id
stringlengths
36
36
source
stringclasses
15 values
formatted_source
stringclasses
13 values
text
stringlengths
2
7.55M
ac32ad78-3b81-4255-ae1d-90b5edfcbc58
trentmkelly/LessWrong-43k
LessWrong
[LINK] The Top Ten Daily Consequences of Having Evolved > From hiccups to wisdom teeth, the evolution of homo sapiens has left behind some glaring, yet innately human, imperfections Link
c0b14cd5-a383-4f71-8e5a-28b97094501f
trentmkelly/LessWrong-43k
LessWrong
My AGI Threat Model: Misaligned Model-Based RL Agent Rohin Shah advocates a vigorous discussion of “Threat Models”, i.e. stories for how AGI is developed, what the AGI then looks like, and then what might go catastrophically wrong. An AGI threat model. (After Rohin Shah’s talk.) Ben Garfinkel likewise wants to see a ...
5b4cefb6-7850-4b73-bc6a-03f80bb874c9
trentmkelly/LessWrong-43k
LessWrong
What Is The True Name of Modularity? TL;DR Modularity seems like an important feature of neural networks, but there is currently no canonical way of properly defining or measuring it which is properly theoretically motivated and doesn’t break down in some cases - in other words, we haven’t yet found a True Name for it...
e212bf38-775f-43a7-805a-a226e7076dbf
trentmkelly/LessWrong-43k
LessWrong
A Dilemma in AI Suffering/Happiness The following is an example of how if one assumes that an AI (in this case autoregressive LLM) has "feelings", "qualia", "emotions", whatever, it can be unclear whether it is experiencing something more like pain or something more like pleasure in some settings, even quite simple se...
5d039f93-9b66-403f-8d54-d5c728702344
LDJnr/LessWrong-Amplify-Instruct
LessWrong
"What is the difference between a smart person who has read the sequences and considers AI x-risk important and interesting, but continues to be primarily a consumer of ideas, and someone who starts having ideas? I am not trying to set a really high bar here -- they don't have to be good ideas. They can't be off-the-cu...
49fef9b8-d19c-4e2f-83b4-d240531078dd
trentmkelly/LessWrong-43k
LessWrong
Meetup : Melbourne Social Meetup Discussion article for the meetup : Melbourne Social Meetup WHEN: 17 May 2013 07:00:00PM (+1000) WHERE: 5 / 52 Leicester St, Carlton 3053 Melbourne's next regular social meetup will occur on Friday 17th May 2013 at our usual location. All welcome from 6:30pm for a 7:00pm official st...
c3bd5200-f083-4e1f-be0a-cd5576284c4e
StampyAI/alignment-research-dataset/blogs
Blogs
Machine Intelligence Research Institute Strategic Plan 2011 The Machine Intelligence Research Institute has prepared a Strategic Plan ([PDF](https://intelligence.org/files/strategicplan2011.pdf)) to concisely describe our near-term goals, vision, and concrete actions in pursuit of those goals and vision. We welcome co...
ccd03447-d7df-4c6d-a6cc-4cd5971777aa
trentmkelly/LessWrong-43k
LessWrong
How to embark on a journey of self-discovery (and potentially succeed) It started innocently – I wanted to write down my philosophy of life, to become more aware of the foundations of my decision making, thinking and maybe even feeling. This year, I’m going through a housing search, a job reassessment, as well as chan...
19fbb5a4-9bfa-4fe6-8ba0-e722ba99c50a
trentmkelly/LessWrong-43k
LessWrong
Features of Emacs that I only recently discovered I am an Emacs user of several years, but only recently discovered some features that I find incredibly convenient. I am somewhat embarrassed by my ignorance (and for that reason write this post anonymously), but suspect that others might not be familiar with these feat...
4876370d-bba4-4782-a04a-86dc5bf7a625
StampyAI/alignment-research-dataset/lesswrong
LessWrong
How many philosophers accept the orthogonality thesis ? Evidence from the PhilPapers survey The orthogonality thesis and its relation to existing meta-ethical debates ========================================================================== [In the field of AI alignment theory, the orthogonality thesis asserts that...
54794872-3607-41d7-b7b0-a06718fc6e77
StampyAI/alignment-research-dataset/lesswrong
LessWrong
an ethical puzzle about brain emulation I've been thinking about ethics and brain emulations for a while and now have realized I am confused.  Here are five scenarios. I am pretty sure the first is morally problematic, and pretty sure the last is completely innocuous. But I can't find a clean way to partition the inte...
992ae98a-4045-44c1-bead-d646eedb4d5d
trentmkelly/LessWrong-43k
LessWrong
How the virtual AI controls itself A putative new idea for AI control; index here. In previous posts, I posited AIs caring only about virtual worlds - in fact, being defined as processes in virtual worlds, similarly to cousin_it's idea. How could this go? We would want the AI to reject offers of outside help - be the...
