In my experience, the best principal engineers are those who think of themselves as just engineers and simply focus on the best ways to make a difference without much preconceived notion of what that entails. The worst are those who have these strong preconceived notions about what they are supposed to be doing, often focusing exclusively on the more visible, politically expedient tasks in the name of having organizational impact, while eschewing hard technical tasks. Well if that was all that was needed, we don't need the technical ladder - managers that used to be engineers are perfectly capable of handling most of those. It's important to realize why principal engineers are given such autonomy and organizational influence - they are supposed to the best engineers that the organization has to offer and other engineers look up to them because they can do the same things they do but better. What you value in principal engineers should not be something you don't value in other engineers and vice versa. This idea they should be working on different types of tasks altogether and focus less on technical things specifically makes a mockery of the technical ladder.

So the worst part about this article, to me, isn't necessarily this notion that principal engineers shouldn't be all about tech - that I think is relatively uncontroversial. But the reality is that if your organization needs to stress this at that level, this should also be true at every level. If you want to have a great engineering organization, nobody, not even the most junior engineer should merely be converting tickets to code - they should ideally be doing all the things that she thinks principal engineers are doing. And whether someone on the team works more on soft, cross-team stuff or hard technical tasks shouldn't depend on one's level - it's entirely possible that interfacing with other teams or stakeholders is something your mid-level engineer can do well and also solving a very narrow technical problem is something you want your principal engineer should work on because it's extremely difficult. I will go even farther and say that if your engineering organization's challenges are such that the hardest problems that your best engineers should focus on are exclusively non-technical, you should reconsider why you need principal engineers instead of much cheaper project managers.

In one specific instance I observed at this organization, principal engineers were, on the whole, less technical than their managerial counterparts or senior engineers. Because the organization bought into this "principal engineers ought to focus on high-level, organizational, cross-team stuff" - their principal engineers were for the most part too busy trying to direct other teams' and people's work to do anything themselves. They had no deliverables beyond being the multiplier so they were also even more removed from the technical reality than line managers, who are at least forced to deal with what their reports are doing.