Dev Blog/Our Approach
15 posts
Reducing the Data Gap with Synthetic Data
What published results show about training and evaluating models on generated data, and what they suggest for process histories
SDCPNs: A Common Language for Complex Systems
From plain Petri Nets to a unified representation of continuous and discrete behaviour, structured state, concurrency, and randomness
Inverse problems in Agent-Based Models
Technical challenges and relevance to a Process Foundation Model
How not to build a Process Foundation Model
Why classic AI model architectures fail to handle the unique properties of process graphs
Deriving slugs from embeddings: vec2slug
Building a tiny transformer decoder to extract URL slugs from vector embeddings
Embeddings as Encodings
Correctly conceptualizing and handling vectorization in knowledge graphs
Design-Led AI
UX rules for building AI-powered professional tools
Collaborative data modeling
Allowing users to build a shared, structured understanding of the world
File structuring using fractal trees
Scalable code organizing for large projects and monorepos
A real-life tech tree
Laying out our open-source development roadmap for HASH, in public
Real-time collaboration across arbitrary blocks
Facilitating multiplayer interactions within novel, unrecognized UI components
The non-negotiable principle
The utility of hard constraints in product design
Open source at HASH
How our approach to open source informs what we do