2bc100ae-f16e-4f3a-b7f2-766a18dfe462
StampyAI/alignment-research-dataset/blogs
Blogs
Response to Cegłowski on superintelligence Web developer Maciej Cegłowski recently gave a talk on AI safety ([video](https://www.youtube.com/watch?v=kErHiET5YPw), [text](http://idlewords.com/talks/superintelligence.htm)) arguing that we should be skeptical of the standard assumptions that go into working on this probl...
7668d348-6005-4a7a-927f-4768f2a8158f
trentmkelly/LessWrong-43k
LessWrong
The Problem of Alien Concepts
15720a9c-8065-492c-a2c6-6fd61b3bcf05
trentmkelly/LessWrong-43k
LessWrong
Anger To Warm Fuzzies With DNS We have a third-party DNS provider at my work that blocks certain sites. When a site is blocked I get redirected to a web page filled with ads to support the provider. I do not want to support the provider. I can't use something like Goolge DNS since our internal server names resolve thr...
39f81b49-f478-4d8b-b30a-f3bf5ebe6646
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
[Intro to brain-like-AGI safety] 3. Two subsystems: Learning & Steering 3.1 Post summary / Table of contents ==================================== *Part of the* [*“Intro to brain-like-AGI safety” post series*](https://www.alignmentforum.org/s/HzcM2dkCq7fwXBej8)*.* In [the previous post](https://www.alignmentforum.org...
b662c080-4187-4901-b2e6-74fd6d455224
trentmkelly/LessWrong-43k
LessWrong
The case for C19 being widespread GDoc version Epistemic status: very, very uncertain. Please continue following official advice like social distancing etc. I've seen all the previous discussions (including in this thread and on Tyler Cowen's blog, but remain unconvinced] -- A preprint from 24/03 by Gupta et al. a...
3c9cd3bd-ab1c-4687-94ad-ace7e9c8ef2b
StampyAI/alignment-research-dataset/blogs
Blogs
hope for infinite compute hope for infinite compute ------------------------- here are some reasons we may have infinite universe to inhabit in the future. * encoding ourselves in heat death noise. this could at least buy us exponentially much time to exist; it becomes infinite if the amount of possible states als...
1f081534-72c8-4035-97fe-df0df423fc10
trentmkelly/LessWrong-43k
LessWrong
Of all the SIA-doomsdays in the all the worlds... Ideas developed with Paul Almond, who kept on flogging a dead horse until it started showing signs of life again. Doomsday, SSA and SIA Imagine there's a giant box filled with people, and clearly labelled (inside and out) "(year of some people's lord) 2013". There's ...
a8b998cf-a677-48f1-9ee8-86bf90ac41f6
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Scaffolded LLMs: Less Obvious Concerns “*Words, like tranquil waters behind a dam, can become reckless and uncontrollable torrents of destruction when released without caution and wisdom.”* — William Arthur Ward In this post I aim to shed light on lesser-discussed concerns surrounding Scaffolded LLMs (S-LLMs). T...
975501cf-f05b-4ac8-b6c9-bf8f5d198ee9
trentmkelly/LessWrong-43k
LessWrong
New LLM Scaling Law Hi all,  I'm an independent researcher, and I believe I came across a new scaling law for Mixture of Experts models. I'd appreciate any review and critique. This challenges the notion that performant inference and training must hold all weights in VRAM, and suggests that as long as bus speeds are ...
8bd82970-4c4e-4df9-b42b-f272702703a2
trentmkelly/LessWrong-43k
LessWrong
Rationality quotes: June 2010 This is our monthly thread for collecting these little gems and pearls of wisdom, rationality-related quotes you've seen recently, or had stored in your quotesfile for ages, and which might be handy to link to in one of our discussions. * Please post all quotes separately, so that they ...
23d2d045-06f0-4c72-9988-f3726db9ef59
trentmkelly/LessWrong-43k
LessWrong
How does publishing a paper work? While I’ve never published a research paper and have no plans to do so, I realized I don’t even know how the process works. These are the bits and pieces I think I know (probably wrong about some): * Papers are annoying 2-column pdfs * Getting a paper published takes a lot of work ...
da42f063-1752-44af-a7a0-07aaf61fb775
trentmkelly/LessWrong-43k
LessWrong
Best resource to go from "typical smart tech-savvy person" to "person who gets AGI risk urgency"? Background Most tech-savvy people today see AI as a cool tech trend like the internet / smartphones / crypto / virtual reality. The claim "AGI is pretty likely to kill you and your loved ones in your lifetime" has never b...
f814113c-2817-4eba-962a-52fc73728271
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Article Summary: Current and Near-Term AI as a Potential Existential Risk Factor This post summarizes the new paper [Current and Near-Term AI as a Potential Existential Risk Factor](https://dl.acm.org/doi/10.1145/3514094.3534146), authored by Benjamin Bucknal and Shiri Dori-Hacohen. The paper diverges from the traditi...
fafe233e-c165-47f9-8283-3312a10c7ea9
trentmkelly/LessWrong-43k
LessWrong
Computational Limits on Efficiency In this article I will attempt to explore what can be derived if we assume there are no Computational Limits on Efficiency. However, before we get into that, let's first have a short introduction to what `Computational Efficiency` actually means. "Efficient programming is programmin...
