
CsUtil.com
@csutil_com
Software Engineer working on #opensource in my free time. See the full project list at http://csutil.com
You might like
Google Maps v2.0? 🌍 Here's @PlayCanvas streaming a scene with 2 BILLION Gaussians! 💪 Coming sooooon!
gm Native Project, Faulty Topography Theme, Frag 005 Expressing both Optics Logic and Phantom 3D gestures Collected by overflowing.eth Via @fellowshiptrust Daily Program Season 2
PlayCanvas Engine v2.12.0 is here 🚀 Streamed LOD system for 3D Gaussian Splats enters beta. Supports the streaming of 100s of millions of Gaussians! Release Notes: github.com/playcanvas/eng…
Why do most humanoid motion trackers still slip, skate, or clip through objects? Because the real bottleneck isn’t RL, it’s bad retargeting… OmniRetarget flips the script: instead of compensating with dozens of reward terms, it generates high-quality, interaction-preserving…
We're excited to see so many of you creating new worlds in our Marble beta preview! Here's a tour through some of your latest creations… To join the fun, sign up at worldlabs.ai/waitlist — we're letting in new users as fast as we can 🔥
Introducing Alterego: the world’s first near-telepathic wearable that enables silent communication at the speed of thought. Alterego makes AI an extension of the human mind. We’ve made several breakthroughs since our work started at MIT. We’re announcing those today.
Today we're unveiling F.02 loading the dishwasher autonomously
Ha. The fact that this works is just great.



Robots don’t wait. So why should your model? Large VLAs run in real time… with no training-time changes. ❗️Worth reading if you’re working on real-world deployment of large models in robotics. I found this write-up on Real-Time Action Chunking (RTC) from Physical Intelligence,…
LichtFeld Studio in Action! GPLv3 - no subscriptions, no data scraping, forever free! - Trains in 20 minutes at 4k for 60k steps with 2m Gaussians and MCMC densification. - With enabled bilateral guidance (CUDA version), it totals 60 minutes. - Pretty much floater free. Makes…
Business Insider: Tesla has moved away from motion-capture suits and VR headsets for Optimus data collection, insiders say, and will now “primarily focus on recording videos of workers performing tasks to teach the robot how to do things like pick up an object or fold a t-shirt.”…
Nano Banana: Make Image Daytime and Isometric (Building Only)


Now I’m testing Nano Banana with product replacement. even with product photos that have complex patterns nano banana can still match them perfectly. On average it only takes me 2–3 tries to get a solid result. Well played base image : @letz_ai




I'll leave you with two demos I had the opportunity to see live - a cat exploring this immaculate 3d world, and this slightly bigger cat being photographed by a "wild life photographer." On-the-fly interactive media and gaming is pretty damn close!
Another 3DGUT scene trained on @insta360 X5 fisheyes, testing the PureVideo mode for low light. It did some weird stuff to the highlights, but still pretty good for how dark it was in there. Barbican Centre, London. @NVIDIAAIDev #NVIDIA3DGUT #NVIDIASweepstakes #3DGS
We do stand at the edge of a pronounced crisis of identity, but, only in the fact that it will happen simultaneously for so many. This has happened again and again in small ways to you, and to everyone ever. In my philosophy undergrad, I was lucky enough to have a prof named…

vibe coding giveth vibe coding taketh away
Creative upscaling ftw! Turn high altitude aerial photogrammetry into low altitude drone imagery. Such a cool way to frame the shots you want in Google Earth and use a FLUX Kontext LoRA to take it all the way.
"[video game] as a community theater production" may be one of the most delightful Veo 3 Fast prompts Please enjoy, in order: GTA, Pokemon, Mario Kart, The Witcher 3, Stardew Valley, Tetris, Mortal Kombat, The Sims, & Death Stranding(!) Yes, the whole prompt was the one above.
First, our results: small models trained with RULER+GRPO are more reliable than o3 on 4/4 tasks, despite being 1/20th the cost. Surprisingly, they even beat models trained with hand-crafted reward functions on 3/4 tasks.


United States Trends
- 1. Flacco 68.6K posts
- 2. Bengals 70K posts
- 3. Bengals 70K posts
- 4. Tomlin 18.1K posts
- 5. Ramsey 16.8K posts
- 6. Chase 101K posts
- 7. Chase 101K posts
- 8. #TNFonPrime 4,928 posts
- 9. #WhoDey 5,682 posts
- 10. Teryl Austin 2,123 posts
- 11. #HereWeGo 9,930 posts
- 12. #PITvsCIN 6,343 posts
- 13. Andrew Berry 2,564 posts
- 14. Max Scherzer 11.7K posts
- 15. DK Metcalf 3,711 posts
- 16. Cuomo 70.2K posts
- 17. Zac Taylor 2,297 posts
- 18. Ace Frehley 85.1K posts
- 19. Darnell Washington 2,395 posts
- 20. Cincy 3,830 posts
Something went wrong.
Something went wrong.