I enjoyed this discussion with Elon Musk at Starbase from Everyday Astronaut - he walked through his 5 step process for efficient manufacturing at scale.
Make requirements less dumb - they are probably dumb, everyone is wrong some of the time. If the requirement was given to you by a “smart” person, even more reason to question it. Each requirement or constraint comes with a name, a person (not a department) who takes responsibility for that requirement.
Try very hard to delete the part or process. If you are not occasionally adding things back in, you are not deleting enough stuff. There is a bias to add stuff “in case you need it” which is a slippery slope.
Simplify or optimize - most common error is to optimize a thing that should not exist… so make sure you try to delete it first. We’ve been trained to answer the question rather than say the question is dumb.
Accelerate cycle time. Go faster. But first optimize the first 3 things.
Final step is automation. Making robots faster or better will not fix your problems if you haven’t optimized the above.
Bonus tip:
Axe the in-process testing. You don’t know what is breaking, so you test a broken process at multiple steps. Then people don’t remove testing as time goes on. If “end of line” testing is fine, then remove the in process testing steps. In process testers will choke the cycle time, may have false positives and slow the whole thing down.
The day of biology as an engineering discipline will dawn with the advent of the synthetic cell. Then we can truly begin to implement some of these strategies of rationale design, streamlining, and automating the engineering of life.
That said, we should be applying these steps in biology related design (bioreactors, I’m looking at you) especially when it comes to feature bloat. We’re still designing and building for maximum cleanliness / regulatory compliance - not every biological product is a drug, and they certainly can’t be priced that way, so we have to make manufacturing cheaper.