AI ALIGNMENT FORUM
AF

CDT=EDT?

Dec 07, 2018 by Abram Demski

Mid-2017, I started to examine my reasons for thinking causality is a necessary ingredient in decision theory. I found grounds to reject all the arguments which had previously seemed to speak in favor of CDT, while at the same time, finding that EDT didn't always do what people claimed it did in problems like Newcomb. I began to develop arguments that CDT and EDT were, in the cases which (I claimed) should be of primary interest, the same decision theory.

8Smoking Lesion Steelman
Abram Demski
8y
10
2Smoking Lesion Steelman II
Abram Demski
8y
1
2Smoking Lesion Steelman III: Revenge of the Tickle Defense
Abram Demski
8y
2
3Comparing LICDT and LIEDT
Abram Demski
8y
2
7 Mixed-Strategy Ratifiability Implies CDT=EDT
Abram Demski
8y
0
3XOR Blackmail & Causality
Abram Demski
8y
4
10A Rationality Condition for CDT Is That It Equal EDT (Part 1)
Abram Demski
7y
3
9A Rationality Condition for CDT Is That It Equal EDT (Part 2)
Abram Demski
7y
0
12Dutch-Booking CDT
Abram Demski
7y
6
21CDT=EDT=UDT
Abram Demski
7y
11
35Troll Bridge
Abram Demski
6y
50
22Dutch-Booking CDT: Revised Argument
Abram Demski
5y
20
28My Current Take on Counterfactuals
Abram Demski
4y
36