A state machine approach (pre-backprop) to general AI that actually works and isn't a black box, at the level of animals, in the form of a python library (soon to be open sourced).
We're currently at v0.1.0, supporting clam-level cognition, and we welcome the curious to our closed-beta.
Self-serve signup is currently paused; please book a meeting.
The code is still small right now (you use 2 major classes written in ~600 sloc to spin up agents made up of single-digit to n-number of neurons; eg. our v0.1.0 basic clam has 9 neurons), so we're developing this project with a group that's all looking at the code together, from the ground up, and we are always actively recruiting collaborators who can understand the enormous potential of our mission.
Come poke at the code, train our generally-intelligent virtual clam, try coding up an agent for your application (we'll help!), or ask us anything.
We'd love to hear from you.
Curious developer? To join our closed beta at this stage, the first step is a quick conversation where we tell you more about this project (a lot of context is involved, and we haven't documented everything yet; its also based on AI research that predates wiki).
Book a demo meeting with us by clicking the button below.
If you're not a developer, of course we're still very keen to meet you; here's our calendly page for easy meetings.
Or sign up below to receive occasional updates.