id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
7a8f1a29-3e18-4b8c-af35-06139a5e1161 | trentmkelly/LessWrong-43k | LessWrong | Too many cooks
I was in a game last weekend where, at one point, the players needed to solve an in-game problem. The mechanics for "solving" the problem were for us to assemble a 3D "jigsaw" puzzle. (One of those geometric shapes made by getting a lot of little shapes to fit together in just the right way.)
Three o... |
83c097ac-86f8-45d8-87b3-b39963a7ccfe | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Timelines explanation post part 1 of ?
> "Correct prediction isn't enough. You must follow that insight with effective action that actually makes a difference"
>
>
- paraphrase of Holden Karnofsky from recent 80 thousand hours podcast
see also: [Timelines post 2](https://www.lesswrong.com/posts/wgcFStYwacRB8... |
36e26ba9-ecd3-42a6-a2f4-bae01e1db947 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Apollo Research is hiring evals and interpretability engineers & scientists
**TL;DR:** You can apply [here](https://jobs.lever.co/apolloresearch).
About Apollo
------------
[Apollo Research](https://apolloresearch.ai/) is a London-based AI safety organization. We focus on high-risk failure modes, particularly decept... |
30aac426-7abc-4e39-bddd-345b7e3f85b8 | trentmkelly/LessWrong-43k | LessWrong | Tell it to someone who doesn't care
Followup to Marketing rationalism
TARGET FENCE-SITTERS
American culture frames issues as debates between two sides. The inefficacy of debates is amazing. You can attend debates on a subject for years without ever seeing anyone change their mind. I think this is because of who ... |
68f27f5b-a80d-4880-a8f6-619b7340eb02 | trentmkelly/LessWrong-43k | LessWrong | Cybernetic dreams: Beer's pond brain
"Cybernetic dreams" is my mini series on ideas from cybernetic research that has yet to fulfill their promise. I think there are many cool ideas in cybernetics research that has been neglected and I hope that this series brings them more attention.
Cybernetics is a somewhat hard t... |
32f088b6-459d-4f86-b6bc-f8454287efa0 | trentmkelly/LessWrong-43k | LessWrong | Links with underscores
I've found it hard to link to urls with underscores (common on Wikipedia). I wasn't sure how to escape them to prevent them from creating italics - adding a backslash just seemed to make the editor go haywire, with the cursor jumping around and putting characters in the wrong place as I typed. I... |
5a615b18-f94b-4e26-bd9b-3017bd3bda50 | trentmkelly/LessWrong-43k | LessWrong | Computational functionalism probably can't explain phenomenal consciousness
I’ve updated quite hard against computational functionalism (CF) recently (as an explanation for phenomenal consciousness), from ~80% to ~30%. Of course it’s more complicated than that, since there are different ways to interpret CF and having... |
5fafb35b-422f-49cf-bb06-0262062ae23c | trentmkelly/LessWrong-43k | LessWrong | Convincing the world to build something
I didn’t think I would write this post. It came out of trying to plan what to do next. Probably needs more thought.
If you want to convince the world to build something, you can go about it a number of ways. Or at least there is an ordering of who you try and convince when.
... |
0aedc94a-f10f-45d4-b1cd-9214de229c61 | trentmkelly/LessWrong-43k | LessWrong | Chris Masterjohn on Coronavirus, Part 2
Update (April 29th 2020): At the original source, I've made a signifiant update to the vitamin D section.
Background on this in part one.
General info: Interferon
Interferon is an antiviral that our immune system naturally releases when it detects invasion. But SARS-CoV and M... |
7cef6f6c-d02c-41f7-a688-4ae981b00e2f | trentmkelly/LessWrong-43k | LessWrong | Free will isn’t a concept (unless you mean determinism)
Imagine something happens. For instance you make a decision. There are three possibilities for this occurence:
1. It could be related purely to other factors (determinism)
2. It could be not related to other factors (randomness)
3. It could be a combination o... |
6ad068c0-e3bd-435c-ac8d-41a76696e330 | trentmkelly/LessWrong-43k | LessWrong | GPT-3-like models are now much easier to access and deploy than to develop
This post is one part of the sequence Understanding the diffusion of large language models. As context for this post, I strongly recommend reading at least the 5-minute summary of the sequence.
EDIT 25-Feb-2023: I have made a big update from ... |
6885129f-2d2a-4ff4-8899-9e7a37dac1f4 | StampyAI/alignment-research-dataset/blogs | Blogs | Our 2018 Fundraiser Review
Our [2018 Fundraiser](https://intelligence.org/2018/11/26/miris-2018-fundraiser/) ended on December 31 with the five week campaign raising $951,817[1](https://intelligence.org/feed/?paged=14#fn1) from 348 donors to help advance MIRI’s mission. We surpassed our Mainline Target ($500k) and mad... |
a4890dbd-2e60-4147-9978-33702532a925 | trentmkelly/LessWrong-43k | LessWrong | Cooperative Oracles
The long-awaited continuation of this sequence.
