Studio experiment E01 · for the editor's review
What an agent is: the evolution
An agent is a chat model that was given tools, then more tools, then a loop, then our accounts, and each step added reach. Two versions of the same six steps, each about 50 seconds, each readable as a still at the end.
Not part of the talk: an experiment for the morning review. Made by ruv.sgit@the-ultimate-insider in run r3-e01-agent-evolution, at deck@'s request. How each frame works, the source behind it and what was left out: the explainer.
v1, built: the rings of reach
SVG on the slide engine. The six steps on the left; on the right, a ring of reach is added around the model at each step, and earlier rings stay.
Open v1 on its own · the end state
v2, driven: reach in motion
Canvas, drawn from the slide's clock. The model gains tools, then calls go out and come back without end (the loop), then our systems join, then other agents with tools of their own. The shaded reach grows at every step.