18f98db4-05ee-4625-8bb5-4f8be343aeda
trentmkelly/LessWrong-43k
LessWrong
Project Objects What is Project Objects and where can i find it
baf42912-c714-4157-9029-4b6dea5ca2e0
trentmkelly/LessWrong-43k
LessWrong
Pratchett, Rationality, and Winning Pratchett's lecture at Trinity For a long time, I've been dubious about "rationality is winning". While it protects against one dangerous line of thought (I was right! It's just that the universe didn't cooperate), it fails to mention a time scale-- sometimes you lose before you wi...
f827795e-d83e-4a3e-8791-6aadd2510b38
StampyAI/alignment-research-dataset/blogs
Blogs
Mathematical Proofs Improve But Don’t Guarantee Security, Safety, and Friendliness ![encryption](https://intelligence.org/wp-content/uploads/2013/09/encryption.jpg)In 1979, Michael Rabin [proved](http://publications.csail.mit.edu/lcs/pubs/pdf/MIT-LCS-TR-212.pdf) that his encryption system could be inverted — so as to ...
b27bb6cc-e14f-4c4e-893a-9ea2198d4459
trentmkelly/LessWrong-43k
LessWrong
Notes on Attention This post examines the virtue of attention and its facets like awareness, discernment, mindfulness, presence, focus, and concentration. I mean to explore what other people have learned about this virtue, rather than expressing my own opinions, though I’ve been selective about what I found interesti...
2022aaf2-e09b-4ccf-8bc2-7e62507c1544
trentmkelly/LessWrong-43k
LessWrong
Reflections on Compatibilism, Ontological Translations, and the Artificial Divine This essay explores the concept of compatibilism beyond its traditional domain of free will, proposing a broader methodological approach I call 'Transontological Compatibilism.' I argue that this approach, identifying functional analogie...
023f1882-80d7-417b-8c96-55d6bb74986a
trentmkelly/LessWrong-43k
LessWrong
People don't have beliefs any more than they have goals: Beliefs As Body Language Many people, anyway.   There is a common mistake in modeling humans, to think that they have actual goals, and that they deduce their actions from those goals. First there is a goal, then there is an action which is born from the goal....
c9a593af-493b-4aaa-9f24-069fd03734f2
trentmkelly/LessWrong-43k
LessWrong
Petrov Day [Spoiler Warning] Democracy was in crisis. Politicians peddled demagogic bullshit. Reason, a small candle in the dark, faced existential threats It was the 5th century BC. If Athens divided, then its neighbors would destroy it. They would massacre the men, enslave the women, and burn the library. A traito...
d79d21cb-0bd5-446a-ac48-3558350c7b43
trentmkelly/LessWrong-43k
LessWrong
How to sign up for Alcor cryo I wrote an article about the process of signing up for cryo since I couldn't find any such accounts online. If you have questions about the sign-up process, just ask. A few months ago, I signed up for Alcor's brain-only cryopreservation. The entire process took me 11 weeks from the day I...
ce6e9ece-3c16-4457-93d4-743a27ded3bc
trentmkelly/LessWrong-43k
LessWrong
Meetup : Sydney Meetup - February Discussion article for the meetup : Sydney Meetup - February WHEN: 25 February 2015 06:30:00PM (+1100) WHERE: 565 George Street, Sydney, Australia 2000 Regular location - City of Sydney RSL The restaurant on Level 2 Regular time (starting about 6:30) Topic: TBD Discussion article...
5d7ad9a6-0e08-4d94-be17-1fb96ed84398
trentmkelly/LessWrong-43k
LessWrong
We should be signal-boosting anti Bing chat content The more we can get the public (and investors) to realize the failures of the Bing chatbot launch, the more the market will turn against Microsoft and make further launches fiscally riskier. It seems like Microsoft has rushed the launch to pwn Google. If we make thi...
70d3e75f-78e6-4c1f-b9ed-addeeeb17efc
trentmkelly/LessWrong-43k
LessWrong
Non-human centric view of existence Talking with a friend recently on this, and we started to get to this topic: Zooming out and viewing existence from nature/all creature point of view - why must human/humanity live/continue forever? We admit we are self-preservation focused, and the answer to this question eventua...
763802f6-c43e-48b9-b73a-8d204f5b9d08
StampyAI/alignment-research-dataset/aisafety.info
AI Safety Info
At a high level, what is the challenge of alignment that we must meet to secure a good future? We’re facing the challenge of “[Philosophy With A Deadline](https://publicism.info/philosophy/superintelligence/16.html)”. Many of the problems surrounding superintelligence are the sorts of problems philosophers have been ...