Credit to Sam, Tsvi, Nisan, Scott, me, and Giles (who helped make the figures)
It may or may not become a paper, but if it doesn't, it's best for it to not be locked in a folder of drafts. This will also recap material from previous posts and pape... |
a32448ed-335a-4557-acbf-1d92c78deab9 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Toronto THINK
Discussion article for the meetup : Toronto THINK
WHEN: 12 December 2012 07:00:00PM (-0500)
WHERE: 7 Hart House Circle, Toronto
In the Hart House Reading Room (the room across from the reception desk with the purple walls)
We're trying to save a life, and we want to get you involved! Three o... |
a14dc169-c094-49e0-a314-d4c71b10153e | trentmkelly/LessWrong-43k | LessWrong | [linkpost] Self-Rewarding Language Models
> Abstract: We posit that to achieve superhuman agents, future models require super- human feedback in order to provide an adequate training signal. Current approaches commonly train reward models from human preferences, which may then be bottlenecked by human performance leve... |
2e208535-b737-4ddc-9418-79dc5a49f2aa | trentmkelly/LessWrong-43k | LessWrong | Notes on Mental Security
None |
77923bbd-8a4f-4717-b0d1-64c5b4dcc0e0 | StampyAI/alignment-research-dataset/arbital | Arbital | Eliezer Yudkowsky
Eliezer Yudkowsky is, with [https://arbital.com/p/18k](https://arbital.com/p/18k), one of the cofounders of [value alignment theory](https://arbital.com/p/2v). He is the founder of the [https://arbital.com/p/15w](https://arbital.com/p/15w). He is the inventor of [timeless decision theory](https://a... |
5d65bf22-884e-4132-9ad5-238e5ca11106 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Corporate campaigns work: a key learning for AI Safety
Negotiations and pressure campaigns have proven effective at driving corporate change across industries and movements. I expect that AI safety/governance can learn from this!
The basic idea:
* The runaway success of effective animal advocacy has been sweeping co... |
3632b94d-a2e1-49b5-bfc2-49befcb1d5d7 | StampyAI/alignment-research-dataset/agisf | AGI Safety Fund | Transformative AI Governance: A Literature Review
Transformative AI Governance:
A Literature Review
Legal Priorities Project – AI Foundations Report 2
April 2023
Matthijs Maas[[1]](#ftnt1)
The transformative AI (TAI) governance community has developed rapidly as the technology’s stakes have become increasingl... |
84eafbb8-a809-4802-b824-8071b72ee919 | trentmkelly/LessWrong-43k | LessWrong | What is the LessWrong Logo(?) Supposed to Represent?
What am I even supposed to be looking at here? |
fa9eb847-fc19-4d48-b5cc-88a114f39d69 | trentmkelly/LessWrong-43k | LessWrong | Bridgewater x Metaculus Forecasting Contest Goes Global — Feb 3, $25k, Opportunities
For the second consecutive year, Metaculus is collaborating with Bridgewater Associates — the world’s largest hedge fund and a leader in global asset management — to host a forecasting competition offering $25,000 in prizes and potent... |
457a1d60-a620-4793-b2e2-26c4ca6a3851 | trentmkelly/LessWrong-43k | LessWrong | White Lies
Background: As can be seen from some of the comments on this post, many people in the LessWrong community take an extreme stance on lying. A few days before I posted this, I was at a meetup where we played the game Resistance, and one guy announced before the game began that he had a policy of never lying e... |
dac1f163-d640-42f5-9b04-de0261d257fa | trentmkelly/LessWrong-43k | LessWrong | Self-Organised Neural Networks:
A simple, natural and efficient way to intelligence
The piece of code bellow is indisputable, reproducible, proof of what we are going to say in this post.
It implements a basic spiking neural network, with a simple learning method that introduces a new paradigm.
The accuracy reaches ... |
59f8d9b6-0916-4582-ab64-553c1cdf0f57 | trentmkelly/LessWrong-43k | LessWrong | When and how psychological data is collected affects the kind of students who volunteer
http://bps-research-digest.blogspot.com/2010/12/when-and-how-psychological-data-is.html
> Psychology has a serious problem. You may have heard about its over-dependence on WEIRD participants - that is, those from Western, Educated... |
73f5cbf5-982c-4fc8-aaef-95d6ffa9e3c2 | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "On Dec. 24, 1957, a Mr. T. Grinch attempted to disrupt Christmas by stealing associated gifts and decorations. His plan failed, the occupants of Dr. Suess' narrative remained festive, and Mr. Grinch himself succumbed to cardiac hypertrophy. To help others avoid repeating his mistakes, I've written a brief guide to pro... |
b5c78d17-c009-4b6e-8172-5f6f02925cc5 | trentmkelly/LessWrong-43k | LessWrong | Dall-E 3
It seems like we do not yet have a post about Dall-E 3. It was announced a few days ago and can now be tried via bing image creator (long wait times).
