The core statement (in my view):

When it comes to strategy, planning and so forth, they’ll go to the suit. Even though the engineer might be much better suited to discuss these topics.

So the problem is the estimation

P(Strategy | Suit) > P(Strategy | Tech)

Humans are pretty good at picking up on regularities, so assuming that this is indeed a consistent pattern, there must be a reason for this judgement.

• Assuming that most strategy people do wear suits, I’d guess that P(Suit | Strategy) is pretty large.
• Assuming that most strategy people are not tech guys, I’d guess that P(Tech | Strategy) is pretty small.

i.e.

P(Suit | Strategy) >>  P(Tech | Strategy)

then via Bayes:

• P(Strategy | Suit) = P(Suit | Strategy) P(Strategy) / P(Suit)
• P(Strategy | Tech) = P(Tech | Strategy) P(Strategy) / P(Tech)

Now, assuming that there’s fewer managers than engineers, i.e.

P(Suit) <= P(Tech)

we immediately get the “problematic” estimation above:

    P(Strategy | Suit) 
=   P(Suit | Strategy) P(Strategy) / P(Suit) 
>   P(Tech | Strategy) P(Strategy) / P(Suit) 
>=  P(Tech | Strategy) P(Strategy) / P(Tech) 
=   P(Strategy | Tech)

In this case, we could say that:

• There’s not enough engineers && suits
• There’s too many suits doing strategy
• There not enough engineers doing strategy
• There should be more suits in tech companies (???)

Modulo sloppy reasoning and faulty assumptions :)

One of my employers really infantilized developers. There were no leadership positions in the developer track. A developer was at the bottom of th org chart up until the last step in the engineering ladder. A team of 8 developers (usually less) had at least 3 manager-types overseeing them. This employer is considered a thought leader in engineering cultuer as well.