Pannag Sanketi (@pannag_) 's Twitter Profile
Pannag Sanketi

@pannag_

Tech Lead Manager / Researcher @GoogleAI @GoogleDeepMind. Open X-Embodiment Co-Lead. @UCBerkeley @iitmadras alum. Tweet about Robotics/ ML, Tech, Sports, etc.

ID: 71741467

linkhttps://www.linkedin.com/in/pannag-sanketi/ calendar_today05-09-2009 06:12:26

678 Tweet

722 Followers

741 Following

AI KATANA (@ai_katana) 's Twitter Profile Photo

Google DeepMind table tennis robot. It’s the first learned robot agent that reaches amateur human-level performance in competitive table tennis 🏓 Where do I buy one?

Ben (@benshindel) 's Twitter Profile Photo

My mom sent me an article a couple days ago about why AI was overhyped, and part of my response was that it is rapidly revolutionizing robotics and automation. I would not have predicted that we'd be at this level in table tennis for a couple years though!

Peyman Milanfar (@docmilanfar) 's Twitter Profile Photo

Every technical person knows about ordinary least-squares (OLS) but most don’t know *total* least-squares (TLS). These measure fitting error differently: OLS minimizes sum of sq. vertical distances whereas TLS minimizes the sum of orthogonal distances from data to fit line 1/2

Every technical person knows about ordinary least-squares (OLS) but most don’t know *total* least-squares (TLS).

These measure fitting error differently: OLS minimizes sum of sq. vertical distances whereas TLS minimizes the sum of orthogonal distances from data to fit line

1/2
Gautam Salhotra (@gautamsalhotra) 's Twitter Profile Photo

High-frequency reactive control, with RL inference in the loop! This work has been ongoing for a while at Google. so many interesting challenges for such a task! Congrats to Pannag Sanketi , Laura Graesser and the team!

Laura Graesser (@lgraesser3) 's Twitter Profile Photo

Tired on a Friday? Take a break and watch a 25min highlight reel of our robot's table tennis matches. 🍵🏓 💫 Highlights: youtu.be/EqQl-JQxToE 💫 Website & paper: sites.google.com/view/competiti… Want to know one of the best things about this project?

Brett Adcock (@adcock_brett) 's Twitter Profile Photo

A ton of AI and Robotics developments this week. Big announcements from Figure, Google DeepMind, Boston Dynamics, OpenAI, Alibaba, and more. Here's everything that happened and how to make sense out of it:

Aviral Bhatnagar (@aviralbhat) 's Twitter Profile Photo

LKG fees have gone up from 2.3L to 3.7L in Hyderabad, mirroring nationally While we focus on house prices, the real inflation has happened in education Inflation adjusted, school fees are up 9x and college fees are up 20x in the last 30 years Education is no more affordable

Walter Isaacson (@walterisaacson) 's Twitter Profile Photo

The brilliant AI pioneer ⁦Demis Hassabis⁩ explains how machines will have to conquer “planning” in order to move from the current models of artificial intelligence to become true artificial general intelligence. ⁦ youtube.com/watch?v=RDpR9i…

alphaXiv (@askalphaxiv) 's Twitter Profile Photo

Amazing work from Google DeepMind! For the first time, robots are able to play table tennis at a level competitive with humans 🏓 Authors David D'Ambrosio Saminda Abeyruwan Atil Iscen Pannag Sanketi are here to answer your questions!

Pannag Sanketi (@pannag_) 's Twitter Profile Photo

It is heartening that our Competitive Table Tennis paper is featured as the "Paper of the week" on Arxiv Labs! Thank you! It has been great to engage with folks so far. Please keep the questions coming!

alphaXiv (@askalphaxiv) 's Twitter Profile Photo

The top 5 actively discussed papers on alphaXiv this week: 1. LoRA: Low-Rank Adaptation of Large Language Models (w/ Edward Hu) 2. The Unbearable Slowness of Being (w/ Jieyu Zheng, Markus Meister ) 3. The Llama 3 Herd of Models (w/ AI at Meta) 4. The AI Scientist: Towards Fully

The top 5 actively discussed papers on alphaXiv this week:

1. LoRA: Low-Rank Adaptation of Large Language Models (w/ <a href="/edwardjhu/">Edward Hu</a>)

2. The Unbearable Slowness of Being (w/ <a href="/JieyuZheng3/">Jieyu Zheng</a>, <a href="/mameister4/">Markus Meister</a> )

3. The Llama 3 Herd of Models (w/ <a href="/AIatMeta/">AI at Meta</a>)

4. The AI Scientist: Towards Fully
Pannag Sanketi (@pannag_) 's Twitter Profile Photo

This is mighty impressive! Don't be fooled if the tasks look simple... These are very hard for robots. The amount of complexity they must've solved for the robot to be this close to a human is HUGE. Plus the dextrous skills. Amazing work Eric Jang and team! 👏💪