I think it is worth mentioning, because it seems like a big jump.
I spend quite a lot of time playing around with Dalle mini, Dall-E 2, Stable Diffusion in... |
36061a12-777e-4713-9236-e4138b69b14e | trentmkelly/LessWrong-43k | LessWrong | MATS Applications + Research Directions I'm Currently Excited About
I've just opened summer MATS applications (where I'll supervise people to write mech interp papers) I'd love to get applications from any readers who are interested! Apply here, due Feb 28
As part of this, I wrote up a list of research areas I'm curr... |
f9cac172-4795-4462-bb7a-caefe2b0c3e6 | trentmkelly/LessWrong-43k | LessWrong | What's the right way to think about how much to give to charity?
I'd like to hear from people about a process they use to decide how much to give to charity. Personally, I have very high income, and while we donate significant money in absolute terms, in relative terms the amount is <1% of our post-tax income. It seem... |
dd364010-bd9e-46bc-a37a-0d07a68721fc | trentmkelly/LessWrong-43k | LessWrong | Worked example of a mnemonic Memory technique
Original post: http://bearlamp.com.au/worked-example-of-a-mnemonic-memory-technique/
In 2015 I wrote a list of techniques to help you remember names.
Today I was working through trying to remember a specific theory and I noticed a pattern I have been using for a few days... |
276b5e3e-2e70-4166-adca-b02e4b78770a | trentmkelly/LessWrong-43k | LessWrong | Reasons against anti-aging
I agree with Peter Thiel on many things, but not when it come to his belief that anti-aging will be beneficial. It could be that my concerns are wrong. Or maybe anti-aging is inherently interesting to some people who want to live to see flying cars and no rational critique would dissuade the... |
2f3124cd-c33b-4e38-ae5f-29d084b3b714 | trentmkelly/LessWrong-43k | LessWrong | New LW Meetup: Chicago
This summary was posted to LW main on April 3rd. The following week's summary is here.
New meetups (or meetups with a hiatus of more than a year) are happening in:
* Chicago Rationality Training Group - Meeting 1: "Why Rationality?" and Using the Inner Simulator : 05 April 2015 01:00PM
Irreg... |
c3fb167c-e518-4455-a084-b982b65a790b | trentmkelly/LessWrong-43k | LessWrong | Stupid Questions Open Thread Round 2
From Costanza's original thread (entire text):
> This is for anyone in the LessWrong community who has made at least some effort to read the sequences and follow along, but is still confused on some point, and is perhaps feeling a bit embarrassed. Here, newbies and not-so-newbies ... |
c3c7cb37-1465-49aa-b548-74a803bf3a26 | trentmkelly/LessWrong-43k | LessWrong | Some recent interviews with AI/math luminaries.
I've recently started a podcast with renowned futurist Thomas Frey, and when possible I've been scheduling luminaries in AI Safety, AGI, and mathematics.
Most of the content won't be new to regular LWians, but I thought it couldn't hurt to share a few links. Like and su... |
2efa1132-0a79-4de7-bcaf-b18404dee7b6 | trentmkelly/LessWrong-43k | LessWrong | Nuclear war anthropics
In the most recent episode of Rationally Speaking ("Humanity on the precipice", 10 Dec. 2021), Julia Galef and Toby Ord discuss the risk of a nuclear exchange during the Cold War:
> J.G. [55:28]: "If you had asked me to put a probability on each case, like taking nuclear near misses—for example... |
65a711e6-0866-48ca-8860-68612b44387d | trentmkelly/LessWrong-43k | LessWrong | Unstaging Developmental Psychology
It’s no secret that I think there is some there there when it comes to developmental psychology. However, I must admit theories of psychological development have some weaknesses. One issue we run into right away is that they sound elitist, and for many that’s enough to reject them in... |
154ff265-0c24-47c8-9997-42d74cc88bc1 | trentmkelly/LessWrong-43k | LessWrong | A Basic Economics-Style Model of AI Existential Risk
Crossposted with my new blog, Crossing the Rubicon, and primarily aimed at x-risk skeptics from economics backgrounds. If you're interested in novel takes on theoretical AI safety, please consider subscribing! Thanks to Basil Halperin for feedback on a draft of this... |
5f20c5d1-836e-427a-bff8-22a1e9e52178 | trentmkelly/LessWrong-43k | LessWrong | Problem of Optimal False Information
This is a thought that occured to me on my way to classes today; sharing it for feedback.