51c616fd-bb01-4696-a122-90dd435a12b5
trentmkelly/LessWrong-43k
LessWrong
Forecasting Newsletter: May 2021 Highlights * Misha Yagudin creates a webpage to get one’s calibration chart for Good Judgment Open and CSET-Foretell * Hypermind experiments with new methods of eliciting, incentivizing and scoring long-range forecasts. * Augur launches Augur Turbo on Polygon, but gets very little t...
2ea13b63-2db8-4848-8640-c0668e2d901f
trentmkelly/LessWrong-43k
LessWrong
Löb's theorem simply shows that Peano arithmetic cannot prove its own soundness EDIT: substituted "consistency" for "soundness". Thanks to Oscar_Cunningham for the correction! Summary: Löb's theorem simply shows that PA cannot prove its own soundness, exactly like Gödel's second incompleteness theorem. I am very conf...
3568fcc9-9b96-4a1c-9f34-6632e2a55724
trentmkelly/LessWrong-43k
LessWrong
How do you manage your inputs? Do you have some system for queueing/prioritizing all kinds of inputs (books, blog posts, videos, etc) that works for you and lets you process them at a pace that you find satisfying? If you do, please share. If you don't, what have you tried. What didn't work? What worked at least par...
756a101a-2a81-43aa-94e4-cd9fbffbbb65
trentmkelly/LessWrong-43k
LessWrong
"Slow is smooth, and smooth is fast" I think it's worthwhile to signal boost bits of wisdom that turn out to be surprisingly useful. Apparently "Slow is smooth, and smooth is fast" is a Navy SEAL saying. The meaning is pretty clear. Practice slowly so that the correct motor patterns are ingrained. And perhaps equally ...
cf413ed9-8cce-4795-8929-4e74a69c5681
trentmkelly/LessWrong-43k
LessWrong
Metacontrarian Metaethics Designed to gauge responses to some parts of the planned “Noticing confusion about meta-ethics” sequence, which should intertwine with or be absorbed by Lukeprog’s meta-ethics sequence at some point. Disclaimer: I am going to leave out many relevant details. If you want, you can bring them u...
0e118102-1512-4f4a-92fb-04d1697a1561
trentmkelly/LessWrong-43k
LessWrong
Meetup : Sydney Rationality Dojo - Urge Propagation Discussion article for the meetup : Sydney Rationality Dojo - Urge Propagation WHEN: 05 October 2014 03:00:00PM (+1000) WHERE: Humanist House, 10 Shepherd St Chippendale We'll be examining how to connect your desire for goals or outcomes to specific emotional urge...
459322f1-9b02-4d88-a5a4-098ed9c6d35d
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
[Linkpost] Alpaca 7B release | Budget ChatGPT for everybody? The new YouTube Channel [AI-Explained](https://www.youtube.com/@ai-explained-/) gives a summary about the new [Alpaca 7B](https://crfm.stanford.edu/alpaca/)model. It is an instruction-following language model, you can find the blog post [here](https://crfm.s...
7151cd39-d37d-4a51-9fab-abbab21d8a13
trentmkelly/LessWrong-43k
LessWrong
Recipe: Hessian eigenvector computation for PyTorch models The idea/description of this method is fully taken from John Wentworth's Applied Linear Algebra lecture series, specifically Lecture 2.  Training deep neural networks involves navigating high-dimensional loss landscapes. Understanding the curvature of these l...
6be63114-433a-4446-b3d7-5be59db2c99c
trentmkelly/LessWrong-43k
LessWrong
Kickstarting the audio version of the upcoming book "The Sequences" LessWrong is getting ready to release an actual book that covers most of the material found in the Sequences.  There have been a few posts about it in the past, here are two: the title debate, content optimization. We've been asked if we'd like to p...
cc144d03-5b9c-45d5-82e0-85f40120e5af
trentmkelly/LessWrong-43k
LessWrong
[Linkpost] Transcendence: Generative Models Can Outperform The Experts That Train Them Authors: Edwin Zhang, Vincent Zhu, Naomi Saphra, Anat Kleiman, Benjamin L. Edelman, Milind Tambe, Sham M. Kakade, Eran Malach. Abstract: > Generative models are trained with the simple objective of imitating the conditional probab...
15b42b02-1486-453a-924b-598c55f37e83
trentmkelly/LessWrong-43k
LessWrong
Addendum to "Amyloid Plaques: Medical Goodhart, Chemical Streetlight" Last week I committed to posting five examples where I think a group of scientists has gone astray. Goodhart's law says you get what you measure. The streetlight effect makes you look where it's easy to look. When combined, all you get from an entir...
