Малий Ґорґан

Heard that phrase a number of times. Everyone who was every saying it looked so smug that it gave me a toothache. Basking in own self-righteousness and and projecting that implication that "if you _really_ understand this shit, you should be able to explain it well, and if you can't it only speaks ill of you."

No, don't take it wrong. Ever since my graduate advisor's co-author once told me that "every good idea should be explainable by hand-waving", I took this idea to my heart and ran with it. But there is one very big issue with it -- it's too easy to abuse and it's being abused. The problem is obvious: if I explain it to you like you are a five-year old, then you would know if like a five-year old, right? There is a reason that we are trying to teach our kids stuff when they are older than five -- a five-year-old's understanding is primitive. There are only two reasons you would want to know something only at the very basic level: (a) Either you don't really care about it and want to know just enough to pick the right dog reflex about it (whether you should eat it, fuck it, piss on it, or bark at it) (b) Or you want to get a starting point about learning more about that something.

But in a modern corporation a "decision-maker" manager never has enough bandwidth yet cannot ever say "I don't know". So they make others explain to them every little thing that would happen under their amazing management... only explain to them like they are five. And they, based on this five-year-old-like level of understanding they are trying to make actual decisions.
That's why the world is in such as sinkhole and accelerating downward. People who make decisions think that they know their shit, because they think that they know what the world around them is about, but they know it at the level of a five year old. And very few of them have tried to read Janusz Korczak's classic, and even if they did, yet fewer would see the parallels. So, we are all doomed.

In case you wonder what is the right solution? Should each manager know everything they manage to the last detail? Should everyone become a rocket scientist? What is the way? Well, I have some answers for you:
1. I don't think there is a general answer, if it exists, I don't know it.
2. In the very narrow scope of human activities, where I do think my knowledge may be relevant the answer is three-prong (a) know the limits to your knowledge; (b) outside of those limits do not be afraid to admit your ignorance and do not pretend that your ideas matter; (c) if you don't want to admit that you don't know something and can't manage something, delegate

Thank you for your attention :)
I think it's the same in politics, but I can only project as I am not in politics. This is outside of my limits.

The term "analysis-paralysis" is supposed to describe a situaton, when one cannot start working on a project because one keeps trying to predict how that project would go, without generating any knowledge that would change the outcome of the project.
This is bullshit. I mean, this is an imaginary case, in reality, "analysis-paralysis" is something your managers say in the following scenario:
1. They want to "start working" on a project. (Which means, discuss stuff in meetings and report progress up.) But since it is you who would do the work, they want you to tell them now how the project will go. Typically in a format like: "we will do X with N headcount and will achieve Y% increase in metrics A and B".
2. And they don't do uncertainties, only hard numbers. If you don't have a hard number but only likelihoods, and try to explain them, how we cannot be sure now, but how we can become more sure.... they yell "analysis-paralysis! anathema!", force you to stop thinking, give them any estimate, and start the project.
3. The trick is that it is your number, not theirs, so if the project fails, it's your fault. Yet they will never take an uncertain answer either, because it will require them to make the decision and that would be anathema -- the responsibility always has to be someone else's.

Result: if the project does work, everyone is happy, the manager is rewarded for being biased for action and getting things done. If the project fails, no one remembers the uncertainty portion, the engineer or the data scientist is blamed for giving the wrong numbers, while the manager is moving on to start working on the next project.


Comment 1: No one claims that managers just cannot understand you. Sometimes -- yes, but often they are quite intelligent and knowledgeable... but they still don't want to listen. I gave up on trying to tell apart those who are too dumb from those who just don't care, the outcome is exactly the same.

Comment 2: An debate-prone reader can ask "what if your analysis does take too long?" This is an interesting hypothetical scenario. I say, if the managers do think that the analysis takes too long and it's better to move on despite the uncertainty, they can always do that. All it takes is saying: "this may fail, and if it fails, we'll take the responsibility". But, alas, in my experience this scenario has never been anything more than a hypothetical.

Літав по роботі з одного узбережжя на інше, по дорозі вирішив спробувати написати допис для ЛінкдІна на професійні теми. Але видно в душі сидить образа на деяких колишніх колег, бо зловив себе на тому, що захотів написати таке собі уточнення про одну штуку, яка мене особливо дратує у виконанні деяких менеджерів. Питання в аудиторію: якщо в ЛінкдІні написати отаке, воно сильно вдарить по репутації? (Сам допис я ще не дописав... каламбур, однако)

My little commentary-turned-rant: A “mental model” is a modern corporate phenomenon that involves a less technical manager or a business leader and a more technical concept (e.g., an algorithm), let us call it “X”. Said leader claims to have built a mental model of X when they have learned just enough about it to support a conversation. Often, in that conversation they start actively sharing their ideas about how to develop X, how to productize it, and so on, even if their level of competence on the topic is still grossly inadequate. In a perfectly cooperative environment such approach is fine, it may help fix problems and produce innovations. However, in a corporate environment that is competitive and/or hierarchical, such mental model-based inputs tend to get abused, leading to wrong decisions, shifted responsibilities, and covered-up problems. Since perfectly cooperative environments are rare, these mental models do more harm than good. For what it’s worth, my advice to leaders in high-tech business world is this: if you hear anyone say “my mental model of X is ... ”, put very little value into what they say about that X afterwards.