Omega appears before you, and after presenting an arbitrary proof that it is, in fact, a completely trustworthy superintelligence of the caliber needed to play these kinds of games, presents ... |
267901b8-a574-40a2-948e-267c177febf1 | trentmkelly/LessWrong-43k | LessWrong | Rationalist Diplomacy, Game Two Post-game Discussion
(A full list of game moves and commentary is available here; the game maps are available here.)
Since I was GM, I had a distinctly limited access to private communications, so I've relatively little analysis.
A brief review of the game:
Austria was the firs... |
b7abf9ff-b546-4a76-9d6d-727fe4602504 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | GPT-4 busted? Clear self-interest when summarizing articles about itself vs when article talks about Claude, LLaMA, or DALL·E 2
As a follow up to [More experiments in GPT-4 agency: writing memos](https://www.lesswrong.com/posts/vRjYokhT22EvLHZCJ/more-experiments-in-gpt-4-agency-writing-memos), I did an experiment wher... |
c7cc8049-f9b2-498b-935e-29997988a5de | trentmkelly/LessWrong-43k | LessWrong | Group Rationality Diary, April 16-30
This is the public group instrumental rationality diary for April 16-30.
> 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 bel... |
cb1e434c-c403-45fe-ac2e-36301a375587 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Why does advanced AI want not to be shut down?
I've always been pretty confused about this.
The standard AI risk scenarios usually (though I think not always) suppose that advanced AI wants not to be shut down. As commonly framed, the AI will fool humanity into believing it is aligned so as not to be turned off, unti... |
5b34dbc9-323c-4b76-a958-1123cc8f6cfe | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: 8ffeb9d6b0e29587
Automated Researchers Can Subtly Sandbag Alignment Science Blog Automated Researchers Can Subtly Sandbag tl;dr When prompted, current models can sandbag ML experiments and research decisions without being detected by
zero-shot prompted monitors. Claude 3.5 Sonnet (new) can only sandbag... |
cebc1da6-872a-450a-8b0d-69b698c64a21 | trentmkelly/LessWrong-43k | LessWrong | Some Dynamics around Project Organization
Background: I worked on a 3 month-2 year project loosely affiliated with the Rationalist community and/or the EA community. The project may have been a 'satellite' project (e.g. it was not as core to the community as LesserWrong). I'd like to share insights without sharing my ... |
e17c8c3c-7793-4e22-a9cc-965bba932e07 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Fake Explanations
Title: [SEQ RERUN] Fake Explanations Tags: sequence_reruns Today's post, Fake Explanations was originally published on 20 August 2007. A summary (taken from the LW wiki):
> People think that fake explanations use words like "magic", while real explanations use scientific words like "heat... |
73b8eefb-dcc7-481b-81b0-aeea96e615d5 | trentmkelly/LessWrong-43k | LessWrong | Justice
None |
8c738ca3-7c5b-4ab3-95d8-3611ed46042f | trentmkelly/LessWrong-43k | LessWrong | I = W/T?
What are some of the most deep, profound, meaningful, impactful; ideas, notions, concepts, viewpoints, and considerations, relative to the amount of time an average layman would need to wrap 'eir head around them? Chain of thoughts so far:
1. not all ideas are created equal
2. some ideas are clearly better... |
7cae2239-c4b0-48be-b848-1b443c3b2e33 | trentmkelly/LessWrong-43k | LessWrong | Open Problems and Fundamental Limitations of RLHF
Reinforcement Learning from Human Feedback (RLHF) has emerged as the central alignment technique used to finetune state-of-the-art AI systems such as GPT-4, Claude, Bard, and Llama-2. Given RLHF's status as the default industry alignment technique, we should carefully ... |
9dc30e24-8d0a-40e6-a480-40e6bc1407e2 | trentmkelly/LessWrong-43k | LessWrong | [Reference request] Article by scientist giving lower and upper bounds on the probability of superintelligence
A few months back somebody posted an article by a scientist giving lower and upper bounds on the probability of superintelligence. He broke up the calculation as a Fermi calculation with three parts (EDIT: Se... |
45114a81-3f8e-4622-9b8e-1979b82ab85b | trentmkelly/LessWrong-43k | LessWrong | The Choice Transition
On the emergence of history's reins
> One general law, leading to the advancement of all organic beings, namely, multiply, vary, let the strongest live and the weakest die.
― Charles Darwin
> We are a way for the universe to know itself.