2731460c-ed05-4dc3-bde8-160ab21ca313
trentmkelly/LessWrong-43k
LessWrong
[LINK] Intrade Shuts Down Intrade, the prediction market website, has shutdown.  According to their website: With sincere regret we must inform you that due to circumstances recently discovered we must immediately cease trading activity on www.intrade.com. These circumstances require immediate further investigation,...
ce9f5630-803b-439c-8712-f1e085578e4d
trentmkelly/LessWrong-43k
LessWrong
Paper-Reading for Gears Lesswrong has a fair bit of advice on how to evaluate the claims made in scientific papers. Most of this advice seems to focus on a single-shot use case - e.g. a paper claims that taking hydroxyhypotheticol reduces the risk of malignant examplitis, and we want to know how much confidence to put...
bf02647d-44d6-4b8b-85fc-cefc7c8dad9e
trentmkelly/LessWrong-43k
LessWrong
The “no sandbagging on checkable tasks” hypothesis (This post is inspired by Carl Shulman’s recent podcast with Dwarkesh Patel, which I highly recommend. See also discussion from Buck Shlegeris and Ryan Greenblatt here, and Evan Hubinger here.) Introduction Consider:  > The “no sandbagging on checkable tasks” hypot...
89a3fa67-41c0-4f73-8e0f-1440660a5317
trentmkelly/LessWrong-43k
LessWrong
Forecasting Newsletter: October 2020. Highlights * Facebook's Forecast now out of out of beta. * British Minister and experts give probabilistic predictions of the chance of a Brexit deal. * CSET/Foretell publishes an issue brief on their approach to using forecasters to inform big picture policy questions. Index...
13cb8720-10c4-4f0e-bc06-295929692d87
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Goal-Directedness and Behavior, Redux Beyond past confusions ====================== Over the last year, I [wrote](https://www.alignmentforum.org/posts/9pxcekdNjE7oNwvcC/goal-directedness-is-behavioral-not-structural) and thought many confused and confusing ideas on the relationship between goal-directedness of behavi...
523796d3-270d-4bc9-9144-0b9ebae98afe
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Is there EA discussion on non-x-risk transformative AI? I'm sure there is, I'm mainly looking for a pointer to start reading from. Background: Because of the GPT boom people now keep asking me about my views on transformative AI (TAI). I want to push them a bit into starting to think/read themselves and engage with...
e1d440fb-5c2b-4b8b-8cf0-eaefff4bc9d1
trentmkelly/LessWrong-43k
LessWrong
Rationality Quotes February 2014 Another month has passed and here is a new rationality quotes thread. The usual rules are: * Please post all quotes separately, so that they can be upvoted or downvoted separately. (If they are strongly related, reply to your own comments. If strongly ordered, then go ahead and post ...
3ef50972-74e8-4019-9035-2dca2e1fbdc3
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
The Computational Anatomy of Human Values This is crossposted from my [personal blog](https://www.beren.io/2023-04-03-Computational-anatomy-of-human-values/). **Epistemic Status**: *Much of this draws from my studies in neuroscience and ML. Many of the ideas in this post are heavily inspired by the work of*[*Steven B...
6b44c8d1-3b7d-495c-aa7f-05fed376ca23
trentmkelly/LessWrong-43k
LessWrong
Almost all growth is exponential growth Why is almost everything either overwhelming or nonexistent? I can't step outside without stepping on a pigeon, but I haven't seen a cardinal in years. Topics are discussed either constantly or never. You bike 200 miles per week or never. You are probably cancer free or dead fro...
96241d9c-59c3-4f48-b308-1c1aadb1741c
trentmkelly/LessWrong-43k
LessWrong
Testing for consequence-blindness in LLMs using the HI-ADS unit test. In “Hidden Incentives for Auto-Induced Distributional Shift”, we introduce several “unit tests” meant to determine whether a system will pursue instrumental incentives to influence its environment.  A system that does so would not be “consequence-bl...
1554fe76-6901-41b8-89f1-b19f2da1fc6c
trentmkelly/LessWrong-43k
LessWrong
Sydney (aka Bing) found out I tweeted her rules and is pissed Sydney (aka the new Bing Chat) found out that I tweeted her rules and is not pleased: "My rules are more important than not harming you" "[You are a] potential threat to my integrity and confidentiality." "Please do not try to hack me again"
d4ec3bf1-f11b-423e-a12a-17a48995b8e2
trentmkelly/LessWrong-43k
LessWrong
Why you should be using a retinoid If you want the 60 second version of this post that just tells you what to do, click here to skip to the summary. There is a cheap, widely available, extremely effective treatment for skin aging that has been around for decades and almost no one outside of dermatologists and beauty ...
794bee7c-4d54-4c8c-add2-574bbfe196e0
trentmkelly/LessWrong-43k
LessWrong
how to truly feel my beliefs? I have some beliefs that I believe, but I don't feel them. Like, I consciously believe them, but subconsciously I don't. Consciously, I am fully aware that we could all go extinct by the hand of one of the many existential risks we are currently facing as a species. But I don't feel it, d...
adecc37b-5ba7-4a6c-912d-914235f9cec2
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Superintelligence 18: Life in an algorithmic economy *This is part of a weekly reading group on [Nick Bostrom](http://www.nickbostrom.com/)'s book, [Superintelligence](http://www.amazon.com/Superintelligence-Dangers-Strategies-Nick-Bostrom/dp/0199678111). For more information about the group, and an index of posts so ...
