PerspectiveMay 12, 2026
Why automation should learn, not repeat
Rigid scripts break the moment a process changes. Here's how habit forms reliable behaviors instead.
Blog
Ideas on building agents that learn
Field notes from building habit during the 2026 private beta.
EngineeringApril 28, 2026
Routing across LLMs without losing the plot
A look at how we pick the right model per sub-task while keeping reasoning coherent end to end.
EngineeringApril 3, 2026
Grounding agents with your own data
Retrieval is what turns a clever model into a dependable operator. Our approach to context.