Hacker News Reader

Comment by drewcoo | Hacker News Reader

In the prisoners' dilemma, there is a mutual adversary plotting to turn them against one another. In the prisoners' there is no communication possible. Those are actually the key elements of the puzzle!

Employer/employee relationships one would hope that there is open communication and that there's not some agent trying to pit them against each other.

I think the analysis breaks down right where it starts for work.

I also think the author should seek counseling if they see work that way. That's just a very unhealthy outlook.

View on HN

From what I've seen, the problem is that if an employee communicates their concerns to their boss, then their boss can use that communication against them. E.g. by punishing them for "not being a team player". In practice, a situation where no communication is possible and a situation where it isn't safe to communicate are very similar.

Right, and ironically the bosses/orgs I've been in that talk about "psychological safety" had the most psycho managers.

Why are these people in their positions? Why are better managed companies not replacing them?

Because tech mostly promotes perceived "high performing" ICs to management. Which has no correlation with being a good manager. And then those people move up and promote more people like them.

I've tried it on for size a few times, don't love it, don't think I'm great at it, but fortunately moved into areas that can pay ICs well.

Otherwise most ICs eventually fall into the trap of having to pretend to want to be a manger in order to keep moving up in the pay scale.

Given the choice, I'll stay out of management, but if not doing it meant capping out on pay.. then sure, I'd have to do it again.

I don’t think so. The prisoner’s dilemma is just a payoff matrix. Adding communication doesn’t change it as long as each player can’t see the other’s choice before making their own. And whether the payoff matrix arises because of a mutual adversary or some other way doesn’t seem like it should make a difference to me.

Make it an iterated prisoner’s dilemma, or add some other way for players to punish each other after seeing their choice, and that does change the game.

The relationship between a boss and an employee is so far removed from a classical single-round prisoner's dilemma that I don't think you'll draw reliably accurate conclusions from it. As it sounds like you know, constraints like iteration change the game significantly, as would making it a game with many actors and reputational effects.

All models are wrong, some models are useful, and in this case it seems more wrong than useful.

Edit: thought you were OP, whoops

>Adding communication doesn’t change it as long as each player can’t see the other’s choice before making their own.

Saying that communication doesn't change the prisoner's dilemma so long as each player can't communicate is a contradiction. The only difference is you used the word "see" instead of the word "communicate", but "see"ing is a form of communication.

The prisoners' dilemma is one of a whole class of thought experiment "games" that are defined by the structure of the payoff matrix. These games are generally framed in terms of two or more "players" making simultaneous choices.

In that context, not being able to objectively see the choice made by other players making their own choice is functionally equivalent to making simultaneous choices. You can re-order when the players make choices and communicate at will without changing the payoff matrix and mathematical model.

They're saying there's a distinction of being able to communicate freely (and possibly lie) while still not showing your final action to defect or cooperate. Which yes, leaves the prisoners dilemma much the same as the 0 forms of comms version.

>They're saying there's a distinction

But OP is saying there is no distinction.

Reply on news.ycombinator.com