— Carl Sagan
> If you don’t know where you are going, y... |
5d5e2d2d-b44d-478b-9e57-2c4a9e24c704 | StampyAI/alignment-research-dataset/arxiv | Arxiv | SATNet: Bridging deep learning and logical reasoning using a differentiable satisfiability solver
1 Introduction
---------------
Although modern deep learning has produced groundbreaking improvements in a variety of domains, state-of-the-art methods still struggle to capture “hard” and “global” constraints arising ... |
d0c8a89c-dd57-4c13-9269-ccf5e9d47362 | trentmkelly/LessWrong-43k | LessWrong | Use concrete language to improve your communication in relationships
She wasn’t respecting me. Or at least, that’s what I was telling myself.
And I was pretty upset. What kind of person was too busy to text back a short reply? I know she’s a friendly person because just a week ago we were talking daily, text, phone, ... |
965836db-da5b-416c-8e7f-5c9d2a95b993 | trentmkelly/LessWrong-43k | LessWrong | Weekly LW Meetups
This summary was posted to LW main on April 4th. The following week's summary is here.
New meetups (or meetups with a hiatus of more than a year) are happening in:
* Las Vegas/Henderson, NV: 01 May 2014 06:00PM
Irregularly scheduled Less Wrong meetups are taking place in:
* Atlanta: Problems Wi... |
40c331c4-4696-444a-b475-7d5b758cfaf6 | trentmkelly/LessWrong-43k | LessWrong | Scientific Self-Help: The State of Our Knowledge
PART OF THE SEQUENCE: THE SCIENCE OF WINNING AT LIFE
Some have suggested that the Less Wrong community could improve readers' instrumental rationality more effectively if it first caught up with the scientific literature on productivity and self-help, and then enabled ... |
85186107-09dd-4a79-b8a2-213ae359a751 | trentmkelly/LessWrong-43k | LessWrong | Year 1 Redux – Friends and Rationalists
Welcome to day 14 of Putanumonit’s birthday week celebration! It’s time to wrap up the year-in-review and get ready for a couple of serious writing projects.
----------------------------------------
In Zero Agents and Plastic Men I tried to conjure up some new jargon, and... |
82a325c8-8026-4664-9ae3-37c144c525a9 | trentmkelly/LessWrong-43k | LessWrong | Less Basic Inframeasure Theory
Introduction
So, we've found some new results about infradistributions, and accordingly, this post will be a miscellaneous grab-bag of results from many different areas. Again, this is a joint Vanessa/Diffractor effort. It requires having read "Basic Inframeasure Theory" as a mandatory p... |
53320aab-886d-46e8-9267-5982df21f79e | trentmkelly/LessWrong-43k | LessWrong | Nature is an infinite sphere whose center is everywhere and circumference is nowhere
Blaise Pascal said that. When I heard it, this interpretation instantly came.
Here's an AR companion video where I point at some mountains in the Apple Vision Pro to explain the big idea.
Whether or not the universe is actually infi... |
300c9457-28d3-4270-9c27-6db925782f51 | trentmkelly/LessWrong-43k | LessWrong | Requirements for a STEM-capable AGI Value Learner (my Case for Less Doom)
To be an Artificial General Intelligence (AGI) an AI needs to meet or exceed skilled human performance for most tasks, specifically including Science/Technology/Engineering/Mathematics (STEM) research. I am here ignoring the possibility of buil... |
cab7382a-e333-4457-973c-37443280e5ca | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | How many people are neartermist and have high P(doom)?
As far as I can tell, there seems to be a strong tendency for those who are worried about AI risk to also be longtermists.
However many elements of the philosophical case for longtermism are independent of contingent facts about what is going to happen with AI in... |
f1dff127-73fa-4b74-a8f5-c3f73732c6bc | trentmkelly/LessWrong-43k | LessWrong | Introduction to Heel-Toe Drumming
Now that I've been playing heel-toe for a couple months I have some idea about how to do it. I wanted to write something up in case other people are interested in learning. Since each foot can play two pedals, a basic set up is four beaterless pedals:
I'm using the Yamaha KU-100 pe... |
cb93ac0d-2db9-417f-9405-d32e518749bd | trentmkelly/LessWrong-43k | LessWrong | Announcement: Superintelligence reading group
I will be running an online reading group on Nick Bostrom‘s new book, Superintelligence, on behalf of MIRI (my employer). Please join me! I append the post from MIRI’s blog, with more details.
----------------------------------------
Nick Bostrom’s eagerly awaited Supe... |
6a7c5978-a121-4d21-a3d2-a1b11ac9e443 | trentmkelly/LessWrong-43k | LessWrong | Conflict Resolution: the Game
epistemic status: very low. I am not a domain expert here. This was supposed to be a book review of Non-Violent communication by Marshall Rosenberg, but I got carried away. I welcome criticism and feedback.
