Hacker News Reader

Comment by Bromskloss | Hacker News Reader

That doesn't seem like an equivalent situation. I'm talking about a tragedy-of-the-commons situation, where an individual has a negligible influence over the climate, and so might rationally go ahead and instead spend his effort on something that he will actually get some benefits from.

View on HN

This is also something Yudkowsky has something interesting to say about

https://arbital.com/p/logical_dt/

Summary: LDT is a decision theory that is able to recognise that the decisions of all agents who implement LDT will be correlated. When it makes a decision it knows that it's making the decision for its entire decision-theoretic equivalence class. If it ever decided to knowingly do a thing that is "individually rational" but collectively disastrous, in cases where the personal cost of the disaster outweighs the personal benefits of defecting, then it would be malfunctioning.

I put quotes around "individually rational" because, in light of LDT, CDT is not really rational. It performs worse. It doesn't win.

And, personally, I don't think the old model of decisionmaking was closer to what humans intuitively consider to be sane, effective decisionmaking, I think we had to be taught to be ruthless parasites who can't keep promises, and I hope we can unlearn that now.

Individuals may be acting in their own self-interest, but civilisation is acting irrationally.

One of the roles of government is to impose rules and regulations that steer individual (and corporate) self-interest in directions which are widely agreed to be in the interests of civilisation (like preventing/mitigating climate change) but which would otherwise be ignored or worsened if left to individuals or corporations.

> Individuals may be acting in their own self-interest, but civilisation is acting irrationally.

Yes, individuals acting rationally is just the thing. If we look closely enough at a civilisation or other group of people, it does break apart into individuals. The individuals are the ones who ultimately has the capacity to act, and has goals (a utility function) and knowledge to base their acts upon.

Right, individuals are the point. Local optimization, no matter how well done, may not get you to a global optimum.

Are you sure it is a matter of local versus global optima? Isn't it rather that what you refer to as the global optimum isn't the optimum for the individual actors, and that each individual benefits from deviating from it (at the expense of everyone else)?

This is the tragedy of the commons viewed as an optimization problem. A global optimum provides benefits essentially perpetually, making it more valuable than the local optimums which are arrived at by each actor individually. (It does require extending the time though the actor's descendents.)

Hence the old saw's insistence on "enlightened self interest".

Granted, but it would still seem rational for everyone to prioritize aligning our short- and long-term incentives so that this sort of tragedy doesn't become worse than it has to be.

Reply on news.ycombinator.com