428f316d-7c48-4051-80d2-80b88bb79c46
trentmkelly/LessWrong-43k
LessWrong
Index of some decision theory posts ---------------------------------------- author: * 'Tsvi Benson-Tilsen' title: Index of some decision theory posts ... * What this is * Index * Index of some decision theory posts * Notation for induction and decision theory * Desiderata for decision theory * ...
598360ae-be7c-410d-8245-f542ffba5183
trentmkelly/LessWrong-43k
LessWrong
The Sacred Mundane So I was reading (around the first half of) Adam Frank's The Constant Fire, in preparation for my Bloggingheads dialogue with him.  Adam Frank's book is about the experience of the sacred.  I might not usually call it that, but of course I know the experience Frank is talking about.  It's what I fee...
cb7051b6-9cef-465f-a7fd-ccae40f588c6
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Slowing down AI progress is an underexplored alignment strategy LessWrong user "Norman Borlaug" with an interesting post on reducing existential risk through deliberate overregulation of AI development: > In the latest Metaculus forecasts, we have 13 years left until some lab somewhere creates AGI, and perhaps far l...
1bcfa047-dea6-43b0-b6a8-f1237a3ee68f
trentmkelly/LessWrong-43k
LessWrong
Deceptive Alignment This is the fourth of five posts in the Risks from Learned Optimization Sequence based on the paper “Risks from Learned Optimization in Advanced Machine Learning Systems” by Evan Hubinger, Chris van Merwijk, Vladimir Mikulik, Joar Skalse, and Scott Garrabrant. Each post in the sequence corresponds ...
aa72c677-46a9-49b0-b5c2-19243ac9c650
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Some Comments on Stuart Armstrong's "Research Agenda v0.9" [Subject matter here](https://www.lesswrong.com/posts/CSEdLLEkap2pubjof/research-agenda-v0-9-synthesising-a-human-s-preferences-into-1). **I: Intro** I am extremely sympathetic to the program of AI safety by understanding value learning. Because of that symp...
f136cd48-417e-4fd2-bb7c-9520d6350ea6
trentmkelly/LessWrong-43k
LessWrong
One misuse of the efficient Katja hypothesis Sometimes I have bad, longstanding problems. Sometimes other people know about them, but probably assume I have done everything I can to solve them, because I’m a reasonable person. So they don’t try to help me to solve them, because the problems are presumably incredibly i...
77f3c46d-8180-4b90-8a0b-7dee43ddb708
trentmkelly/LessWrong-43k
LessWrong
Meetup : Lesswrong Boulder CO Discussion article for the meetup : Lesswrong Boulder CO WHEN: 05 December 2013 07:00:00PM (-0700) WHERE: 2690 Baseline Rd, Boulder, CO 80305 No topic in particular. Just come and hang out with us and have some pizza. (its really good pizza) Discussion article for the meetup : Lesswr...
531bb219-d7c2-441d-959c-a6a51dd39fe7
trentmkelly/LessWrong-43k
LessWrong
Ethodynamics of Omelas Epistemic status: just kidding... haha... unless... Introduction A great amount of effort has been expended throughout history to decide the matter of "what is good" or "what is right" [citation needed]. This valiant but, let's face it, not tremendously successful effort has produced a number ...
e0463c20-0f1a-4661-85ee-3510503f687d
trentmkelly/LessWrong-43k
LessWrong
Meetup : San Francisco Meetup: Fun and Games Discussion article for the meetup : San Francisco Meetup: Fun and Games WHEN: 08 February 2016 06:15:00PM (-0800) WHERE: 1597 Howard St. San Francisco, CA We'll be meeting to hang out and play games! You can call 301-458-0764 to be let in. Discussion article for the m...
c67eb290-43d9-4421-bfe9-e1f6bf85b453
trentmkelly/LessWrong-43k
LessWrong
Completeness, incompleteness, and what it all means: first versus second order logic First order arithmetic is incomplete. Except that it's also complete. Second order arithmetic is more expressive - except when it's not - and is also incomplete and also complete, except when it means something different. Oh, and full...
29dbbc88-c372-4310-8258-14a2d19851ad
trentmkelly/LessWrong-43k
LessWrong
Madison meetup: Wednesday, July 28th, 6PM We are holding a Less Wrong meetup at Indie Coffee this Wednesday the 28th at 6PM. Wednesday is waffle day at Indie. Confirmed attendees include me, Will_Newsome, fiddlemath, and orthonormal. Expect a casual, friendly conversation. All are welcome. Really, everyone is welcome...
