This is incredibly difficult and I think more represents the ideal, rather than a consistently achievable reality (at least I have never actually seen this in my own experience). Inevitably your upper leadership becomes infiltrated and then taken over by people who's primary skill is gaming the system of incentives. Once entrenched, they are difficult if not impossible to remove, and jealously gatekeep against anyone more competent that might threaten their position from below. New ideas and reforms are seen as threats, as most of the leadership is only interested in these things to the extent that they might provide an opportunity to increase their own social status, wealth, or power. In a company, this is takes on the form of the depressingly banal, but it is as true at the highest levels of power in every state and such things are behind many of the human-initiated catastrophes we see throughout recorded history.
To your point about the 90/10 ratio, I really don't get the sense that Elon thinks this way at all, however. I also agree that I don't think he is really getting the people he wants. As you said, those that would be most likely to stay under these circumstances are those who can't really go somewhere else (like those with a visa) or who are workaholics. Neither of these neatly intersect with the group of productive and skilled engineers and support staff. Instead of gross competency you are selecting for the desperate and the pathological.
This is why most of us could never be in the kind of position like Elon Musk, Bill Gates and Steve Jobs. They see things for what they are and make the cold calculations. They still make mistakes but are directionally correct.
Power laws like the Pareto principle assert themselves unless you actively push against them. Twitter was not. Everyone knew they were overstaffed for years. The 50% cut when Elon took over had little effect. They were not a lean machine where the 90% were pulling the weight.
I also said if the goal is to rebuild twitters culture in a different mold then this is a way to do it. But it is not about identifying the 10%, it’s about identifying the people who map to the intended culture. That’s fine too. But they are driving away the force multipliers by doing this, and creating a reputation that will drive them away. That’s because force multipliers necessarily love the people individually they work with. They aren’t making the cold calculations you’re talking about, and they are generally repulsed by them. And you can’t have an excellent engineering culture based on mercenaries.
You don’t have to believe me. This is just my observations after decades of being involved in these things directly and being the one making the cold calculations.
What she also ended up doing is to smooth over tense situations between very competent people who are working in a high-stake environment where the urge to lash out/to fight for your vision above others' is getting increasingly strong as fatigue sets in.
That value isn't written down in her performance review. Yet, she's diffused a number of conflicts, allowing easier/quicker resolution of issues in a business where a mistake can (will) cost thousands of dollars. Because she's more approachable than her cross-shift, her subordinates are also more susceptible to ask her if they don't know something, instead of doing a potentially costly mistake or letting suboptimal situations carry on.
Personal skill-wise, she's not a magical 10% -- she's decently skilled, still learning her trade. However, she does help multiply everyone's performance.
As parent said, a gear needs many parts to turn.
This idea of 10% and 90% is fundamentally wrong. It’s like saying the only meaningful part of a gear system is the meshing teeth. That’s after all where all the magic happens. But the interior of the gear is importing to position the teeth correctly. Where the gear attaches is more crucial - you can function fine with a few broken teeth but if the attachment is too tight or too lose it won’t work. More importantly the empty space in the system is crucial to the function.
I think the ratio is generally reversed - 90% of the people are making the magic happen, and 10% are redundant. It’s true there’s always a small percentage of folks that demonstrate remarkable effectiveness, but they can’t do that without the rest of the team essentially carrying out the work they are too valuable to be distracted from. They provide the cover, the headroom, and while the exceptional performers are force multipliers, there must be a force to multiply. They can’t multiply themselves alone. A good manager knows how to integrate those people into the rest of the gear system such that their force multiplication is magnified to its maximum. But the force itself is not redundant.
Due to their rarity the force multiplying engineers make more money and receive outsized recognition. The same is true for say a star quarterback (I know nothing about sportsball fwiw). The quarterback is famous and compensated better than the rest of the team if they’re amazing. But they can’t win the game alone.
The other thing to note about exceptional engineers, which I modestly count myself as one after many years of impostor syndrome, they can’t stand bullshit and absurd decision making. They also respect and even love their coworkers and team and will go to no end to see them individually successful. They are generally not mercenaries nor are they obsessively driven by a vision at all cost. There are very competent engineers who are like that. But they tend to be poisonous to the team and largely ineffectual because they’re not force multipliers. More than anything, no one trusts them - not even other engineers like them. They usually get stuck midcareer and either become bitter and withdrawn or end up as the star at some small shop that couldn’t afford a truly amazing engineer.
Now, I don’t work at twitter and never have. They might not be a serious engineering shop and his strategy is to dynamite the existing culture and rebuild it as a more serious engineering shop. But that won’t be successful if he can’t recognize the value of the 90%. And he shouldn’t fool himself that his force multipliers clicked yes. He ended up with people in precarious personal situations that can’t afford to lose their job and mercenaries that were bitter how they were marginalized while the force multipliers were magnified.
He will have a lot of rebuilding, and building of trust once lost is the hardest to rebuild of all. These events can haunt recruiting efforts for decades.