Thinking, in public.
Short perspectives from the practice of building useful AI and dependable software.
NOTE 01 / Applied AI
Start with the first useful agent.
A narrow, measurable workflow is a better starting point than an enterprise-wide promise.
Read field noteNOTE 02 / EngineeringOwnership is an architecture decision.
A handover works when independence is built into the engagement from the start.
Read field noteNOTE 03 / DeliveryThe distance from demo to dependable.
A prototype proves a possibility. Production proves that it survives contact with the real world.
Read field noteA useful conversation is a good start.Bring the hard problem.
Start a conversation