id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
4d5d36c6-789d-44ab-8f9b-4ca7da595714 | trentmkelly/LessWrong-43k | LessWrong | If you choose not to decide, you still have made a choice.
Pick one. There will be a quiz.
ADT – Acausal Decision Theory. Figure out what other agents want and do it, even if you can’t talk to them and they might not exist, if and only if they would do the same for you even if you don’t exist. Which hopefully you do.... |
25c2f0c5-bcac-4af4-bec5-53698eaf6b22 | trentmkelly/LessWrong-43k | LessWrong | Idea: Open Access AI Safety Journal
This is short because I'm mainly trying to gather feedback on how much the idea might be worth pursuing, but how much value does it seem the existence of an open access AI safety journal would provide? Some reasons I can think of in favor:
* Have a journal focused on AI safety, gi... |
89f510c1-96da-4ea2-a228-2f83f3c6f56b | trentmkelly/LessWrong-43k | LessWrong | Should you go with your best guess?: Against precise Bayesianism and related views
(This post draws heavily on earlier writing co-authored with Jesse Clifton, but he’s not listed as an author since he hasn’t reviewed this version in detail.)
Should we always be able to say whether one outcome is more likely, less lik... |
5b7b28e7-8d32-465e-95ed-bacbe889c27e | trentmkelly/LessWrong-43k | LessWrong | AI and the paperclip problem (or: Economist solves control problem with one weird trick!)
None |
8b467afa-817b-4249-a17c-19c7ec5d23b9 | trentmkelly/LessWrong-43k | LessWrong | building character isn't about willpower or sacrifice
I always used to think that character was hard to build; that it required Herculean feats of self-restraint, sacrifice, and discipline to maintain and uphold. Others always seemed to be able to do the right thing far more consistently than I did, and it felt like i... |
42f46e04-33a6-40e0-adf2-3c75507d2f2c | trentmkelly/LessWrong-43k | LessWrong | Distinct Configurations
,,,,,,,,
The experiment in the previous essay carried two key lessons:
First, we saw that because amplitude flows can cancel out, and because our magic measure of squared modulus is not linear, the identity of configurations is nailed down—you can’t reorganize configurations the way you can re... |
0dd5312e-2cc7-47c4-8d09-3ffb88476e46 | trentmkelly/LessWrong-43k | LessWrong | [Classifieds] What are you doing to make the world a better place and how can we help?
None |
a384a754-4df6-492c-a4df-830834fe1b5f | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | A Mechanistic Interpretability Analysis of Grokking
**A significantly updated version of this work is** [**now on Arxiv**](https://arxiv.org/abs/2301.05217)
*aka, how the best way to do modular addition is with Discrete Fourier Transforms and trig identities*
*If you don't want to commit to a long post, check out th... |
212b8db2-5d7a-4389-8493-1344d15abd86 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Contra "Strong Coherence"
[Polished from my shortform](https://www.lesswrong.com/posts/zhhYwM7gk8LZsDzxj/dragongod-s-shortform?commentId=ikCNN8625qmfTCJye)
See also: [Is "Strong Coherence" Anti-Natural?](https://www.lesswrong.com/posts/c4GrmECzui2zT4fMq/is-strong-coherence-anti-natural)
---
Introduction
======... |
83637ed0-f2a4-433b-af6a-aaf4f4a075a9 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Conjunction Fallacy
Today's post, Conjunction Fallacy was originally published on 19 September 2007. A summary (taken from the LW wiki):
> Elementary probability theory tells us that the probability of one thing (we write P(A)) is necessarily greater than or equal to the conjunction of that thing and a... |
1d743987-062b-42ed-87ff-9a1b27b9bde0 | trentmkelly/LessWrong-43k | LessWrong | Questions on the human path and transhumanism.
I had a waking nightmare: I know some of you reading this just went "Oh great, here we go..." but bear with me. I am a man who loves to create and build, it is what I have dedicated my life to. One day because of the Less Wrong community I was prompted to ask "What if the... |
125f21e6-57ee-42ff-857f-0357a515a257 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Subjective Knowledge and Reasoning about Agents in Multi-Agent Systems
1 Introduction
---------------
The research community has successfully used Kripke models to model epistemic situations in multiagent systems. But it seems
necessary to have a broader notion of epistemic models that may not always be defined wit... |
f41a09b4-9151-4626-9c1c-593458cde035 | trentmkelly/LessWrong-43k | LessWrong | Predictable Exploration
An idea I've been kicking around for a while without getting very far:
The big problem with epsilon-exploration as a way of getting reasonable counterfactuals is that it only tells the agent what would happen if it took an action unpredictably, rather than predictably; as a result, it doesn't ... |
58dcfdaa-8f34-4a07-98bc-2139b09b218e | trentmkelly/LessWrong-43k | LessWrong | Aligned Objectives Prize Competition
The goal of this prize is to generate ideas and plans for a long-term, positive future. The starting prize pool is $500 USD, though others are welcome to add to it.
