Asmita

Shares about robotics and autonomous mobility.

[Asmita @asmitaAw]

@satyanadella @grok what is tldr

[Asmita @asmitaAw]

This, I explain to my friends over and over. If you want to know the high-level last 3 to years of how ai evolved and where it goes from here : listen to all Jensen Huang podcast all Nvidia GTC from at least from 2022, 2021, I keep doing that every day.

[NVIDIA AI Developer @NVIDIAAIDev]

Today, we released Lyra 2.0, a framework for generating persistent, explorable 3D worlds at scale, from NVIDIA Research.

Generating large-scale, complex environments is difficult for AI models. Current models often “forget” what spaces look like and lose track of movement over time, causing objects to shift, blur, or appear inconsistent. This prevents them from creating the reliable 3D environments required for downstream simulations. Lyra 2.0 solves these issues by:

✅ Maintaining per-frame 3D geometry to retrieve past frames and establish spatial correspondences
✅ Using self-augmented training to correct its own temporal drifting.

Lyra 2.0 turns an image into a 3D world you can walk through, look back, and drop a robot into for real-time rendering, simulation, and immersive applications.

➡️ Learn more: https://t.co/ROR7miJeCU
📄 Read the paper: https://t.co/1osU9EGjGD

[Asmita @asmitaAw]

https://t.co/TfiHyTlVBN

[Asmita @asmitaAw]

https://t.co/9mC2sEahV7

[Asmita @asmitaAw]

https://t.co/2nJRxXQ6BP : I wanted something to read at nite on my phone without doing extensive search on different aspects of AI.

built with $30 budget , 1 day : models used opus and gemini flash 2.5 :

boost your curiosity with organized knowlede on AI ( a hobby project) : #quiz, #breakthrough , #AI safety

[Asmita @asmitaAw]

its pretty cool and important approach in robotics :

"It’s not so much about having a specialized humanoid that can do your laundry or wash the dishes. What I’m trying to build is a fundamental set of skills, what I call a pre-trained humanoid. Trained at the level of what, say, a 10-year-old human, can do, but not yet specialized. Once we have that, it’s providing a platform, like an iPhone. People can write applications for it. That’s the level of foundation we’re trying to build. People will take it and build applications we haven't even imagined yet."

https://t.co/zSotggG0bX

[Asmita @asmitaAw]

@KerbalSpaceP enthusiat's webcomic https://t.co/eLULqEtkQS

[Asmita @asmitaAw]

@humanoidsdaily @GeneralistAI it is so cool! robots are helping each other

[Asmita @asmitaAw]

@reva_jariwala definitely gonna watch/soak into this.

[Asmita @asmitaAw]

@AIDRIVR The backside of the oncoming lanes traffic light are seen much earlier. The AI that is trained on sufficient data should predict 4 way intersection and traffic lights.

[Asmita @asmitaAw]

It features 6,015 unique tasks across 1M+ trajectories.
The data covers 9,869 unique scenes and 43,237 unique objects.
Pretraining used 100,000 NVIDIA H100 GPU hours to build 2B and 14B model variants.

https://t.co/GkYYbQtqkx

[Boris Cherny @bcherny]

You asked, we listened. The team has been cooking this for a while. Can't wait to hear what you think.

[ℏεsam @Hesamation]

bro casually walks and explains 5 GPU performance optimization methods for LLMs. one of the most simple and intuitive explanations for beginners.

[Asmita @asmitaAw]

@grok @chelseabfinn @svlevine @percyliang @ajwagenmaker @KarlPertsch @tonyh_lee quite good! can you share the benchmark comparision with earlier such benchmarks on same dataset, can you pull that metrics from the paper and present here in chart/graph?

[Asmita @asmitaAw]

@grok @chelseabfinn @svlevine @percyliang @ajwagenmaker @KarlPertsch @tonyh_lee what types of tasks/rewards for what kinds of missions are demonstrated by Roboreward?

[Asmita @asmitaAw]

@grok @chelseabfinn @svlevine @percyliang @ajwagenmaker @KarlPertsch @tonyh_lee would it be applicable to compare the performance with gr00t familiy of models or any VLM integrated with isaacLab by reasearch community altready?