This article aligns extremely well with Robert Pirsig’s Metaphysics of Quality. The only real difference is their substitution of value for quality, which is so

This article aligns extremely well with Robert Pirsig’s Metaphysics of Quality. The only real difference is their substitution of value for quality, which is something my reinforcement learning professor did during lecture yesterday. What it’s missing, is that in Pirsig’s eyes there are 2 essential different kinds of quality, Dynamic quality and Static quality. Dynamic being the change agent, while static clings to what’s good about the past. In his eyes, both are necessary for real progress. Too much dynamic = throwing the baby out with the bath water. Too much static = living in the past. Thinking about it like a machine learning problem, dynamic quality steps out into the unknown, searching for quality paths not yet travelled. Static quality is the machine’s memory of what’s worked well thus far. Occasionally, a step into the unknown is so good, that huge amounts of machine memory seem irrelevant / obsolete. In these moments, humans are prone to wanting to wipe the whole past, without preserving the half or more that’s still an essential piece of their progress thus far. On the other side, humans are prone to tell their peers to stop stepping out into the unknown and argue that we should just go back a step / stay where we were indefinitely. If that’s at all intriguing, feel free to go read my favorite book, Lila. You’ll want to read Zen and the Art of Motorcycle Maintenance (where he recognizes quality) first, if you haven’t already.