A lot of others have talked about the UML aspect of this post, but I want to speak to something that was more in the latent space of my reading of the post and

A lot of others have talked about the UML aspect of this post, but I want to speak to something that was more in the latent space of my reading of the post and its dislike for agile trends and a kind of supposition of something being lost by losing UML. It seems to me that people emphasize the success of engineering disciplines outside of software engineering as a way to show the weakness of software engineering, but I think this way of looking at software engineering artifacts is misleading. Optimization algorithms have some pretty dismal outcomes when you choose to analyze at a particular point in time, especially when you choose a failing artifact, because those failures are absolutely intentional and critical to the proper functioning of many types of optimization processes. As an example generative discriminating processes, like GANs and like Software Creation + Chaos Engineering as popularized by Netflix, thrive on the creation of failure. Yet if you focus on the examples of failure you're missing the forest for the tree whose death is fertilizing the soil of the forest and resulting in a thriving ecosystem. Yes there was failure, but that failure wasn't an indication that the practice itself wasn't holistically valid. The failures were the means by which success was obtained. This falls directly out of the mathematics of learning. The field envy of other engineering disciplines is misplaced. The limit of these optimization processes when expressed in code tends to be machine learning. These optimization processes result in learned creations of superior quality to what is currently produced and it seems highly likely that most fields that don't embrace this will find themselves out-competed by disruptive upstarts who aren't blind to the benefit of new technologies and workflows. Likely, some of these upstarts will be startup founders who frequent this site. When you look at a time slice, failures look terrible. When you look at a longer period of time: failure is smarter than success. Risk aversion is self-destruction; die to live or you'll live to die.