This paradigm is the basis of eXtreme Programming, Scrum, all agile development as well as the management of DOD 5000.02 ACAT1 (> $5B) programs. It is the basis of all successful project management processing starting with the principles of (1) What does Done look like? (2) How can we get to done? (3) Do we have enough resources to reach done as planned? (4) What impediments will we encounter along the way to done? and (5) How can we measure progress toward done?
The 5th Principle is the mechanism of feedback in all project management methods. It creates a closed loop control system.
Here's a nice simple explanation of closed and open loop control systems. If we assume we are managing or controlling as system without stating a desired outcome based on the inputs we are simply fooling ourselves. The desired state of a software development project is the production of a capability that is valued by the customer for the input of our work efforts, budget, and period of time to develop this value.
So when we hear about developing software without making estimates of cost and schedule, we're operating in an open loop control system. If there are stories placed in a queue in a Kanban style process and we know the capacity for work, we can estimate how long it will take to complete all the work in the queue. If the stories are approximately the same size and the arrival of new stories matches the departure of completed stories, than Little's Law applies.
But this just pushes the needed knowledge of cost and schedule back to those generating the stories. If we want to control the work - and not just be open loop working till we run out of time and money - then we need to know the desired goal. We need a set point for the system. We need to know what done looks like in units of measure meaningful to the decision makers.
The video speaks to thing for machines. Developing software is the same paradigm. We have a goal - a desired capability. A capacity for work - our through put. The output of our work is in response to the goal of the customer to own the capability to do something of value with the software. How long will to take to get that capability and how much will it cost is the desired goal of the software development control system?
Open Loop projects are characterized as discovery processes with no specific outcomes other than to discover what we don't know. I'd love to work on those.