Following1.8K
- Followers13.7K
- Likes7.2K
I spend all week sitting at a laptop. On the weekend, I love getting to hang out with my daughter and share our love for horses.
The comments on this are EPIC!
I always feel bad when a company launches something and gets eclipsed when a competitor outdoes them that same day.
Porsche will be fine, people who buy Porsche are not the same people as those who buy Tesla.
Whoa! That was a cool trick! Almost looked like it was a visual generated from a world model. It just didn't add up. Way to go Tesla! It's always a show.
Born before 1985 - Used 5.25" floppy "A" drive
Born before 1995 - Used 3.5" floppy "B" drive
Born after 1995 - Never used a floppy
This guy is clearly too young to have used floppy discs.
Keep an eye on this new scanner. It’s literally the biggest competition to XGRIDS so far. No pricing yet, but based on the hardware and the current market, it will be priced for the professional market… sorry hobbyists, this is a serious tool.
New gaussian splatting camera from @KosmoAIOfficial. I've been using it for a few months and the captures coming out are consistently epic.
Wow, the details and consistency of this scene are amazing. But even more amazing is showing the background notes. The depth map shows you a bit more about what is going on behind this output.
This is obviously old footage. All Figure 02 models jumped into a cauldron of molten steel 2 weeks ago.
He has a point. I gave my Google login info to my Muse agent so it can log in for me. Muse can also monitor my MacBook messages so it can handle the 2FA for me.
I’m playing chess while everyone is is still playing checkers.
Can someone explain to me why it’s easier to train a robot to do this than it is to train a robot to pick up a pen off the floor?
An intriguing new world model and they are focused on Real2Sim2Real and Physical AI 😮
Hey friends at Gracia how do I get access? I would love to do an eval video.
I’ve always been intrigued by VPS, but I always assumed it would only work when a company had a mass amount of imagery of the location. Niantic Spatial pulled it off thanks to Pokémon Go
This could be very very interesting for setting up high quality real2sim environments. My only question is are all of those objects SimReady or just look good?
Moving Atoms (YC S26) : Scan to BIM is here:
Send a LiDAR or SLAM scan from any scanner. Within 48 hours every object comes back separated, measured and editable, in IFC, Revit, STEP, USD or GLB: ready for your BIM team, your digital twin.
Who this is for:
-> Data Center Requiring a Digital Twin
→ Manufacturing floors and production lines
→ Warehouses and logistics / fulfilment centres
→ Wet labs and scientific facilities
→ Commercial kitchens and food production
→ Any site where you plan to put a robot and want to test it there first
🤯 a year ago I would say photogrammetry will always be better than 3DGS to mesh. But now I might have to talk that back.
I still think lidar/photogrammetry fusion is the ultimate winner though.
Thoughts?
Unpopular opinion: Waymo drives better than humans.
Here’s my first ride. The car last minute ditched from a wrong lane, missed the turn into my destination. Both moves cut several minutes off the ride time 😆
Super important work. Robots will fall. It shouldn’t be catastrophic when they do. I’ve seen this a Disney doing research in the same area. How can it fall and not break? Disney goes a step further. They make the fall look whimsical.
This paper visualised: "Multiple contact planning for minimizing damage of humanoid falls" (Ha & Liu, IROS 2015) turns falling into a planning problem.
- Model fall as a chain of inverted pendulums
- Solve for contact order with dynamic programming
- Minimise largest single impact, not the total
- Fall like a judoka!
2 reaction to Argon for Robotics:
1. It’s clear these large models that are taking advantage of scaling laws are going to change the robotics industry.
2. Not impressed, cornhole bags must be thrown from 27 feet away or no points. Rules are rules.