We are going to pretend for this exercise that the technical part of AI Alignment is already solved. Your mission, s... |
1f89ad31-1df5-4fb7-96ec-8e37ec979315 | trentmkelly/LessWrong-43k | LessWrong | Class consciousness for those against the class system
Each person is special. Amanda is in a class of her own, Fred is in a class of his own. Amanda has many different properties: where she came from, what she looks like, what behaviors she habitually does, what she knows and doesn't know, what her plans are, how she... |
bc72477c-c3f3-495d-b1f6-43f5193c516d | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Is the argument that AI is an xrisk valid?
Hi folks,
My supervisor and I co-authored a philosophy paper on the argument that AI represents an existential risk. That paper has just been published in Ratio. We figured LessWrong would be able to catch things in it which we might have missed and, either way, hope it migh... |
f8fa3f7a-7583-44d7-950e-5c79b4c71b66 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | A mesa-optimization perspective on AI valence and moral patienthood
Edit: I no longer endorse this view as written.
-----------------------------------------------
### I now see this view as a necessary (but not sufficient) condition for valence.
1. Introduction
---------------
*Summary:* I argue that mesa-optim... |
da6523e8-a0c0-4370-8511-56b636afb618 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Vancouver Bias and Such
Discussion article for the meetup : Vancouver Bias and Such
WHEN: 19 January 2013 01:00:00PM (-0800)
WHERE: 2505 West broadway vancouver
Hello everybody.
We had a blast when jsalvatier came to visit last weekend.
This weekend, I think we should talk about the biases. So come out a... |
ebdb1325-9c11-4c90-8c32-39829a2eeb58 | trentmkelly/LessWrong-43k | LessWrong | Factored Cognition
Note: This post (originally published here) is the transcript of a presentation about a project worked on at the non-profit Ought. It is included in the sequence because it contains a very clear explanation of some of the key ideas behind iterated amplification.
------------------------------------... |
e0ba2ddb-3994-43f1-8677-735a3574917f | StampyAI/alignment-research-dataset/arxiv | Arxiv | Toward Building Science Discovery Machines
1 Toward Building Science Discovery Machines
Abdullah Khalili1* and Abdelhamid Bouchachia1
1 Department of Computing and Informatics, Bournemouth University, United Kingdom.
*Correspondence: a.m.khalili@outlook.com
Abstract The dream of building machines that ca... |
ab38cb2d-ecc9-4dd2-b775-ca6d89ed765f | trentmkelly/LessWrong-43k | LessWrong | The Best Educational Institution in the World
Crossposted on The Memeticist and the EA Forum.
Note: I originally intended to publish this at the end of February. I decided to hold off until now because of COVID-19.
What would happen to college dorms if all colleges disappeared? Imagine some crazy parallel Universe i... |
f30c180d-3c2c-4265-a8b4-dd89065760bb | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Epilogue: Atonement (8/8)
Today's post, Epilogue: Atonement (8/8) was originally published on 06 February 2009. A summary (taken from the LW wiki):
> The last moments aboard the Impossible Possible World.
Discuss the post here (rather than in the comments to the original post).
This post is part of ... |
b3c426c0-4d3e-4d80-bc21-59c5c49fadb8 | trentmkelly/LessWrong-43k | LessWrong | Does AI governance needs a "Federalist papers" debate?
During the American Revolution, a federal army and government was needed to fight against the British. Many people were afraid that the powers granted to the government for that purpose would allow it to become tyrannical in the future.
If the founding fathers ha... |
2e97d44f-b97c-48a7-b2a5-86655444fc61 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Anytime Decision Making with Imprecise Probabilities
470
Anytime Decision Making with Imprecise Probabilities
Michael Pittarelli
Computer Science Department
SUNY Institute of Technology
Utica, NY 13504-3050
mike@sunyit.edu
Abstract
This paper examines methods of decision making
that are able to accommodate li... |
5df3d8d6-429f-4bae-9cf2-513051c69b3b | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Faster Than Science
Today's post, Faster Than Science was originally published on 20 May 2008. A summary (taken from the LW wiki):
> Is it really possible to arrive at the truth faster than Science does? Not only is it possible, but the social process of science relies on scientists doing so - when the... |
36964824-1386-49e4-94cc-b31588f84029 | trentmkelly/LessWrong-43k | LessWrong | Natural wireheadings: formal request.
This post is a formal request for everybody in this forum, who are the most likely humans to produce an FAI in the future, that in the possible resulting utopia some activity will still be available, even if they are not based on challenges of increasing complexity. I call these a... |
d441dbae-5236-406f-b1f0-726718073ab9 | trentmkelly/LessWrong-43k | LessWrong | [LINK] Reverse priming effect from awareness of persuasion attempt?