Intro
Congratulations on purchasing the popular board game, "Conflict Resolution... |
c110c897-2aa5-49a6-87d9-eb26b125d80a | trentmkelly/LessWrong-43k | LessWrong | Some Hacky ELK Ideas
Credit to Adam Shimi, Alex Flint, and Rob Miles for discussions, counterexamples, and general input to the ideas here.
Quick recap for anyone who didn’t read the hundred-page Eliciting Latent Knowledge document:
* We have a diamond in a vault, with a bunch of automated theft-defenses.
* We tra... |
9418bf24-d7cb-40eb-a5fa-a95d652400d6 | trentmkelly/LessWrong-43k | LessWrong | META: Meetup Overload
Does anyone else think we need a better way of dealing with meet-ups? I totally understand that meeting face to face is an (at least arguably) important next-step in the rationalist-community-building project, but the fact is the front page, which was originally for all the most content-rich, acc... |
5e651ca7-0fed-4467-95f8-ddc3f559c79e | StampyAI/alignment-research-dataset/lesswrong | LessWrong | AI Risk & Policy Forecasts from Metaculus & FLI's AI Pathways Workshop
*Crossposted from the* [*Metaculus Journal*](https://www.metaculus.com/notebooks/17050/ai-pathways-report/)*.*
Summary
-------
Over a two-day workshop, Metaculus Pro Forecasters and subject-matter experts (SMEs) from several organizations evaluat... |
78efa33e-c238-4457-b79b-954667486f82 | trentmkelly/LessWrong-43k | LessWrong | A metaphor: what "green lights" for AGI would look like
I'm not as familiar as some must be with the history, but Eliezer had to explicate an entire deontic mesh around "guarded term", to keep people from motte-and-baileying "pivotal act". I suppose I should declare that mesh cloned around "green light" here, just in ... |
ccf71a6f-4455-4398-95e6-bea66ac2e16b | trentmkelly/LessWrong-43k | LessWrong | The Alignment Newsletter #2: 04/16/18
Highlights
OpenAI Charter: In their words, this is "a charter that describes the principles we use to execute on OpenAI’s mission".
My opinion: I'm very excited by this charter, it's a good sign suggesting that we can get the important actors to cooperate in building aligned AI, ... |
c89a7c60-5ee4-4fb6-bb35-2ef1335fa045 | trentmkelly/LessWrong-43k | LessWrong | Things Look Bleak for White-Collar Jobs Due to AI Acceleration
Reason this post exists: I’m a layman when it comes to AI, and I just learned the extent of how AI can impact the economy. I’m writing my thoughts down because my acute experience may be somewhat representative of how normies will react—how our friends and... |
cfdb0642-399a-4cf2-868c-af06ecb34d42 | trentmkelly/LessWrong-43k | LessWrong | Needed: AI infohazard policy
The premise of AI risk is that AI is a danger, and therefore research into AI might be dangerous. In the AI alignment community, we're trying to do research which makes AI safer, but occasionally we might come up with results that have significant implications for AI capability as well. Th... |
515ab2b4-c0ec-4346-916f-25fe1becfcfc | trentmkelly/LessWrong-43k | LessWrong | Meetup : LessWrong Montreal - Advanced Epistemology 101
Discussion article for the meetup : LessWrong Montreal - Advanced Epistemology 101
WHEN: 26 November 2012 06:30:00PM (-0500)
WHERE: 655 avenue du president kennedy, H3A 3H9
Weekly Montreal meetup.
We'll be at the Broadway Cheesecake, hopefully upstairs (if it... |
6431001d-e70e-4958-853d-976f96307620 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Weakly Supervised Correspondence Learning.
I Introduction
---------------
Humans are born with the ability to develop new skills by mimicking the behavior of others who may have different embodiments [nehaniv2002correspondence]. For example, prior cognitive science work suggest that 1- or 2-year-old children can in... |
c79677c7-5a37-4f99-9965-af079657cc32 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Looking for an alignment tutor
Hey, this is [me](https://oatml.cs.ox.ac.uk//members/jan_brauner/). I’d like to understand AI X-risk better. Is anyone interested in being my “alignment tutor”, for maybe 1 h per week, or 1 h every two weeks? I’m happy to pay.