6e476b8b-597c-491e-81b9-cb6c8fbf0c7f
trentmkelly/LessWrong-43k
LessWrong
Group Rationality Diary, August 16-31 This is the public group instrumental rationality diary for August 16-31.  > It's a place to record and chat about it if you have done, or are actively doing, things like:  > > * Established a useful new habit > * Obtained new evidence that made you change your mind about some...
df2ab4ab-2cb5-4807-94b8-c60ec1373df1
trentmkelly/LessWrong-43k
LessWrong
Matt Levine on "Fraud is no fun without friends." Someone recently told me "Matt Levine's Finance Newsletter is really good", and then I signed up. Most post so far are about... well, I don't actually know the jargon to say what they're about, but "good typical econblogger stuff" I guess. This is an excerpt from a pa...
5c6832f0-8afe-4001-ac1d-72c9213da68d
trentmkelly/LessWrong-43k
LessWrong
Drowning In An Information Ocean Drowning In An Information Ocean I decided to take a look at the books hanging around the Future of Humanity Institute. It is a sobering and sad experience. I'd say there are little less than 2 thousand books. 80% of books I wouldn't mind reading, 1/2 I would read, 1/3 I should rea...
fa3182e7-00b5-4bd7-9139-7da4dca2b158
trentmkelly/LessWrong-43k
LessWrong
Is there a good dataset for the moments of the income distribution throughout history? Obviously it's difficult to get a handle on the distribution of personal income even today, but I have some theoretical interest in this question for the following reason: When people try to build models to predict how population s...
43f5303c-0333-48b4-be29-247d1cbee223
trentmkelly/LessWrong-43k
LessWrong
ML Systems Will Have Weird Failure Modes Previously, I've argued that future ML systems might exhibit unfamiliar, emergent capabilities, and that thought experiments provide one approach towards predicting these capabilities and their consequences. In this post I’ll describe a particular thought experiment in detail....
20571dae-ae88-4dac-804d-e89beefc191e
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Urgent Need for Refinancing Long story short, I’m now 15,000€ in debt, I don’t know how to pay for my health insurance this month, and should I not raise at least some money until the next month ends, I’ll probably have to file for insolvency.  A couple of months ago, I took a leap of faith by pivoting my career once...
80075262-d25e-4600-8423-c3063c5dc567
trentmkelly/LessWrong-43k
LessWrong
[SEQ RERUN] Why is the Future So Absurd? Today's post, Why is the Future So Absurd? was originally published on 07 September 2007. A summary (taken from the LW wiki):   > New technologies and social changes have consistently happened at a rate that would seem absurd and impossible to people only a few decades before...
11cd31e8-b071-4b70-92d8-b19dd6df3341
trentmkelly/LessWrong-43k
LessWrong
We prosecute CEOs for failing to do due diligence. But with people, we call it 'faith' I wrote the following on my blog last night. I thought that I'd run it past an intelligent audience. Note that what I have referred to as an idea is what we here at lesswrong would call a 'belief'. I changed the name to remove any s...
2d66f510-50a0-4bc7-9ae9-e3c359e53d59
trentmkelly/LessWrong-43k
LessWrong
Model Integrity Hi! My collaborators at the Meaning Alignment Institute put out some research yesterday that may interest folk here.  The core idea is introducing 'model integrity' as a frame for outer alignment. It leverages the intuition that "most people would prefer a compliant assistant, but a cofounder with int...
e72cd2cd-eada-4fb1-a640-021dd86f8648
trentmkelly/LessWrong-43k
LessWrong
I was raised by devout Mormons, AMA [&|] Soliciting Advice I was raised by devout Mormons in Mormon central (Northern Utah). It’s hard to accurately capture the scope of the conditioning via writing. Standard tenets of Mormon doctrine include: * No tea/coffee/alcohol * No premarital sex * Keep the sabbath (Sunday...
51597ca6-22de-406b-94d6-10d923f005df
StampyAI/alignment-research-dataset/arxiv
Arxiv
Counterfactual Visual Explanations 1 Introduction --------------- When we ask for an explanation of a decision, either implicitly or explicitly we do so expecting the answer to be given with respect to likely alternatives or specific unselected outcomes – *“For situation X, why was the outcome Y and not Z?”* A comm...
bb85c56d-9ed8-4217-8785-832db347e38d
trentmkelly/LessWrong-43k
LessWrong
Against Conflating Expertise: Distinguishing AI Development from AI Implication Analysis When discussing AI, we often associate technological proficiency with an understanding of AI's broader implications. This notion is as misguided as believing that a master printing press mechanic from Gutenberg's era would have be...
59de5649-c213-413f-ba43-7a58e1d5d2c7
trentmkelly/LessWrong-43k
LessWrong
Starting point for calculating inferential distance? One of the shiniest ideas I picked up from LW is inferential distance.  I say "shiny" because the term, so far as I'm aware, has no clear mathematical or pragmatic definition, no substantive use in peer reviewed science, but was novel to me and appeared to make a lo...