Recently came across this blog post on Language Log summarizing this recent paper by Laran et al. Super-short version: When people are aware that a slogan is trying to persuade them, reverse-priming effects in which they avoid doing as it suggests can... |
cdd45a5f-5734-4ae3-901d-7b4354883d57 | trentmkelly/LessWrong-43k | LessWrong | Why The Insects Scream
Crossposted on my blog.
1 Beneath the giants’ heels
The giants ruled the world and
They reshaped it a lot
Nature warped beneath their hands
Vast palaces were built, but not
Much heed was paid to those below
Those trampled by their feet
They didn’t really care to know
What those tiny b... |
2620573f-e14e-48ab-a19c-e28c0dddb310 | trentmkelly/LessWrong-43k | LessWrong | Large introductory science classes
The undergraduate course experience at a number of colleges and universities across the world includes introductory science classes. Some colleges require only science-based majors (including sciences and engineering) to take the introductory science classes, whereas others require s... |
f749da46-0c77-48ee-be97-43eab7bc02db | StampyAI/alignment-research-dataset/blogs | Blogs | New paper: “Proof-producing reflection for HOL”
[](https://intelligence.org/files/ProofProducingReflection.pdf)MIRI Research Fellow Benya Fallenstein and Research Associate Ramana Kumar have co-authored a new paper on machine reflection, “[**Proo... |
a8e6f221-fd4c-4b04-8ac0-a70ce9094f93 | trentmkelly/LessWrong-43k | LessWrong | Terrorist leaders are not about Terror
From "Academics Doubt Impact of Osama bin Laden’s Death":
> "...Fifty-three percent of the terrorist organizations that suffered such a violent leadership loss fell apart — which sounds impressive until you discover that 70 percent of groups who did not deal with an assassinatio... |
3d59bdf2-f203-4e32-ab7d-1335338fc525 | trentmkelly/LessWrong-43k | LessWrong | Does there exist a detailed Bayesian COVID tracker?
[bounty: $100 for recommending a tool that I use for more than two weeks]
I would love (and happily pay for) a piece of software that I could tell (a) a bunch of recent interactions between people ("Alice, masked, spent 3 hours, indoors, distanced, with Bob, unmaske... |
2454f90e-4d10-44df-9bf3-ef3f1e085f0c | StampyAI/alignment-research-dataset/special_docs | Other | Safe Pareto Improvements for Delegated Game Playing
Safe Pareto Improvements for Delegated Game Playing
Caspar Oesterheld
Department of Computer Science, Duke University
caspar.oesterheld@duke.eduVincent Conitzer
Department of Computer Science, Duke University
conitzer@cs.duke.edu
ABSTRACT
A set of players delegate pl... |
92604a66-d7b1-4402-82bc-44638dbd1b1c | trentmkelly/LessWrong-43k | LessWrong | The Economics of p(doom)
Transformative AI poses an imminent existential risk to humanity, and yet we're spending close to zero dollars on mitigating this risk. This feels wrong, but by how much?
In a recent paper "The Economics of p(doom): Scenarios of Existential Risk and Economic Growth in the Age of Transformativ... |
530c7b62-0ff5-419a-971d-e4149f45327c | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Prospects for AI safety agreements between countries
This post summarizes a project that I recently completed about international agreements to coordinate on safe AI development. I focused particularly on an agreement that I call “Collaborative Handling of Artificial intelligence Risks with Training Standards” (“CHART... |
aaedc197-51a2-4a3c-b38f-7778eb093453 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Mathematical Measures of Optimization Power
In explorations of AI risk, it is helpful to formalize concepts. One particularly important concept is intelligence. How can we formalize it, or better yet, measure it? “Intelligence” is often considered mysterious or is anthropomorphized. One way to [taboo](/lw/nu/taboo_you... |
a2d696b3-ab98-46e2-bdc1-27ca0bfb9f5b | trentmkelly/LessWrong-43k | LessWrong | The Role of Groups in the Progression of Human Understanding
This post isn't intended to be anything stunningly original, but simply to collect a few basic, but under-utilized observations together in one place.
The question of how we can progress human knowledge and understanding is one that is of particular interes... |
d8219c7b-c973-42f8-b2b5-06d6ee8f92ae | trentmkelly/LessWrong-43k | LessWrong | Creating a Text Shorthand for Uncertainty
Most things I find I discuss are highly uncertain, but it can be really confusing and wordy to state that uncertainty in writing. In this last sentence for example I felt the need to write “I find” to point out uncertainty, for example.
First, people are really bad at agreein... |
40dff9a2-7b05-4de2-8911-ab2f9d0e8e41 | trentmkelly/LessWrong-43k | LessWrong | On Not Pulling The Ladder Up Behind You
Epistemic Status: Musing and speculation, but I think there's a real thing here.
I.
