Matching pennies essays

Rewards vs goals - The wirehead problem and forecasting first;
Command line - The Matching Pennies Command Line Protocol;
Probabilities - Revealing probability estimates;
Stream virtues - The virtues of streaming;
Generality - Maintaining generality while matching pennies;
Forecasting as compression - The link between forecasting and compression;
Inductive inference is like lift - Comparing machine intelligence to powered flight;
The broad scope of inductive inference - The implications and significance of inductive inference;
Manipulative forecasting - Forecasters may try to make the world simple;
Cybernetic diagrams - Diagrams of a simple forecasting agents;
Wirehead analysis - Will forecasting agents wirehead?
Collective intelligence - Strategies can join forces;
Code squeezing - On squeezing the model;
The virtues of generality - Why be a "jack of all trades"?;

Machine intelligence essays

Universal instrumental values - Some instrumental values are widespread;
Expected utility maximisers - Basic optimisation architecture;
Self-improving systems - Systems thet feed on themselves;
Solomonoff induction - On near-optimal inductive inference;
Universal agents - Universal agents, sensation and action;
Oracles, sages, genies - A pantheon of superintelligences;
Utility counterfeiting - The problems of fake utility;
The one true razor - The optimality of Occam's razor;
On AIXI - On AIXI: a model of synthetic intelligent agents;
Reversibility and machine intelligence - Advanced machine intelligence and reversible computation;
Open source - On open source machine intelligence;
Beyond Occam's razor - Occam's razor has been superseded;
Machine faith - Self-fulfilling prophesies;
Cryptocurrencies - Cryptocurrencies and machine intelligence
Turing Test Tournaments - Designing Turing tournaments
Architectural limits to computation - A roadmap for overcoming computational bottlenecks
Deep prediction - Using neural networks for forecasting
Intelligence augmentation - Machine intelligence on a human foundation;
Angelic foundations - Building angels;
Intelligent design vs random mutations - How does intelligent design stack up against GAs;
The second superintelligence - One superintelligence - or many?;
The awakening marketplace - Stockmarket AI;
Superintelligence - the driving problem - The path to superintelligent machines;
The wirehead problem - Superintelligence vs addiction;
On parallelism -  Computer architecture: the path to a parallel world;
Ubiquitous digitalisation - On the progressive digitalisation of the world;
Evaluation under simulation - How evolution performs evaluations under simulation;
Self improving systems are here already - Companies and governments as SISs;
Optimization - A brief introduction;
The number of insights before superintelligence - Number of insights required to produce superintelligence;
Against strong AI - The term "Strong AI" - and why you should probably avoid using it;
How quickly will machine intelligence take off? - How quickly will machine intelligence "take off";
Will machines take our jobs? - Will the advent of intelligent machines make humans redundant;
Handicapped superintelligence - How utility functions of intelligent agents may affect their survival;
Stopping superintelligence - Superintelligent agents that don't object to being switched off;
Ethical Artificial Intelligence - A review of Bill Hibbard's book;
Belief aggregation - A collective intelligence technique;
Bayesianism - Probability theory basics;
Far future humans - Instrumental erasons for keeping humans around;

Tim Tyler | Contact |