Anusha Avyukt, PhD

❓Decisions under Uncertainty 🎯 Stats and RL 🪙Tokenomics 🧬DeSci, FinTech 🚀Blue Oceans and 🌻Aha! moments✨

[Delulu Labs @delululabs_]

Playing with Seedance 2.0 and it's next-level fun.
Making movies just got a whole lot easier.
All that matters now are soul and taste.

Oneshotted this previously complex shot. Prompt: Animate the frame with the kid running fast toward the camera as the camera pulls back faster to reveal the night sky with twinkles and yellow/gold illuminated pathway.

[Anusha Avyukt, PhD @atomsha42]

@minmsme Applied and made the payment on 7th September for MSME Udyam certificate. No further update. Where and when do I get the certificate?

[Ross Harkness @THEROSSHARKNESS]

It’s supposed to be hard.

[Jim Fan @DrJimFan]

Not every foundation model needs to be gigantic. We trained a 1.5M-parameter neural network to control the body of a humanoid robot. It takes a lot of subconscious processing for us humans to walk, maintain balance, and maneuver our arms and legs into desired positions. We capture this “subconsciousness” in HOVER, a single model that learns how to coordinate the motors of a humanoid robot to support locomotion and manipulation.

We trained HOVER in NVIDIA Isaac, a GPU-powered simulation suite that accelerates physics by 10,000x faster than real time. To put the number in perspective, the robots undergo 1 year of intense training in a virtual “dojo”, but take only ~50 minutes of wall clock time on one GPU card. The neural net then transfers zero-shot to the real world without finetuning.

HOVER can be *prompted* for various types of high-level motion instructions that we call “control modes”. To name a few:

- Head and hand poses: can be captured by XR devices like Apple Vision Pro.
- Whole-body poses: via MoCap or RGB camera.
- Whole-body joint angles: Exoskeleton.
- Root velocity command: Joysticks.

What HOVER enables:
- A unified interface for us to control the robot using whichever input devices are convenient at hand.
- An easier way to collect whole-body teleoperation data for training.
- An upstream Vision-Language-Action model to provide motion instructions, which HOVER translates to low-level motor signals at high frequency.

HOVER supports any humanoid that can be simulated in Isaac. Bring your own robot, and watch it come to life!

It's a big teamwork from NVIDIA GEAR Lab and collaborators: 🧵

[Sam Altman @sama]

big congrats!!

[OpenAI @OpenAI]

🌐ChatGPT search 🌐is starting to roll out to all Free users today.

Search the web in a faster, better way—available globally on https://t.co/nYW5KO1aIg and our mobile and desktop apps for all logged-in users.

[Sam Altman @sama]

you can just do things

[Josep Ferrer @iamjosepferrer]

The Transformer's encoder clearly explained 👇🏻

[The News Agents @TheNewsAgents]

"The truth is rare, fantasy and lies are cheap. If we flood the world with information, the truth will sink to the bottom and not rise to the top."

How do we navigate a world rife with misinformation?

Author of Nexus @harari_yuval talks to @maitlis.

[Yuval Noah Harari @harari_yuval]

Who controls your information?

Tiberius moved to Capri for safety, but Sejanus used the isolation to control all communication and rule Rome. In our digital age, information gatekeepers hold the same power.
#RomanHistory #NexusBook #DigitalAge #Information

[Yuval Noah Harari @harari_yuval]

Who holds more power: the fact-checkers or the storytellers?

History proves that stories, not just facts, shape societies and inspire action.

Watch the full interview from the 18th of October: https://t.co/QSmWuqFeLt

#BigThink #Storytelling

[Yuval Noah Harari @harari_yuval]

Are we solving the right problems?
Find out why solving one problem after another doesn’t always improve the situation. Are we focusing on the wrong issues? Watch the full interview here: https://t.co/sKshdnI0wq

#NexusBook #JordanHarbingerShow @JordanHarbinger

[Vlad @VladTheInflator]

Steve Jobs talks about how worthless consultants are

[Jordan Noone 🇺🇸 @jordannoone]

There are two types of people in the world

[BURKOV @burkov]

Give me a dataset large enough and a GPU to finetune an LLM on and I will beat your stupid benchmark.

[Jiaxuan You @youjiaxuan]

🚀Excited to share our latest paper, *ResearchTown*, a framework for research community simulation using LLMs and graphs
🌟We hope this marks a new step toward automated research
🧠Paper: https://t.co/xz18Kpc6q5
💻Code: https://t.co/y1Dn9A0LQZ
🤗Dataset: https://t.co/UKOsZqOFgc

[Anusha Avyukt, PhD @atomsha42]

Good social media is good sociology.

Someone said this, and I am curious what sociology can tell us about what makes some seemingly simple posts viral and how some communities growing and build loyalty.

[Anusha Avyukt, PhD @atomsha42]

violin plots should not exist. https://t.co/IwR7Mjw9Sr via @YouTube

[Divya Mittal @divyamittal_IAS]

The lesson?

Sometimes, the key to transformation lies in seeing the potential in what you already have and betting on that transformation. For NVIDIA, that meant turning gaming GPUs into the engines of the AI revolution.

[Divya Mittal @divyamittal_IAS]

What an amazing run by the Indian relay team as they qualify for #Olympics! Wouldn’t tell the result else it kills the excitement..watch it yourself..

Looking forward and cheering for them to get the Gold medal in Paris