Dmytry

On accepting an argument if you have limited computational power.

It would seem rational to accept any argument that is not fallacious; but this leads to consideration of problems such as Pascal's mugging and other exploits. I've had a realization of a subconscious triviality: for me to accept an argument as true, it is not enough that I find no error in it. The argument must also be so structured that I would expect to have found an error if it was invalid (or I myself must make such structured version first). That's how mathematical proofs work - they are so structured that finding an error requires little computational power (only knowledge of rules and reliability); in the extreme case an entirely unintelligent machine can check a proof. In light of this I propose that those who want to make a persuasive argument should try to structure the argument so it'd be easy to find flaws in it. This also goes for the thought experiments and hypothetical situations. Those seem rather often to be constructed with entirely opposite goal in mind - to obstruct the verification process or to try to prevent the reader from trying to find flaws. Something else tangentially related to the arguments. The faulty models are the prime cause of decision errors; yet the faulty models are the staple of thought experiment; nobody raises an eyebrow as all models are ultimately imperfect. However, to accept an argument based on imperfect model one must be capable of correctly propagating the error and estimating the error in the final conclusion, as a faulty model may be so constructed as to itself differ non substantially from the reality but in such a way that the difference diverges massively along the chain of reasoning. My example of this is the Trolley Problems. The faults of original model are nothing out of ordinary; simplified assumptions of the real world, perfect information, etc. Normally you can have those faults in model and still arrive at reasonably close outcome. The end result is throwing of fat people onto tracks, cutting up of tra

32Jan 11, 2012

Conjunction fallacy and probabilistic risk assessment.

26Mar 8, 2012

Rationality of sometimes missing the point of the stated question, and of certain type of defensive reasoning

22Dec 29, 2011

Evolutionary psychology: evolving three eyed monsters

20Mar 16, 2012

Dmytry

Message

817

798

16y

A belief propagation graph

I drew an illustration of belief propagation graph for the AI risk, after realizing that this is difficult to convey in words. Similar graphs are applicable to many other issues. The issue, in brief: Ultra low latency (i.e. low signal delay) propagation from biases to AI risks, slightly longer latency...

May 10, 20124

Complexity based moral values.

Preface: I am just noting that we people seem to be basing our morality on some rather ill defined intuitive notion of complexity. If you think it is not workable for AI, or something like that, such thought clearly does not yet constitute a disagreement with what I am writing...

Apr 6, 2012-12

Pascal's mugging and Bayes

Suppose that your prior probability that giving $1000 to a stranger will save precisely N beings is P(1000$ saves N beings)=f(N) , where f is some sort of probability distribution. When the stranger makes a claim that he will torture N beings unless you give him the $1000 , the...

Mar 30, 20123

Deadlines and AI theory

The AI is a real-time algorithm - it has to respond to situation in the real time. The real-time systems have to trade time for accuracy, and/or face deadlines. The straightforward utility maximization may look viable for multiple choice questions, but for write-in problems, such as technological innovation, the number...

Mar 23, 20120

Open question on the certain 'hot' global issue of importance to FAI

A question: why anything about global warming gets downvoted, even popularly readable explanation of the fairly mainstream scientific consensus? edit: Okay, this is loaded. I should put it more carefully: why is the warming discussion generally considered inappropriate here? That seems to be the case; and there are pretty good...

Mar 23, 2012-9

Better to be testably wrong than to generate nontestable wrongness

Abstract: Test the world-models [at least somewhat] scientifically by giving others and yourself opportunity to generate straightforwardly and immediately testable factual predictions from the world-model. Read up facts to make sure you are not wrong before posting, not only to persuade. I have this theory: there are people with political...

Mar 20, 2012-7

Saturating utilities as a model

Okay, it is a very raw idea, but consider the utility processing that works as following: 1: The utility i'm speaking of is not 'happiness', nor is it 'strength of the compulsion', the utility is only used for the purpose of comparing between futures to pick the one with larger...

Mar 19, 20120

Load More (7/25)

LESSWRONG
LW

LESSWRONG
LW

Dmytry

Dmytry

Dmytry

On accepting an argument if you have limited computational power.

Conjunction fallacy and probabilistic risk assessment.

Rationality of sometimes missing the point of the stated question, and of certain type of defensive reasoning

Evolutionary psychology: evolving three eyed monsters

Dmytry

A belief propagation graph

Complexity based moral values.

Pascal's mugging and Bayes

Deadlines and AI theory

Open question on the certain 'hot' global issue of importance to FAI

Better to be testably wrong than to generate nontestable wrongness

Saturating utilities as a model

On accepting an argument if you have limited computational power.

Conjunction fallacy and probabilistic risk assessment.

Rationality of sometimes missing the point of the stated question, and of certain type of defensive reasoning

Evolutionary psychology: evolving three eyed monsters

A belief propagation graph

Complexity based moral values.

Pascal's mugging and Bayes

Deadlines and AI theory

Open question on the certain 'hot' global issue of importance to FAI

Better to be testably wrong than to generate nontestable wrongness

Saturating utilities as a model