Alan E Dunne

This is a special post for quick takes (aka "shortform"). Only the owner can create top-level comments.

Alan E Dunne1yQuick Take

On catastrophic risk and effective persuasion:

https://us06web.zoom.us/webinar/register/WN_j555baQqRjeWhiefEAQ82Q#/registration

-2

Replying toMy Current Thoughts on the AI Strategic Landscape

My Current Thoughts on the AI Strategic Landscape

So far as the slave carries out immediate work from fear of consequences they are locally aligned with the master's will.

Replying toPublic Opinion on AI Safety: AIMS 2023 and 2021 Summary

Public Opinion on AI Safety: AIMS 2023 and 2021 Summary

How did you get respondents? Why are they "nationally representative"?

Replying toWhat is to be done? (About the profit motive)

What is to be done? (About the profit motive)

1/ evidence for these statements?

2/ in what sense is it profitable to throw away food or maintain empty dwellings that is distinct from "maintaining everyone else's quality of life"?

3/ if the evil is that some people's needs are not valued enough could that not be remedied by giving them money and making it profitable to meet their needs?

Replying toPolarization is Not (Standard) Bayesian

Polarization is Not (Standard) Bayesian

Is martingale different from conservation of expected evidence?

https://www.lesswrong.com/posts/jiBFC7DcCrZjGmZnJ/conservation-of-expected-evidence

Replying toEvaluating GPT-4 Theory of Mind Capabilities

Evaluating GPT-4 Theory of Mind Capabilities

With Respect

Given that in more than a third of the cases where GPT and the answer set disagreed you thought GPT was right and the answer set was wrong, did you check for cases where GPT and the answer set agreed on an answer you thought was wrong?

Yours Sincerely

Replying toReview & rebuttal of "Why machines will never rule the world: artificial intelligence without fear"

Alan E DunneJul 08, 2023

Review & rebuttal of "Why machines will never rule the world: artificial intelligence without fear"

Astral Codex Ten: https://astralcodexten.substack.com/p/your-book-review-why-machines-will

Replying toWhat are the best non-LW places to read on alignment progress?

What are the best non-LW places to read on alignment progress?

This seems to have stopped in July 2022.

Replying to[Linkpost] Introducing Superalignment

[Linkpost] Introducing Superalignment

"Finally, we can test our entire pipeline by deliberately training misaligned models, and confirming that our techniques detect the worst kinds of misalignments (adversarial testing)."

Replying toIdea: medical hypotheses app for mysterious chronic illnesses

Idea: medical hypotheses app for mysterious chronic illnesses

In a "heatplot" or plots cf https://www.elsblog.org/the_empirical_legal_studi/2023/05/heatplots-for-correlation-coefficients-graphs.html

Evaluation Evidence Reconstructions of Mock Crimes Submission 3

Evaluating Evidence Reconstructions of Mock Crimes -Submission 2

Epistemic Status: Exploratory

Introduction

I want to train people to decide the truth about events, through practices reconstructing events about which the truth is independently known

I am asking the help of people who are interested in applying reason to understand how reason and the understanding of it should shape how training for applying reason to solve a class of practical problems should be evaluated

At present:

-no one is specifically trained to bring disputes about the facts of events to a final conclusion

-we have no way to tell how accurate factfinders or factfinding methods are

-dispute resolution is shuffled to people not specifically trained or designated for it

Alternatives considered:

-common sense -anyone can do it

-learning through experience... (read 683 more words →)

-1

Evaluating Evidence Reconstructions of Mock Crimes

Epistemic Status: Exploratory

Introduction

I want to train people to decide the truth about events, through practices reconstructing events about which the truth is independently known

2. At present:

-no one is specifically trained to bring disputes about the facts of events to a final conclusion

-we have no way to tell how accurate factfinders or factfinding methods are

-dispute resolution is shuffled to people not specifically trained... (read 682 more words →)

-1