Newcomb-like problem—specifically a case of Parfit’s Hitchhiker
Parfit’s Hitchhiker is not a “Newcomb-like problem”. In fact, it’s not even obvious that it is actually a proper decision problem: the only decision maker is the driver and you can’t control their decision. You only get to decide if you can precommit.
Anyway, Google doesn’t turn any reference to Parfit’s Hitchhiker independent of LessWrong. Has this dilemma really originated from Parfit or did EY make it up?
Anyway, Google doesn’t turn any reference to Parfit’s Hitchhiker independent of LessWrong. Has this dilemma really originated from Parfit or did EY make it up?
Google points to this as the original reference. It’s paywalled, so I cannot check.
The paper also apparently mentions Kavka’s toxin puzzle, an interesting exercise in rationality and precommitment occasionally discussed on LW.
1) Individuals who make certain decisions seem to win at higher rates than individuals who do not. 2) As far as you know, the act of decision doesn’t causally effect the likelihood of a win.
These two requirements seem inconsistent.
I’d define a decision problem to be Newcomb-like if the payoff and the agent mental state (preferences, beliefs, decision procedures) are not independend conditional on the agent’s decision.
Some of the problem on the list you linked are Newcomb-like, other are committment problems, other aren’t even decision problems.
My def. isn’t inconsistent. Those who buy computers are less likely to die of malaria,
^that’s an instance of Solomon’s problem, which is considered a newcomblike problem.
It’s the same as the fact that those who one-box are more likely to get more money in Newcomb’s. A third factor (socioeconomic status, CTGA allele, agent’s mental state prior to decision) accounts for the variance.
My def. isn’t inconsistent. Those who buy computers are less likely to die of malaria,
Once you condition on all the available evidence, such as the socioeconomic status, these two events become independent. Likewise, in the Solomon’s problem, a genetic test that detects the lesion would destroy the Newcomb-like structure of the dilemma.
Anyway, Google doesn’t turn any reference to Parfit’s Hitchhiker independent of LessWrong. Has this dilemma really originated from Parfit or did EY make it up?
AFAICT Kavka’s toxin puzzle is isomorphic to it (except that in this case the billionaire’s motives are alien).
Parfit’s Hitchhiker is not a “Newcomb-like problem”. In fact, it’s not even obvious that it is actually a proper decision problem: the only decision maker is the driver and you can’t control their decision. You only get to decide if you can precommit.
Anyway, Google doesn’t turn any reference to Parfit’s Hitchhiker independent of LessWrong. Has this dilemma really originated from Parfit or did EY make it up?
Yes, it does. Try using syntax similar to this:
Thanks!
Google points to this as the original reference. It’s paywalled, so I cannot check.
The paper also apparently mentions Kavka’s toxin puzzle, an interesting exercise in rationality and precommitment occasionally discussed on LW.
Thanks. I’ve also found this one, which is not paywalled.
It’s in Reasons and Persons
Thanks!
Then you’re defining it differently from the way I, and others, are.
My req’s for a Newcomblike problem:
1) Individuals who make certain decisions seem to win at higher rates than individuals who do not.
2) As far as you know, the act of decision doesn’t causally effect the likelihood of a win.
what where your reqs?
These two requirements seem inconsistent.
I’d define a decision problem to be Newcomb-like if the payoff and the agent mental state (preferences, beliefs, decision procedures) are not independend conditional on the agent’s decision.
Some of the problem on the list you linked are Newcomb-like, other are committment problems, other aren’t even decision problems.
My def. isn’t inconsistent. Those who buy computers are less likely to die of malaria,
^that’s an instance of Solomon’s problem, which is considered a newcomblike problem.
It’s the same as the fact that those who one-box are more likely to get more money in Newcomb’s. A third factor (socioeconomic status, CTGA allele, agent’s mental state prior to decision) accounts for the variance.
Once you condition on all the available evidence, such as the socioeconomic status, these two events become independent.
Likewise, in the Solomon’s problem, a genetic test that detects the lesion would destroy the Newcomb-like structure of the dilemma.
AFAICT Kavka’s toxin puzzle is isomorphic to it (except that in this case the billionaire’s motives are alien).