When I was a kid, a friend of mine had a tree fort. If you've never seen such a fort, imagine a series of wooden boards secured to a tree, creating a platform about fifteen feet off the ground ... |
f118eb07-2e69-434f-9010-2ea3bf2d7e1b | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Intermittent Distillations #1
This is my low-budget version of Rohin's [Alignment Newsletter](https://rohinshah.com/alignment-newsletter/).
A critique of pure learning and what artificial neural networks can learn from animal brains (Anthony M. Zador)
=================================================================... |
a02eb119-c01e-470f-8a53-0f34e4e6dd85 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Poll: Which variables are most strategically relevant?
Which variables are most important for predicting and influencing how AI goes?
Here are some examples:
* Timelines: “When will crazy AI stuff start to happen?”
* [Alignment tax](https://forum.effectivealtruism.org/posts/63stBTw3WAW6k45dY/paul-christiano-current... |
0b26d534-39a1-4921-b3e2-3142580db653 | trentmkelly/LessWrong-43k | LessWrong | Building AGI Using Language Models
> Despite the buzz around GPT-3, it is, in and of itself, not AGI. In many ways, this makes it similar to AlphaGo or Deep Blue; while approaching human abiliy in one domain (playing Chess/Go, or writing really impressively), it doesn’t really seem like it will do Scary AGI Things™ an... |
dda0a345-db5e-4bbe-8c0c-bdefd3d5452a | trentmkelly/LessWrong-43k | LessWrong | Life's Story Continues
Followup to: The First World Takeover
As last we looked at the planet, Life's long search in organism-space had only just gotten started.
When I try to structure my understanding of the unfolding process of Life, it seems to me that, to understand the optimization velocity at any given point,... |
8a498287-ad59-46b2-8e38-428c16521622 | trentmkelly/LessWrong-43k | LessWrong | Investigating the learning coefficient of modular addition: hackathon project
As our project at the Melbourne hackathon on Singular Learning Theory and alignment (Oct. 7-8), we did some experiments to estimate the learning coefficient of the single-layer modular addition task at a basin, an invariant that measures the... |
6aa19e9f-fa22-4f3f-84f6-db94f4354ec2 | trentmkelly/LessWrong-43k | LessWrong | Never Leave Your Room
Related to: Priming and Contamination
Psychologists define "priming" as the ability of a stimulus to activate the brain in such a way as to affect responses to later stimuli. If that doesn't sound sufficiently ominous, feel free to re-word it as "any random thing that happens to you can hijack y... |
ef483a62-72ba-4251-9dff-2f4578c7dd52 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Value Alignment Verification
1 Introduction
---------------
If we desire autonomous agents that can interact with and assist humans and other agents in performing complex, risky tasks, then it is important that humans can verify that these agents’ policies are aligned with what is expected and desired. This alignme... |
0de45f66-b3af-41a8-87d5-2be704417120 | trentmkelly/LessWrong-43k | LessWrong | LINK: Guinea worm disease close to eradication
The disease, that is, maybe not the worm itself. Anyway, Team Human scores its second point against Team Disease:
http://www.sciencealert.com/guys-we-re-really-close-to-eradicating-the-second-disease-ever-from-the-planet |
6e7d57c5-d732-4b93-b703-749b77b3fa1e | trentmkelly/LessWrong-43k | LessWrong | Russian x-risks newsletter fall 2021
It is estimated that excess morality during Covid pandemics will reach 1 million in Russia to the end of 2021, according to demograph A.Raksha. He also said that in October 2021 daily morality in Russia was highest since 1945. He uses numbers from local death-registration offices (... |
704fcbea-a99b-4d8b-94e3-151d35efd769 | trentmkelly/LessWrong-43k | LessWrong | [video] "Transhuman", featuring Sandberg and Bostrom
Transhuman, a 20-minute Dutch documentary about Anders Sandberg, is now available online. From Sandberg's blog:
> Here is the chance of see my beetle collection, my grubby kitchen, and an absolutely stunning combination between a supercomputer center and an oxfor... |
09b729f4-7c2d-44d9-b32f-25fdb02801ab | trentmkelly/LessWrong-43k | LessWrong | How good are LLMs at doing ML on an unknown dataset?
I just ran two evaluation tests on each of the three leading LLM chatbots, GPT4o, Claude Sonnet 3.5 and Gemini advanced. In the challenge the models were presented with a novel dataset, and were asked to develop a ML model to do supervised classification of the dat... |
db36c6b4-e0bb-4d71-8fe7-567688107857 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Any Philosophy PhD recommendations for students interested in Alignment Efforts?
I'm currently planning my Philosophy PhD applications.
If I'm interested in a traditional area of Philosophy (Epistemology, Ethics, Phil of Language), I can easily check recommendations from Philosophy Gourmet Reports, rely on testimoni... |
743b31cd-c12e-4ec9-8c88-e351454dd7b2 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | How harmful are improvements in AI? + Poll
This post was written by Marius Hobbhahn and Tilman Räuker.
