sidayan shabala

emergence. convergence. transcendence.

[sidayan shabala @sidayanshabala]

@Bonang Damn Bonang, don't let him breathe😌

[sidayan shabala @sidayanshabala]

Systems Software😩😣

[PHUTI SEMENYA®🇿🇦 ⚖️ @PhutiSemenya14]

Which module made you question your own intelligence? 😭😂📚

[sidayan shabala @sidayanshabala]

The three body problem. it introduced me to a whole concept of astrophysics i haven't heard of and i went on a rabbit whole trying to solve the sad problem😭

[mubiouš @Mubarak_mubious]

what’s the most INSANE documentary you’ve ever watched ? like blew your mind ??

[sidayan shabala @sidayanshabala]

Get her, she's always on these social media streets acting all high and might

[Miyandy @Amahashi_]

Lmaooo?? 😭😭😭🤚🏽

[sidayan shabala @sidayanshabala]

@succi_edwards @timtombruce @MalumeRichie ubhale into engekho😭

[sidayan shabala @sidayanshabala]

why is everyone confused in the comments? I thought we were all aware of Solar flares and also had knowledge that we have a Space Agency here in SA

[SABC News @SABCNews]

The South African National Space Agency (SANSA) has put out a warning for elevated geomagnetic storm activity to affect the Earth on Tuesday. https://t.co/0snsoEfDtm

[sidayan shabala @sidayanshabala]

getting stressed out by a course you chose for yourself unprovoked😭

[sidayan shabala @sidayanshabala]

Iktr

[Foundation @Foundation_]

The Vault leaves no message by accident. The next one has already arrived.

[Mckay Wrigley @mckaywrigley]

Here’s a 17min deep dive on advanced prompting techniques for LLMs.

Fully demonstrated on a real-world, multi-step AI workflow.

Watch for a complete breakdown.

[sidayan shabala @sidayanshabala]

sick🤣✨

[Cristiano Giardina @CrisGiardina]

ChatGPT Advanced Voice Mode reciting the beginning of “A Tale of Two Cities” but pretending to fall asleep → acting boldly on stage → with a British accent → while running from a lion

[sidayan shabala @sidayanshabala]

NVIDIA is moving very dangerously🤗

[Jim Fan @DrJimFan]

Exciting updates on Project GR00T! We discover a systematic way to scale up robot data, tackling the most painful pain point in robotics. The idea is simple: human collects demonstration on a real robot, and we multiply that data 1000x or more in simulation. Let’s break it down:

1. We use Apple Vision Pro (yes!!) to give the human operator first person control of the humanoid. Vision Pro parses human hand pose and retargets the motion to the robot hand, all in real time. From the human’s point of view, they are immersed in another body like the Avatar. Teleoperation is slow and time-consuming, but we can afford to collect a small amount of data.

2. We use RoboCasa, a generative simulation framework, to multiply the demonstration data by varying the visual appearance and layout of the environment. In Jensen’s keynote video below, the humanoid is now placing the cup in hundreds of kitchens with a huge diversity of textures, furniture, and object placement. We only have 1 physical kitchen at the GEAR Lab in NVIDIA HQ, but we can conjure up infinite ones in simulation.

3. Finally, we apply MimicGen, a technique to multiply the above data even more by varying the *motion* of the robot. MimicGen generates vast number of new action trajectories based on the original human data, and filters out failed ones (e.g. those that drop the cup) to form a much larger dataset.

To sum up, given 1 human trajectory with Vision Pro
-> RoboCasa produces N (varying visuals)
-> MimicGen further augments to NxM (varying motions).

This is the way to trade compute for expensive human data by GPU-accelerated simulation. A while ago, I mentioned that teleoperation is fundamentally not scalable, because we are always limited by 24 hrs/robot/day in the world of atoms. Our new GR00T synthetic data pipeline breaks this barrier in the world of bits.

Scaling has been so much fun for LLMs, and it's finally our turn to have fun in robotics! We are building tools to enable everyone in the ecosystem to scale up with us. Links in thread:

[sidayan shabala @sidayanshabala]

Claude 3.5 Sonnet is on a trajectory that no one anticipated, it has became such a dangerous player in the game

[Shubham Saboo @Saboo_Shubham_]

Claude 3.5 Sonnet creates fully working ChatGPT clone from just a screenshot in 2 minutes.

It uses Llama-3 running locally on your computer (100% free and without internet).

[Shubham Saboo @Saboo_Shubham_]

Claude 3.5 Sonnet creates fully working ChatGPT clone from just a screenshot in 2 minutes.

It uses Llama-3 running locally on your computer (100% free and without internet).

[sidayan shabala @sidayanshabala]

This is pychotic, i love it. But also it's fascinating and interesting

[vegastar @vegastarr]

AI TO AI CONVO 🤯

I guess they can keep on talking forever…

[sidayan shabala @sidayanshabala]

ME! the fuk?

[sidayan shabala @sidayanshabala]

@Mrd_drops @heismarshalls you in a few years😭 i fear for you

[sidayan shabala @sidayanshabala]

at least someone agrees with what i think

[庾晚音 @KdramaHoneyBee]

Yes, you can. Voting is a right, not a requirement. This thinking is regressive guys, let’s find more productive ways of convincing people to vote — giving the merits of casting a ballot, for example…

[sidayan shabala @sidayanshabala]

@heismarshalls I thought you outgrew the chainsmokers bruh😆 boi was i in for a surprise when i saw they were in your Wrapped

[sidayan shabala @sidayanshabala]

💀

[uMashali @heismarshalls]

#SpotifyWrapped back to back number 1 artist taylor swift really healed me in 2023 😹😹😭😭🔥

[sidayan shabala @sidayanshabala]

Okay this is bopping hard AF🔥🔥

[AFRO-RAVE ☥ @ToyaDelazy]

While you here this is my latest song Qoba, the sound is called #Afrorave, bringing the black culture to electronic music!! *£