What if rigor was easy to scale?
This has everything to do with AI. If you can do it once, you can do it as many times as you want.
The hard part is doing it once.
Building the pipeline is real engineering. Data Mint has solved most of it, and other well-made systems will too. What remains is the codebook. That's the hard part.
A codebook has to carry your full intent: every question, every definition, every edge case, every decision about what the document didn't say. No oral tradition. No training sessions. No shared habits that never got written down. Nothing hides. This is the craft.
Which leaves two questions. (1) Why don't people trust it yet? (2) How do we know when to trust it?
@karlrohe