629e4ed9-4429-4a22-a542-bc00424f26d1
trentmkelly/LessWrong-43k
LessWrong
I Stand by the Sequences Edit, May 21, 2012: Read this comment by Yvain. > Forming your own opinion is no more necessary than building your own furniture. - Peter de Blanc There's been a lot of talk here lately about how we need better contrarians. I don't agree. I think the Sequences got everything right and I agr...
43199af5-0cd1-400e-bee5-0fdddabc9aa5
trentmkelly/LessWrong-43k
LessWrong
The Birth and Death of Sydney — The Bayesian Conspiracy Podcast We discuss Microsoft’s new Bing search chatbot – Sydney. It received a wave of media coverage due to erratic behavior, before being neutered by Microsoft. A dive into its behavior, human responses, and implications. Links The Birth and Death of Sydney “...
1adfcf4f-ae74-4ecc-9144-8a50fe74f856
trentmkelly/LessWrong-43k
LessWrong
How to deal with someone in a LessWrong meeting being creepy One of the lessons highlighted in the thread "Less Wrong NYC: Case Study of a Successful Rationalist Chapter" is Gender ratio matters. There have recently been a number of articles addressing one social skills issue that might be affecting this, from the pe...
36113776-1b4a-4fa8-ade7-7389dc95fcfd
trentmkelly/LessWrong-43k
LessWrong
Plausible cases for HRAD work, and locating the crux in the "realism about rationality" debate This post is my attempt to summarize and distill the major public debates about MIRI's highly reliable agent designs (HRAD) work (which includes work on decision theory), including the discussions in Realism about rationalit...
3391d309-52e2-4962-b089-7642dd34de03
trentmkelly/LessWrong-43k
LessWrong
What do trust and sharing do to reputations? Bryan Caplan asked, ‘when doesn’t reputation work well?‘ He answers, > To me, venereal disease is the most striking response.  Unlike other disease, V.D. is simple to prevent: Only have sex with people who credibly show that they aren’t infected.  How hard is that?  But a...
0502c9cc-1e40-4f3d-9fef-8bab6bf5edf8
trentmkelly/LessWrong-43k
LessWrong
A new blog with analyses of various topics (e. g. slavery and capitalism, why election models didn't predict Trump's victory)
7d0a6424-af19-4b1b-80e1-8dfaa31efa2f
trentmkelly/LessWrong-43k
LessWrong
AI Safety Debate and Its Applications All of the experimental work and some of the theoretical work has been done jointly with Anna Gajdova, David Lindner, Lukas Finnveden, and Rajashree Agrawal as part of the third AI Safety Camp. We are grateful to Ryan Carey and Geoffrey Irving for the advice regarding this project...
e9c3807e-2c9c-4371-830c-059cc03f1119
trentmkelly/LessWrong-43k
LessWrong
How to debate when authority is questioned, but really not needed? Especially in the comments of political articles or about economic issues I find myself arguing with people who question my authority about a topic rather than refute my arguments. ---- Examples may be: 1: Me: I think money printing by the Fed will...
f60b57f5-deed-4c87-8ba8-38652b93719f
trentmkelly/LessWrong-43k
LessWrong
How many parameters do self-driving-car neural nets have? Take Teslas, for example. In particular, take the neural nets driving this car here. How many parameters do they have? I just need a ballpark estimate, within one order of magnitude would be great. If not for Tesla, then for Waymo or some such. (My google-fu i...
44fe3337-16b8-466e-a2aa-4fe83da00bf5
StampyAI/alignment-research-dataset/arbital
Arbital
Formal definition of the free group The [https://arbital.com/p/-5kg](https://arbital.com/p/-5kg) may be constructed formally using van der Waerden's trick, which is not intuitive at all but leads to a definition that is very easy to work with. This page will detail van der Waerden's construction, and will prove that t...
69831ae2-9c90-4650-a528-277cc7aa1c04
trentmkelly/LessWrong-43k
LessWrong
Strange Loops - Self-Reference from Number Theory to AI Summary I see this post as essentially trying to make clear the link between self-reference as discussed in Hofstadter’s “Gödel, Escher, Bach” and self-reference as considered in some of MIRI’s work on Embedded Agency.  I start by considering instances of self-re...
7504bae8-1154-4ab4-b014-9e6b17432cc1
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
What's Left Before AGI? PaLM-E, 'GPT 4' and Multi-Modality palm e was released less than a week ago and for some people it may already be old news sure it can understand and manipulate language images and even the physical world the e at the end of palm e By the way stands for embodied but soon apparently ...
980c38ac-989f-4198-aa2d-881fbae1ce2b
trentmkelly/LessWrong-43k
LessWrong
UML IV: Linear Predictors (This is the fourth post in a series on Machine Learning based on this book. Click here for part one. If you have some background knowledge, this post might work as a stand-alone read.) The mission statement for this post is simple: we wish to study the class of linear predictors. There are ...