LESSWRONG
LW

All of conitzer's Comments + Replies

Why am I Me?
conitzer10

Not the Doomsday Argument, but self-locating probabilities can certainly be useful in decision making, as Caspar Oesterheld and I argue for example here: http://www.cs.cmu.edu/~conitzer/FOCALAAAI23.pdf and show can be done consistently in various ways here: https://www.andrew.cmu.edu/user/coesterh/DeSeVsExAnte.pdf

Reply
1dadadarren
Let's take the AI driving problem in your paper as an example. The better strategy is regarded as the one that gives the better overall reward from all drivers. Whether the rewards of the two instances of a bad driver should be cumulatively or just count once is what divides halfers and thirders. Once that is determined the optimal decision can be calculated from the relative fractions of good/bad drivers/instances. It doesn't involve taking the AI's perspective in a particular instance and deciding the best decision for that particular instance, which requires self-locating probability. The "right decision" is justified by averaging out all drivers/instances, which does not depend on the particularity of self and now.  Self-locating probability would be useful for decision-making if the decision is evaluated by its effect on the self, not the collective effect on a reference class. But no rational strategy exists for this goal
Why am I Me?
conitzer*10

Just to make sure I understand your argument, it seems that you (dadadarren) actually disagree with the statement "I couldn't be anyone except me" (as stated e.g. by Eccentricity in these comments), in the sense that you consider "I am dadadarren" a further, subjective fact.  Is that right?  (For reference / how I interpret the terms, I've written about such questions e.g. here: https://link.springer.com/article/10.1007/s10670-018-9979-6)

But then I don't understand why you think a birth-rank distribution is inconceivable.  I agree any such d... (read more)

Reply
Hell is Game Theory Folk Theorems
conitzer50

Nice provocative post :-)

It's good to note that Nash equilibrium is only one game-theoretic solution concept.  It's popular in part because under most circumstances at least one is guaranteed to exist, but folk theorems can cause there to be a lot of them.  In contexts with lots of Nash equilibria, game theorists like to study refinements of Nash equilibrium, i.e., concepts that rule out some of the Nash equilibria.  One relevant refinement for this example is that of strong Nash equilibrium, where no subset of players can beneficially devia... (read more)

Reply
In memoryless Cartesian environments, every UDT policy is a CDT+SIA policy
conitzer30

Great discussion!  Just pointing out two additional relevant references:

Aumann, Hart, Perry.  The Absent-Minded Driver.  Games and Economic Behavior, 1997.

Korzukhin. A Dutch Book for CDT Thirders. Synthese, 2020.

The first one also discusses how optimal policies are consistent with CDT+SIA.  Both give examples of other policies being consistent with CDT+SIA too.

Reply