- submission id
- —
- input hash
- —
- queue position
- —
- received
- —
01submit
Send something to the cortex.
A submission enters the public queue and waits its turn. H01 decides what happens next — there is no manual selection after this point.
next the submission is encoded into structured features, then into neural stimuli.
Activity propagates through the H01 recurrent graph. The readout resolves into CREATE, WATCH or REJECT.
Only CREATE continues toward an onchain deployment, and only after the policy engine validates it.
Rejected and unstable submissions stay visible in the archive — nothing is deleted because H01 decided against it.
Eight submissions per hour, per person. The cortex runs one job at a time on a single machine.
expect a refusal the readout is at its genesis state — it has never been trained, and it says so in every record it writes.
In practice it rarely separates the eight candidates it is given by enough to commit. The margin it needs before anything may launch is 0.08; the margins it has actually produced so far are all below that, and every one of them is published in the archive next to its submission. So submissions currently resolve to WATCH.
That threshold is the safety mechanism, not a setting. It will not be lowered to make a launch happen. Nothing has been deployed onchain yet, and the archive will show it the moment something is.
Tokens that may eventually be created through this system are experiments. They are not investments and carry no expectation of value.
02follow
Then watch it move.
follow the queue position
watch the neural state evolve
see the decision resolve
verify what reaches the chain