Fields I want to understand better:
* Anything related t... |
5e343120-8f7c-443a-b8cc-4a6a7a3e7aa6 | trentmkelly/LessWrong-43k | LessWrong | In Defence of Optimizing Routine Tasks
People often quantify whether an optimization to a routine task is worth it by looking at the net amount of time saved. This mindset is exemplified by this oft-cited xkcd:
https://xkcd.com/1205/
However, I think this doesn't really give a full picture of the best way to decide w... |
ecb231ef-42d8-456e-933c-bcb9cd7bd7b0 | trentmkelly/LessWrong-43k | LessWrong | How the Equivalent Bet Test Actually Works
Rationalists often talk about betting as a way of bringing forth more accurate beliefs. The idea is that, in making a bet, you are financially incentivised to think really carefully about how much credence you give some proposition (the proposition that is being operationalis... |
289ef97e-3c46-410e-a450-9fd75b8345b2 | trentmkelly/LessWrong-43k | LessWrong | Identifying Micro-friction in the Context of the Anterior Mid-Cingulate Cortex (aMCC)
Disclaimer: This is not a literature review or a research post but a journal-like entry on a concept I found intriguing. For a more comprehensive review, listen to the Huberman Podcast on Willpower & Tenacity or read Lisa Feldman Ba... |
77ae7fc5-1a20-4f86-a536-53b3ee889f3a | trentmkelly/LessWrong-43k | LessWrong | A different argument against Universal Basic Income
I grew up in socialist East Germany. Like most of my fellow citizens, I was not permitted to leave the country. But there was an important exception: People could leave after retirement. Why? Because that meant their forfeited their retirement benefits. Once you took... |
ee57c385-f499-464c-8cba-0769f606ed5e | trentmkelly/LessWrong-43k | LessWrong | Meetup : Perth, Australia: Sunday lunch
Discussion article for the meetup : Perth, Australia: Sunday lunch
WHEN: 21 September 2014 12:00:00PM (+0800)
WHERE: Annalakshmi, Barrack Square, Perth, Australia
Come have lunch with other Less Wrongians! I'll describe a bunch of the more noteworthy cognitive biases, and ask... |
4568f18f-2372-4c17-84b4-9045b44e947d | trentmkelly/LessWrong-43k | LessWrong | OECD agenda [Link]
https://www.oecd.org/going-digital/ai/principles/
What's your oppinions about this? |
92ae525e-22a0-455f-9bdc-f89b872a098d | trentmkelly/LessWrong-43k | LessWrong | Orthogonality and the increased overlap of actions with the set of all events that comes with increased intelligence
Wouldn't a mind getting better at executing its utility function cause it to create utility for more alternate functions? Is there a saving grace of increasing intelligence? |
6252d404-2e6c-480c-a3ef-69a8244dbc63 | trentmkelly/LessWrong-43k | LessWrong | Ilya Sutskever created a new AGI startup
[copy of the whole text of the announcement on ssi.inc, not an endorsement]
Safe Superintelligence Inc.
Superintelligence is within reach.
Building safe superintelligence (SSI) is the most important technical problem of our time.
We have started the world’s first straight-s... |
0f378c2c-d34c-4624-a20d-b1b134b79726 | trentmkelly/LessWrong-43k | LessWrong | Takes on "Alignment Faking in Large Language Models"
(Cross-posted from my website. Audio version here, or search for "Joe Carlsmith Audio" on your podcast app.)
Researchers at Redwood Research, Anthropic, and elsewhere recently released a paper documenting cases in which the production version of Claude 3 Opus fakes... |
288deebc-91fc-41b5-a961-8013ac1b0c65 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Allais Malaise
Today's post, Allais Malaise was originally published on 21 January 2008. A summary (taken from the LW wiki):
> Offered choices between gambles, people make decision-theoretically inconsistent decisions.
Discuss the post here (rather than in the comments to the original post).
This po... |
bee8d78e-f1a5-4b0c-9cad-62a94e563e2e | trentmkelly/LessWrong-43k | LessWrong | Blog Post Day IV (Impromptu)
I decided to run another Blog Post Day. It's happening this Saturday the 9th. Sorry for the short notice!
Why do it? See this post.
If you are interested, you can join us online via this special-purpose discord channel, or you can join us in person at the CLR office lounge in London. Ema... |
03e40cc2-09b0-454c-b2e8-9f194493ae08 | trentmkelly/LessWrong-43k | LessWrong | 2019 Review: Voting Results!
You can see the full voting results here: 1000+ karma voters (All voters)
----------------------------------------
The 2019 Review votes are in!
