Log inSign up
user avatar
Covariant
@CovariantAI
Empowering robots to see, think, and act.
Berkeley, CA
covariant.ai
Joined November 2018
160
Following
11.6K
Followers
  • user avatar
    Covariant
    @CovariantAI
    Jan 29, 2020
    After 2+ years in stealth, we’re excited to launch today! covariant.ai Thank you to our team, customers, partners and investors, we couldn’t have done it w/o your support and trust. Exciting milestone, even more exciting journey ahead! medium.com/covariant-ai/b…
  • user avatar
    Covariant
    @CovariantAI
    Mar 11, 2024
    Today, we are introducing RFM-1, our Robotics Foundation Model giving robots human-like reasoning capabilities.
    00:00
    118K
  • user avatar
    Covariant
    @CovariantAI
    Oct 25, 2022
    Thanks to @BillGates for stopping by to experience the future of AI robotic automation! Bill challenged the Covariant Brain, our AI Robotics platform, to pick chaotically arranged items. The Covariant Brain did not disappoint, directing the robots to pick with accuracy and ease!
  • user avatar
    Covariant
    @CovariantAI
    May 6, 2020
    We’ve raised $40m in Series B funding led by @IndexVentures w/ AI-focused @Radicalvcfund + existing investor @AmplifyPartners. Grateful for the support of our investors, customers + partners as we continue to bring AI Robotics to the real world!
    Logistics AI Startup Covariant Reaps $40 Million in Funding Round
    Logistics AI Startup Covariant Reaps $40 Million in Funding Round
    From wsj.com
  • user avatar
    Covariant
    @CovariantAI
    Aug 28, 2024
    Our Robotic Foundation Model, trained on data from real-world warehouse operations and simulated scenarios, enables our robots to identify objects, understand 3D spaces, and predict optimal grasping and placement.
    00:00
    34K
  • user avatar
    Covariant
    @CovariantAI
    Apr 9, 2024
    Large language models are trained to predict the next token in text. For robotics, this means training models on physical interaction datasets to build generalized AI that can simulate the physical world. We've built RFM-1, the first commercial Robotics Foundation Model. RFM-1
    00:00
    45K
  • user avatar
    Covariant
    @CovariantAI
    Feb 25, 2020
    Interested in what ABB's competition looked like? We made a new *uncut* 30+ min video to show an ABB/Covariant robot picking lots of difficult items from the original challenge, from rubber 🦆 to 🍎to💊bottles. (Full video here: bit.ly/2v6lViP)
    00:00
  • user avatar
    Covariant
    @CovariantAI
    May 16, 2024
    Scaling Law in Robotics: We see that as we train RFM-1 on more data, our model's performance improves predictably. For example, our research data indicates that by increasing the size of the dataset on which RFM-1 is trained, pick retries can be reduced by 43%. More importantly,
    52K
  • user avatar
    Covariant
    @CovariantAI
    Mar 27, 2024
    RFM-1's latest scaling update enables robots with in-context learning of grasping improvements. The video shows the self-reflective reasoning capability — after a few tries and failing, the robot has an internal dialogue, hypothesizing that its current gripper is not suited for
    00:00
    25K
  • user avatar
    Covariant
    @CovariantAI
    Jun 20, 2024
    The best performing AI matters. @ABBRobotics invited 20 robotics companies to solve 26 complex piece-picking challenges, half of them kept secret. Covariant was the only one to successfully complete them all. See the full video here: covariant.ai/insights/covar…
    00:00
    8K
  • user avatar
    Covariant
    @CovariantAI
    Apr 3, 2024
    A multimodal any-to-any sequence model, RFM-1 is an 8B parameter transformer trained on text, images, videos, robot actions, numerical and sensor data. RFM-1 tokenizes all modalities into a common space and performs autoregressive next-token prediction with a broad range of
    00:00
    11K
  • user avatar
    Covariant
    @CovariantAI
    Jan 29, 2020
    Curious what a Covariant-powered robot looks like in live production? Here are a few videos from the Obeta warehouse near Berlin!
    00:00
  • user avatar
    Covariant
    @CovariantAI
    Feb 21, 2024
    Our co-founders made the bold move to leave the early team at OpenAI. Why? To create AI for the physical world—productized robots that can reason and adapt to real-world scenarios. An early conviction that is echoed by @ilyasut in a discussion on @dwarkesh_sp's podcast.
    00:00
    14K
  • user avatar
    Covariant
    @CovariantAI
    Dec 15, 2020
    Why do robots struggle w/ seemingly simple tasks? In her @NeurIPSConf talk, Covariant AI research scientist Anusha Nagabandi explains why, and argues that model-based deep RL gives robots the reasoning + adaptation skills needed to make sense of the world: bit.ly/2LAyfQa