The factory wall transfers cleanly to AI rollouts: the question is never how advanced the model is, but which side of the wall a task lands on. Your product-variety axis maps onto judgment variance - the more the correct output depends on context that is not in the spec, the longer the task stays with humans. Tesla's 'humans are underrated' reads less as failed ambition than as a mispricing: the floppy parts were never going to be machine work at any level of trying. The organizations getting AI employment right are the ones that draw the wall before they buy the machines.
Spot on. The "context that is not in the spec" is essentially tacit knowledge. Just like in traditional manufacturing, you can't automate what hasn't been standardized first. Too many companies stumble with AI because they expect the models to magically bridge that tacit knowledge gap. "Drawing the wall before buying the machines" is exactly right.
The coordination point is the strongest part. If hierarchy, handovers and approval chains are solutions to expensive information transfer, then faster machines alone do not change the organization. Lower the cost of coordination and the shape of the factory can change with it.
The factory wall transfers cleanly to AI rollouts: the question is never how advanced the model is, but which side of the wall a task lands on. Your product-variety axis maps onto judgment variance - the more the correct output depends on context that is not in the spec, the longer the task stays with humans. Tesla's 'humans are underrated' reads less as failed ambition than as a mispricing: the floppy parts were never going to be machine work at any level of trying. The organizations getting AI employment right are the ones that draw the wall before they buy the machines.
Spot on. The "context that is not in the spec" is essentially tacit knowledge. Just like in traditional manufacturing, you can't automate what hasn't been standardized first. Too many companies stumble with AI because they expect the models to magically bridge that tacit knowledge gap. "Drawing the wall before buying the machines" is exactly right.
The coordination point is the strongest part. If hierarchy, handovers and approval chains are solutions to expensive information transfer, then faster machines alone do not change the organization. Lower the cost of coordination and the shape of the factory can change with it.