Research & Development Beyond Pre-Trained AI
Trainable, relatable intelligence for every application: a new form of machine intelligence
a python library for AI agents based on a neuro-symbolic alternative to deep learning that is continuous, transparent, and robust
Continous learning is part of human intelligence that is completely absent from AI like GPTs, which remain Pre-trained by design. We're building what's missing from first principles.
Why does this matter? Through our research, we've determined that the hallucination problem, or more generally the inability of AI to understand the meaning behind information, is a result of the dominant pre-training paradigm, where "what to learn" is extrinsically determined by a human, necessiating a gap between training and inference.
By contrast, humans and animals learn with an intrinsic motivation that allows for continous learning and is robust to hallucinations in ways AI still lacks.
Our code fills some of those gaps, and at version 0.1.3, supports a level of associative intelligence comporable to a basic animal like a clam. We're building way outside of the deep learning paradigm, so while we use a form of neural networks, we are building our own core library from scratch and evolving more complicated agent designs that would support higher levels of cognition (after the clam is a worm). We're building AGIs from the bottom-up.
Read the docs, try the API, and say hi on our discord for more access.
Try our new MNIST demo here.
How It Works in 2 Steps
1One config to build custom Agents
Decide on the number of input, output, and control neurons for your application. See examples of Agent Archs here.
2One method to train & query Agents, locally or via our Agents-as-a-Service API
Post inputs to get outputs. If an input includes a control signal (like pleasure or pain), that will trigger the Agent to learn.
FAQs
What base model is running under-the-hood?
How is this different than Reinforcement Learning?
What's meant by AI "trained like animals?"
What's the problem with backpropagation or with AI as it is?
How does this solve AI hallucination or bias?
Why does this matter if "AGI" or superintelligence is imminent?
Is your aim to replace deep learning or Generative AI / LLMs?
How do I contribute?
Is AO Labs a startup or research org?
Have we published any papers?
Are we hiring?
We are thankful to have beta testors from top organizations
Think Differently About Thinking