Disclaimer: We have previously posted [this piece](https://forum.effectivealtruism.org/posts/LqWPSNRu2t46fv6hK/ai-acceleration-from-a-safety-perspective-trade-offs-and) on the EA forum. We now post it here because L... |
24136735-670d-461b-801c-387c5d73afe2 | trentmkelly/LessWrong-43k | LessWrong | How to Escape From the Simulation (Seeds of Science)
Seeds of Science (a scientific journal specializing in speculative and exploratory work) recently published a paper, "How to Escape From the Simulation" that may be of interest to the LW community.
Author: Roman Yampolskiy
Full text (open access): PDF
Abstract
M... |
63462f0d-0e17-40f4-8235-25fdb77a7ed7 | trentmkelly/LessWrong-43k | LessWrong | LINK: An example of the Pink Flamingo, the obvious-yet-overlooked cousin of the Black Swan
India vs. Pakistan: the nuclear option is dangerously close, and nobody seems to want to prevent it
http://qz.com/541502/a-nuclear-war-between-india-and-pakistan-is-a-very-real-possibility/ |
ba15c2e1-77c4-4c1d-9eb3-582f1a0d3fdc | trentmkelly/LessWrong-43k | LessWrong | Eliezer's Videos and More
I like watching videos of Eliezer talking and explaining things. So here is a list of videos I have discovered so far. I have not tried to make this list exhaustive. If you know of any more videos, please post them in the comments.
Eliezer Yudkowsky - Less Wrong Q&A Playlist
The main Elieze... |
5329ce21-7fc6-4719-b067-d7076e397ed3 | trentmkelly/LessWrong-43k | LessWrong | Sleeping Beauty Resolved?
,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
[This is Part 1. See also Part 2.]
Introduction
The Sleeping Beauty problem has been debated ad nauseum since Elga's original paper [Elga2000], yet no consensus ... |
a8d6c339-e33c-40b0-9e5d-cc46ed1f498f | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Safety standards: a framework for AI regulation
*Purpose: provide a public resource on safety standards that introduces the ideas I've heard people talk about and some new ideas. I don't know which ideas are new because the discourse has been largely non-public. I think safety standards provide a clear regulatory dire... |
826d1522-503c-4f85-911f-e8f4b9681f9a | trentmkelly/LessWrong-43k | LessWrong | Bayesian Nights (Rationalist Story Time)
Tell us a story. A tall tale for King Solamona, a yarn for the folk of Bensalem, a little nugget of wisdom, finely folded into a parable for the pages.
The game is simple:
1. Choose a bias, a fallacy, some common error of thought.
2. Write a short, hopefully entertaining... |
609e909d-4146-4402-8fc7-8e6308b9c6da | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Notes on Moderation, Balance, & Harmony
This post examines the virtues of **moderation**, **balance**, and **harmony**. I wrote this not as an expert on the topic, but as someone who wants to learn more about it. I hope it will be helpful to people who want to know more about these virtues and how to nurture them.
I ... |
98030159-baf1-4735-8e83-89888136993f | trentmkelly/LessWrong-43k | LessWrong | Ends Don't Justify Means (Among Humans)
> "If the ends don't justify the means, what does?"
> —variously attributed
>
> "I think of myself as running on hostile hardware."
> —Justin Corwin
Yesterday I talked about how humans may have evolved a structure of political revolution, beginning by believing... |
5ca2350a-c65e-4a3f-b340-17dbae90b3b3 | trentmkelly/LessWrong-43k | LessWrong | GWWC Pledge History
Giving What We Can members have pledged to donate at least 10% of what they earn to help others as best they can, but this is broader than it was originally. The pledge was specific to global poverty, and only became cause-neutral in late 2014 after a lot of discussion. I think that discussion is i... |
f3ee95b8-6d6a-4850-a4ba-caac42f8d9de | trentmkelly/LessWrong-43k | LessWrong | The problem of mankind indestructibility in disastrously unpredictable environment
The problem of mankind indestructibility in disastrously unpredictable environment
Concerning development of human race indestructibility roadmap
Kononov Alexandr Anatolievich, PhD (Engineering), senior researcher, Institute of System... |
a74ab878-7a14-4606-a91c-1b131cac2deb | StampyAI/alignment-research-dataset/arbital | Arbital | Axiom of Choice: Guide
## Learning the Axiom of Choice ##
[https://arbital.com/p/multiple-choice](https://arbital.com/p/multiple-choice)
[https://arbital.com/p/multiple-choice](https://arbital.com/p/multiple-choice)
[https://arbital.com/p/multiple-choice](https://arbital.com/p/multiple-choice)
[https://arbital.c... |
d3e3e7c6-218d-4e57-91bd-fc4c17cde312 | trentmkelly/LessWrong-43k | LessWrong | What are some of the best introductions/breakdowns of AI existential risk for those unfamiliar?
