• AI That Can Learn After Training

     

    Stateful, compute-efficient, real time models, powered by weightless neural networks.

     

  • EXPLORE OUR TECHNICAL DEMOS POWERED BY AO_CORE WWNS

  • MNIST on WNNs

    A reference design concept on continuous retraining with unparalleled data efficiency using only 788 binary neurons as trainable parameters, 28x28 input to 4 output neurons

  • LLMs + WNNs

    A hybrid recommender system using embeddings and our ao_core framework for context-aware personalized YouTube recommendations

  • ARC-AGI Prize on WNNs

     

    An attempt at the ARC-AGI challenge leaning on the inherent power of WNNs to learn non-linearities

  • WE ARE THANKFUL TO HAVE BETA TESTERS FROM TOP ORGANIZATIONS

     

    broken image
    broken image
    broken image
  • WHO WE ARE

     

    AO Labs horizontal logo

     

    AO Labs is building a more reliable alternative to deep learning and LLMs using continuously trainable, compute-efficient weightless neural networks. We are building AI that can learn after training.