Tools

Search

Explore

Videos Channels Figures

Atmrix

About

Tools

Search

Explore

Videos Channels Figures

Atmrix

About

JD

Jeff Dean

08/08/17

@ Y Combinator

We created an 'arm farm' where robots learn to grasp objects by trial and error, improving their grasping abilities through pooled sensor data.

Video

YC

Jeff Dean’s Lecture for YC AI

@ Y Combinator

08/08/17

Related Takeaways

DH

Demis Hassabis

05/21/25

@ Google

We're fine-tuning Gemini Robotics to teach robots to perform useful tasks like grasping and following instructions in real-world scenarios.

BM

Bob McGrew

02/01/25

@ Y Combinator

One of the early projects involved teaching a robot hand to solve a Rubik's Cube, aiming to demonstrate that AI could generalize beyond narrow tasks.

JD

Jake Dyson

04/04/23

@ ainslie + ainslie Performance People

We are developing robotic strawberry pickers to automate the labor-intensive process of harvesting, which currently relies on human speed and efficiency.

JD

Jeff Dean

08/08/17

@ Y Combinator

We released a public dataset of 800,000 grasp attempts, which significantly enhances the grasping model's performance compared to smaller datasets.

SP

Sundar Pichai

06/05/25

@ Lex Fridman

AI advancements are crucial for the future of robotics, as they enable us to build models that work safely and effectively in real-world scenarios.

WZ

Wojciech Zaremba

05/18/17

@ Y Combinator

Robots struggle to grasp arbitrary objects, as current solutions are often hand-coded for specific items rather than adaptable to any object.

JF

Jim Fan

05/07/25

@ Sequoia Capital

Data collection for robotics is challenging; we can't scrape the necessary data from the internet, and we rely on expensive teleoperation methods to teach robots.

DH

Demis Hassabis

05/23/25

@ Google for Developers

We are at an exciting moment where advancements in AI algorithms are finally enabling robotics to realize its potential.

PE

Paul Eremenko

05/27/25

@ Sequoia Capital

AI applications that ingest real-world sensor data and robotics are crucial building blocks for achieving physical AI, which is essential for developing engineering AGI.