This browser cannot run WebGL, so the live current is not available here. The article has a rendered version of the same idea.
Zero Point Studio Lab
River of Intelligence
A current that is 99.9 percent autonomous, and the rare human pebble heavy enough to bend it.
● ClickTap to drop a pebble. Hold to make it heavier.
Processes executed
0
Autonomous decisions
0
Human interventions
0
Decision budget
0 / 100
0 past decisions are still steering this river
starting
How to read it
Almost all of the work above is autonomous. The small part that is not changes everything downstream of it.
01
The current
Autonomous AI working
Hundreds of thousands of processes moving at once, computed on the GPU every frame from a flow field rather than stepped one by one on the CPU. It runs whether or not anyone is watching it.
02
The pebble
Human input
Tiny in size, enormous in density. It behaves like a black hole: an event horizon, an accretion disk of captured work, and a visible warp in the field for a long way around it.
03
The course change lasts
The direction you set sticks
The deflection does not decay. Once the current has bent around your instruction it holds that new line all the way downstream, which is the whole point of the metaphor.
04
The swarms
AI teams adapting
Agent clusters plan their own routes and fork on their own schedule. Drop a pebble in front of one and it re-plans around it instead of fighting it, then carries the instruction downstream in coral.
The decision budget
A human gets to make roughly 100 real decisions in a day. That is the whole budget, and it is why the counter above tracks every pebble you drop. Make them count: automate more of the work, put yourself inside the flow, and spend your decisions where they actually change the direction.
Build your own
This is a model of how I already work. The agents behind it are real, and they run in teams: a lead agent that makes most of the calls, and specialists under it doing the work.
CRHQ.ai is how you build those teams