Experative is the memory and judgment that make reliable autonomy possible. It is the operating layer a self-driving laboratory needs before it can run itself, holding what was tried, why, and what it means for the next decision.
Roadmap
Each transition has to be earned. Moving to the next stage requires evidence that decision quality is improving, not merely that more work can be run. At every stage we keep source-linked evidence, decision rationale, outcome interpretation and documented learning.
Capture conditions, provenance and negatives, so the record can be routed later rather than merely stored.
Recommend or refuse, with the supporting evidence in view. A refusal is a correct output, not an error state.
Rank the next experiments and learn from what comes back, so each returned result changes what follows.
Send approved work to lab automation, with the run record attached back to the decision that ordered it.
Propose, run, interpret and document continuously. Every automated experiment stays reconstructable, holding what was known at the decision, why the run was selected, what the outcome means, and how it changes the next action.
