Neuromorphic Agents
Building Safe & Verifiable Neuromorphic Agents
Neuromorphic agents are a new AI substrate with new risks. We think it may also be a route to AI whose behaviour is safe by proof.
Talks, papers, demos, perspectives, and new tools from the Mind Guild Collective.
Milestones, research progress, and announcements, newest first.
Neuromorphic agents are a new AI substrate with new risks. We think it may also be a route to AI whose behaviour is safe by proof.

Katrina Schleisman and Minda Myers presented the five-layer S3Q architecture in Bellevue, WA. Watch the recorded talk and try the simulation in your browser.

Our implementation paper maps each S3Q tenet to published computational primitives, with defined interfaces and falsifiable predictions.
The Implicit Sequence Learning Test now runs on the Mind Guild site, alongside the Memory Palace Trainer and the Basal Ganglia Attention Gate Lab.

Luis El Srouji explores primary and learned value, the PVLV model, and what biologically informed value learning could add to adaptable AI.

A recorded working session on the formal specification of Layer 1: object-centric representations, interfaces, APIs, and data structures.
No updates of this type yet.