Discussion of AI risk has gone mainstream in the past few months, and as usual whenever that happens, many of the people new to the field think they're an expert and proceed to engage in the exact same arguments over and ov... |
e6300c8e-633f-4fc7-8f1d-f0e9f9c4b704 | trentmkelly/LessWrong-43k | LessWrong | How will this recession differ from the last two?
My model is that the last recession (2008) was essentially financial in nature- a bunch of wealth was destroyed on paper but we lost no productive capacity. The 2002 recession was similar- most of what happened was revealing that things were valueless, rather than dest... |
b2bbe09b-dda5-4b23-81ad-6a0fa84dafd8 | trentmkelly/LessWrong-43k | LessWrong | Practical Decision-Making Course: A Retrospective
This is a linkpost for Practical Decision-Making Retrospective, an article looking back on the successes and learnings from this course. This article is part two of a series of ongoing retrospectives on the progress of the Guild of the Rose, an organization for rationa... |
4bf4b3de-8493-4457-bfe1-02d8a136fb4e | trentmkelly/LessWrong-43k | LessWrong | Positive Expectations; how to build Hopefulness
The most hopeful I have ever felt was probably just before I set off to university. The least hopeful I have ever felt was probably around the start of the second wave of covid where I live, which also coincided with me realizing I really don't much like the chemistry de... |
3d3b5af5-c31f-413b-8957-73b4b5d771cf | trentmkelly/LessWrong-43k | LessWrong | Architects of Our Own Demise: We Should Stop Developing AI Carelessly
Some brief thoughts at a difficult time in the AI risk debate.
Imagine you go back in time to the year 1999 and tell people that in 24 years time, humans will be on the verge of building weakly superhuman AI systems. I remember watching the anime s... |
367cafb1-3b61-4d8a-bf45-ee063626e797 | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post3783
A putative new idea for AI control; index here . Corrigibility should allow safe value or policy change. Indifference allows the agent to accept changes without objecting. However, an indifferent agent is similarly indifferent to the learning process. Classical uncertainty over values has the opposite pro... |
1696cb6d-ab7b-43a6-be91-c7b14e7c3117 | StampyAI/alignment-research-dataset/aisafety.info | AI Safety Info | Would it improve the safety of quantilizers to cut off the top few percent of the distribution?
It certainly seems that doing something like this would help.
This idea isn’t discussed in Rob Miles’s [video](https://www.youtube.com/watch?v=gdKMG6kTl6Y) or [the paper](https://intelligence.org/files/QuantilizersSaferAl... |
d1a2f20f-da31-4504-90ea-b5c1061e5f03 | trentmkelly/LessWrong-43k | LessWrong | Three Kinds of Competitiveness
Crossposted from AI Impacts
[Epistemic status: I wrote this for Blog Post Day II. Sorry it's late.]
In this post, I distinguish between three different kinds of competitiveness -- Performance, Cost, and Date -- and explain why I think these distinctions are worth the brainspace they oc... |
cac6806a-340f-490e-bb47-0bfa0f43f79e | trentmkelly/LessWrong-43k | LessWrong | Complex Systems for AI Safety [Pragmatic AI Safety #3]
This is the third post in a sequence of posts that describe our models for Pragmatic AI Safety.
It is critical to steer the AI research field in a safer direction. However, it’s difficult to understand how it can be shaped, because it is very complex and there is... |
1a2407ff-518c-4cd6-a75f-5e68cfa7016f | StampyAI/alignment-research-dataset/lesswrong | LessWrong | How well did Manifold predict GPT-4?
Chat GPT-4 is already here!! Who could have seen that coming… oh wait Manifold (kinda) did? 😅
I thought I’d write a short piece on how Manifold Markets was used to predict the launch of GPT-4 and its attributes. Both its successes and its failures. Disclaimer I work at Manifold.
