Sure, after N years, you might be career-tracked from development into management, but you probably also know where all the bodies are buried. If you're not doing code reviews, trying to mentor new developers, or just butting in and saying "We tried that in 2006 and it has huge non-obvious compliance/performance/maintainability costs!", it lets all that knowledge go to waste.
If is supremely annoying when code review is done by superior who does not produce code, but has opinions based on vague recollections.
In contrast, in UX/Design, Steve generally had very astute insights, knew so, and it was much harder to get him to change his mind.
For a publicly documented example of a bad technical decision by young Steve that the team simply worked around for a while until he changed his mind, see: https://www.folklore.org/StoryView.py?project=Macintosh&stor...
It was terrible. CEOs don’t have the time to keep up with technologies and understand the low-level requirements that drive good decision making. A person doing a CEO job can’t possibly keep up with someone doing full-time engineering work in any fast-moving field.
The only way we could make progress was by subtlety seeding our ideas to the CEO and waiting for him to claim them as his own directives after he realized they were good ideas. It was an exhaustive way to get things done.
The ability to say "this is no longer my job" is crucial if a leader is to focus on what is their job.
(Nothing personal colordrops, I've just seen this take a lot lately.)
It's kind of similar to airline pilots or surgeons, where we have to demand "excellence" as a lowest-common-denominator. At that level, they need to be graded on a reverse curve, where unlike regular grades, where a mere 65% score is a passing grade — for those folks, nothing less than a 95% should be a passing grade.
It's easy to make the argument for airline pilots and surgeons, because the human tragedy that results from their potential incompetence is immediate and obvious. I feel like the tragedy that comes out of incompetent CEOs is a lot more insidious, and sneaky. Poorly run companies, and depressing "waste of life" has ruined a lot of people's lives.
In fact, one could dare to surmise that it's actually, statistically speaking, directly killed a lot of people (an easy low-hanging fruit here would be safety stuff that OSHA doesn't cover), but there's a lot of mental health stuff we don't track as well. This whole line of thinking reminds me a lot of Florence Nightengale — her primary contribution to medical statistics was to take something that people brushed off as irrelevant or not-causal, and prove that it was a primary malefactor. She convinced the british top brass that the bulk of their deaths during war had nothing to do with enemy action, or even hospital deaths from injury, but literally was just caused by petty diseases, due to poor barracks sanitation amongst uninjured men.
If you always aim for the exception, what you actually end up doing is selecting for the third competence; someone who can code-switch. And if that person isn't exceptional, then you just have a CEO who is getting involved things outside their depth.
Isn't that exactly what the parent comment said?
So my quibble is simply that "proof leaders do X" doesn't disprove "doing X is bad leadership," despite parent's implication to the contrary.
That 'established' part is key because early stage companies require wearing so many hats. And knowing when and to whom to let something go it's crucial IME.
But the reality today is that most companies simply cannot afford to hire strong engineers across the board for economic supply-and-demand reasons.
You have to work within the bounds of possibility, and yes, that sometimes means that people have to wear multiple hats.
- the leader cannot emotionally bear to give up engineering work and focus on leadership work. This signals a lack of maturity.
- there is some other serious organizational problem preventing delegation from happening (weak engineering IC hires, lack of clear processes and ownership patterns, etc)