This is an awesome article. Related phenomenon: platformization of a solution before even the first instance of the problem is properly solved. There are a lot of incentives for this in large tech companies for both engineers and managers to show broader impact, but in practice it often results in ill-fitting solutions that impose a tax on everyone else. A good rule of thumb I'd propose would be to require solving a problem three ways before anyone can propose a platform, then extract the best of each approach to build your platform out of and have the three owning teams all sign off that it is the right solution for their specific problem. Of course it's easy to argue against this in the name of redundant effort, but I think it's a small price to pay to avoid over-ambitious but under-qualified architecture astronauts imposing half-baked mandatory indirection whenever they can corner a gullible director.