... |
6ba3312b-72b8-4dae-8dde-1fe53b356cd0 | trentmkelly/LessWrong-43k | LessWrong | The Dial of Progress
“There is a single light of science. To brighten it anywhere is to brighten it everywhere.” – Isaac Asimov
You cannot stand what I’ve become
You much prefer the gentleman I was before
I was so easy to defeat, I was so easy to control
I didn’t even know there was a war
– Leonard Cohen, There is a... |
abf477bb-73a6-4444-ac2c-6f0f18dd4e0e | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | How To Go From Interpretability To Alignment: Just Retarget The Search
When people talk about [prosaic alignment proposals](https://www.lesswrong.com/posts/fRsjBseRuvRhMPPE5/an-overview-of-11-proposals-for-building-safe-advanced-ai), there’s a common pattern: they’ll be outlining some overcomplicated scheme, and then ... |
ce855f23-8f32-4f10-b9f1-d8961fbaad35 | trentmkelly/LessWrong-43k | LessWrong | PSA for LW futurists/academics
Rutgers was planning to offer a "Future of Humankind" course via Coursera, but I received this announcement today:
"Hello Future of Humankind registrants,
It is with great sadness, that we must inform you that due to the untimely death of James Martin, we will be unable to offer the Fu... |
ffc49131-0879-4469-b2f5-8b6e3820478d | trentmkelly/LessWrong-43k | LessWrong | The Value of Uncertainty Reduction - references to academic literature appreciated
My intuition (probably widely shared) suggests that uncertainty about the future is stress inducing and reducing uncertainty about the future is helpful because it allows us to plan. So I started trying to invent thought experiments tha... |
71cdc1ff-b1fc-41df-b752-f93faecd7e7b | trentmkelly/LessWrong-43k | LessWrong | Sam Harris and the Is–Ought Gap
Many secular materialist are puzzled by Sam Harris's frequent assertion that science can bridge Hume's is–ought gap. Indeed, bafflement abounds on both sides whenever he debates his "bridge" with other materialists. Both sides are unable to understand how the other can fail to grasp ele... |
b9105a66-9340-4692-9ec1-4a0247acc575 | trentmkelly/LessWrong-43k | LessWrong | The "public debate" about AI is confusing for the general public and for policymakers because it is a three-sided debate
Summary of Argument: The public debate among AI experts is confusing because there are, to a first approximation, three sides, not two sides to the debate. I refer to this as a 🔺three-sided framew... |
9987141b-d0d8-48fe-bab6-8f73ac682ec5 | trentmkelly/LessWrong-43k | LessWrong | Model-driven feedback could amplify alignment failures
Anthropic, DeepMind and Google Brain are all working on strategies to train language models on their own outputs. For a brief summary of the work so far:
1. Red Teaming Language Models with Language Models (Perez et al., 2022). One model prompts another, seeking... |
906593e9-b79a-47d7-82d5-19b13d12f45d | StampyAI/alignment-research-dataset/arxiv | Arxiv | Finding Neurons in a Haystack: Case Studies with Sparse Probing
1 Introduction
---------------
Neural networks are often conceptualized as being flexible “feature extractors” that learn to iteratively develop and refine suitable representations from raw inputs [[1](#bib.bib1), [2](#bib.bib2)]. This begs the questi... |
1382aa81-422f-4472-95ab-98acd23ed78d | StampyAI/alignment-research-dataset/arxiv | Arxiv | Uncalibrated Models Can Improve Human-AI Collaboration
1 Introduction
---------------
In safety-critical settings like medicine, AI is often integrated as interactive feedback, allowing a human to decide when and to what extent the AI’s “advice” is utilized [[29](#bib.bib29)]. For example, in medical diagnosis task... |
ca25b372-9eea-4ca1-88e6-b5676ecbb048 | trentmkelly/LessWrong-43k | LessWrong | Mini-review: 'Proving History: Bayes' Theorem and the Quest for the Historical Jesus'
I recently received an advance review copy of historian and philosopher Richard Carrier's new book, Proving History: Bayes' Theorem and the Quest for the Historical Jesus.
The book belongs to a two-volume work on the Historical Jesu... |
b57ba66c-8f23-49d4-b344-73cbcedab76c | trentmkelly/LessWrong-43k | LessWrong | Best informative videos on the Internet
Inspired by Yair Halberstadt's similar question about non-textbooks, I'm asking this question to see if people have any videos that they would want to recommend to others.
The criterion is that the videos must be informative or educational. I don't want a playlist of your favor... |
b1651328-1a1b-44aa-ac3c-0c21c703ad46 | trentmkelly/LessWrong-43k | LessWrong | Paper: "A simple function of one parameter (θ) can fit any collection of ordered pairs {Xi,Yi} to arbitrary precision"--implications for Occam?
None |
5322ac4b-c6ad-40cb-a99a-bc86b291c77a | trentmkelly/LessWrong-43k | LessWrong | The (short) case for predicting what Aliens value
The case
Most of the future things we care about – i.e., (dis)value – come, in expectation, from futures where humanity develops artificial general intelligence (AGI) and colonizes many other stars (Bostrom 2003; MacAskill 2022; Althaus and Gloor 2016).
Hanson (2021)... |
f90a5ed7-dd2e-4876-ba55-172093e3ce95 | trentmkelly/LessWrong-43k | LessWrong | DRAFT: Three Intellectual Temperaments: Birds, Frogs and Beavers
Here is a draft of a potential top-level post which I'd welcome feedback on. I would appreciate any suggestions, corrections, additional examples, qualifications, or refinements.
