LESSWRONG
LW

AI Alignment Writing Day 2018

Aug 13, 2019 by Ben Pace

On 10th July 2018, all attendees of the MIRI Summer Fellows Program were given an entire day to write blogposts to the AI Alignment Forum with ideas they'd been thinking about. These are the 28 posts that resulted, in chronological order.

10Choosing to Choose?
Ω
Daniel Herrmann
7y
Ω
7
13The Intentional Agency Experiment
Ω
Alexander Gietelink Oldenziel
7y
Ω
5
11Two agents can have the same source code and optimise different utility functions
Ω
Joar Skalse
7y
Ω
11
28Conditioning, Counterfactuals, Exploration, and Gears
Ω
Diffractor
7y
Ω
1
12Probability is fake, frequency is real
Ω
Linda Linsefors
7y
Ω
7
12Repeated (and improved) Sleeping Beauty problem
Ω
Linda Linsefors
7y
Ω
5
15Logical Uncertainty and Functional Decision Theory
Ω
swordsintoploughshares
7y
Ω
4
15A framework for thinking about wireheading
Ω
theotherotheralex
7y
Ω
4
87Bayesian Probability is for things that are Space-like Separated from You
Ω
Scott Garrabrant
7y
Ω
22
15A universal score for optimizers
Ω
levin
7y
Ω
8
15An environment for studying counterfactuals
Ω
Nisan
7y
Ω
6
55Mechanistic Transparency for Machine Learning
Ω
DanielFilan
7y
Ω
9
43Bounding Goodhart's Law
Ω
eric_langlois
7y
Ω
2
40A comment on the IDA-AlphaGoZero metaphor; capabilities versus alignment
Ω
AlexMennen
7y
Ω
1
27Dependent Type Theory and Zero-Shot Reasoning
Ω
evhub
7y
Ω
3
22Conceptual problems with utility functions
Ω
Dacyn
7y
Ω
12
9No, I won't go there, it feels like you're trying to Pascal-mug me
Ω
Rupert
7y
Ω
0
12Conditions under which misaligned subagents can (not) arise in classifiers
Ω
anon1
7y
Ω
2
58Complete Class: Consequentialist Foundations
Ω
abramdemski
7y
Ω
37
20Clarifying Consequentialists in the Solomonoff Prior
Ω
Vlad Mikulik
7y
Ω
16
11On the Role of Counterfactuals in Learning
Ω
Max Kanwal
7y
Ω
2
28Agents That Learn From Human Behavior Can't Learn Human Values That Humans Haven't Learned Yet
Ω
steven0461
7y
Ω
11
36Decision-theoretic problems and Theories; An (Incomplete) comparative list
Ω
somervta
7y
Ω
0
54 Mathematical Mindset
Ω
komponisto
7y
Ω
5
6Monk Treehouse: some problems defining simulation
Ω
dranorter
7y
Ω
1
24An Agent is a Worldline in Tegmark V
Ω
komponisto
7y
Ω
12
15Generalized Kelly betting
Ω
Linda Linsefors
7y
Ω
5
16Conceptual problems with utility functions, second attempt at explaining
Ω
Dacyn
7y
Ω
5