Birds, Frogs and Beavers
The introduction of Birds and Frogs by Freeman... |
aebc0dc8-b4c6-44b5-8a16-0f80cd48d250 | trentmkelly/LessWrong-43k | LessWrong | The Generalized Product Rule
Imagine we have a company with investment projects A, B, C,....For instance, A might be a new high-speed Internet service, B might be a new advanced computer, C might be a new inventory management software, etc. We are interested in calculating the total return from these investments at th... |
87c4f2f0-e389-4229-9264-b13889aa8d45 | trentmkelly/LessWrong-43k | LessWrong | The Bottom Line
There are two sealed boxes up for auction, box A and box B. One and only one of these boxes contains a valuable diamond. There are all manner of signs and portents indicating whether a box contains a diamond; but I have no sign which I know to be perfectly reliable. There is a blue stamp on one box, fo... |
05922ac3-23b1-4caa-91fb-8b82f2b97e93 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Why Copilot Accelerates Timelines
> "Say we have intelligences that are narrowly human / superhuman on every task you can think of (which, for what it’s worth, I think will happen within 5-10 years). How long before we have self-replicating factories? Until foom? Until things are dangerously out of our control? Until ... |
39b37245-d974-4969-9a64-1604ab5c0508 | trentmkelly/LessWrong-43k | LessWrong | If you say "don't lose focus," what does that actually tell you to focus on?
Crossposted. (At this point the discussion on LW is nearly always better than the discussion on my own blog, and gives me more to think about afterwards. Thank you for that.)
If there’s any reason for you to watch these piano practice videos... |
d75a54fc-1804-40a1-86da-ca5b11dcaf12 | trentmkelly/LessWrong-43k | LessWrong | AI Takeover Scenario with Scaled LLMs
Introduction
In this AI takeover scenario, I tried to start from a few technical assumptions and imagine a specific takeover scenario from human-level AI.
I found this exercise very useful to identify new insights on safety. For instance, I hadn't clearly in mind how rivalries l... |
4f4fd9ff-38a4-48e4-99ba-61d0b1949d37 | trentmkelly/LessWrong-43k | LessWrong | Support for book on finding terminal goals and higher purpose
As part of my project of spreading rationality to a broad audience and thus raising the sanity waterline, I'm writing a book on using rationality-informed strategies to help people find terminal values, with an orientation toward encouraging a positive and ... |
5164a770-270a-4c9b-928e-f176e3992481 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] On Not Having an Advance Abyssal Plan
Today's post, On Not Having an Advance Abyssal Plan was originally published on 23 February 2009. A summary (taken from the LW wiki):
> Don't say that you'll figure out a solution to the worst case scenario if the worst case scenario happens. Plan it out in advance... |
64748958-9936-45b1-8600-2e7d8342aebf | trentmkelly/LessWrong-43k | LessWrong | Drowning children are rare
Stories such as Peter Singer's "drowning child" hypothetical frequently imply that there is a major funding gap for health interventions in poor countries, such that there is a moral imperative for people in rich-countries to give a large portion of their income to charity. There are simply ... |
f9938dc7-a347-4d9e-adc2-a0edb323e009 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Complex problems, limited information, and rationality; How should we make decisions in real life?
Discussion article for the meetup : Complex problems, limited information, and rationality; How should we make decisions in real life?
WHEN: 16 October 2013 07:00:45PM (-0700)
WHERE: West Los Angeles (At the W... |
fff70870-84b2-4c7e-8885-e01ea5e67fd1 | trentmkelly/LessWrong-43k | LessWrong | Some implications of radical empathy
Summary
In my first piece, I settled on conceiving of radical empathy as an object view. I highlight some potential moral implications of object views:
1. More of what someone finds pleasant or is (knowingly or unknowingly) disposed to find pleasant is better for them, all else e... |
75f4d0b8-96ab-4eb3-a825-28e9eba9dd35 | trentmkelly/LessWrong-43k | LessWrong | Posting May Slow
Greetings, fearless readers:
Due to the Oxford conference on Global Catastrophic Risk, I may miss some posts - possibly quite a few.
Or possibly not.
Just so you don't think I'm dead.
Sincerely,
Eliezer. |
568ebca7-8956-4181-9346-822b5ff98890 | trentmkelly/LessWrong-43k | LessWrong | Reopen Protest Sign Survey
With the protests to end the lockdowns, the main pictures I've seen in my social media bubble have been ones with signs like " I need a haircut" or " Let my People Golf". So it's not surprising that people would be writing responses like annacaffeina's:
> People aren't out protesting to b... |
99b99066-b4d3-4de6-b5b5-f593dbaeaf09 | StampyAI/alignment-research-dataset/special_docs | Other | The Myth of the Rational Voter: Why Democracies Choose Bad Policies
[]
THE MYTH OF THE RATIONAL VOTER
THE MYTH OF
THE RATIONAL VOTER
WHY DEMOCRACIES CHOOSE BAD POLICIES
With a new preface by the author
BRYAN CAPLAN
[pub]
Copyright © 2007 by Princeton University Press
Published by Princeton University Press, 4... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.