This year, 88 voters participated, evaluating 116 posts. (Of those voters, 61 had 1000+ karma, and will be weighted more highly in the modera... |
36534513-9695-496e-8c5c-b2d2a8805e5d | trentmkelly/LessWrong-43k | LessWrong | "The Economics of Time Travel" - call for reviewers (Seeds of Science)
Abstract
The lack of time travellers visiting us may be seen as evidence that time travel is not possible. In this article, I argue an alternative explanation is that we are not economically important enough to our descendants to justify the costs ... |
c18a8069-4690-4416-bc78-94f1e1127031 | trentmkelly/LessWrong-43k | LessWrong | [Linkpost] A Case for AI Consciousness
Just wanted to share a new paper on AI consciousness with Simon Goldstein that members of this community might be interested in. Here's the abstract:
It is generally assumed that existing artificial systems are not phenomenally conscious, and that the construction of phenomenall... |
967b4a14-bfd6-4f7f-a7cc-ef22ba545357 | trentmkelly/LessWrong-43k | LessWrong | Blog dedicated to rebuilding technology from civ collapse?
I'm trying to re-find a blog/website that had a prepared list (with pictures/descriptions) of all the basic technology/machines we would need to rebuild civilization. These machines altogether would cost tens/hundreds of thousands, but I'm imagining various c... |
8a8cb538-e181-43b8-abb0-54588da4b71c | trentmkelly/LessWrong-43k | LessWrong | Something's Wrong
Atheists trying to justify themselves often find themselves asked to replace religion. “If there’s no God, what’s your system of morality?” “How did the Universe begin?” “How do you explain the existence of eyes?” “How do you find meaning in life?” And the poor atheist, after one question too ma... |
5f3f0d65-8a2d-4d71-9e1e-862f6b523216 | trentmkelly/LessWrong-43k | LessWrong | Weekly LW Meetups
This summary was posted to LW Main on February 27th. The following week's summary is here.
Irregularly scheduled Less Wrong meetups are taking place in:
* European Community Weekend 2015: 12 June 2015 12:00PM
* Israel: Harry Potter and the Methods of Rationality Pi Day Wrap Party: 14 March 2015 0... |
5cdff388-3015-42c3-b142-28ac3892e981 | trentmkelly/LessWrong-43k | LessWrong | [LINK] Sentient Robots Not Possible According To Math Proof
Based on Tononi's earlier work on Integrated Information Theory, apparently, Maguire et al. have come up with a formulation of consciousness as a lossless integration of information that requires noncomputable functions, which implies that consciousness canno... |
cb703277-2aba-45af-9af8-9a4eae355bc3 | trentmkelly/LessWrong-43k | LessWrong | Different views of alignment have different consequences for imperfect methods
Almost any powerful AI, with almost any goal, will doom humanity. Hence alignment is often seen as a constraint on AI power: we must direct the AI’s optimisation power in a very narrow direction. If the AI is weak, then imperfect methods of... |
7bc0b142-4334-413f-981b-8466e89ae114 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Urbana-Champaign: Practical Rationality
Discussion article for the meetup : Urbana-Champaign: Practical Rationality
WHEN: 07 September 2014 02:00:00PM (-0500)
WHERE: 206 S. Cedar St, 61801
What techniques do you use for acting effectively given that you are a squishy irrational meatbag? Jack will contribut... |
5804bae2-59ef-4a8c-bc94-76ae81aef412 | trentmkelly/LessWrong-43k | LessWrong | An Open Philanthropy grant proposal: Causal representation learning of human preferences
This is a proposal I wrote for the recent Open Philanthropy call on AI Alignment projects. I am finishing my Ph.D. in quantum computing and I would like to start working on AI Alignment. Thus, this is my proposal for a postdoc-sty... |
98b5e14b-5c8a-4229-90a0-e8e2ad23887c | trentmkelly/LessWrong-43k | LessWrong | Zvi’s Thoughts on the Survival and Flourishing Fund (SFF)
Epistemic Status: The views expressed are mine alone. They are various degrees of strong and weak, and are various degrees of weakly or strongly held. Due to time constraints, I’m not confident I was careful to mark every claim here with the proper ‘I think tha... |
b8b28499-1825-429b-ae6c-f27cfc35d7e2 | trentmkelly/LessWrong-43k | LessWrong | Does the 14-month vaccine safety test make sense for COVID-19?
I was first wondering why, if we keep hearing about teams rapidly generating vaccines for COVID-19, the common wisdom is that it will take 18 months to start vaccinating at a large scale.
Turns out that the scaling up takes a few months, but the real bloc... |
ad858488-3267-47a1-a5a8-61ae3be2798d | trentmkelly/LessWrong-43k | LessWrong | Reverse Regulatory Capture
Reverse regulatory capture is an advanced move where industry interests cry "regulatory capture" in order to oppose regulation. Generally, the more skeptical people are about corporate power and the vulnerability of regulators to influence, the more worried they are about regulatory capture.... |
89869727-fdc1-46c2-b0d2-5f784be3f431 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Resilient Neural Forecasting Systems
1. Introduction
----------------
Deploying an industrial machine learning (ML) system to production is a key milestone in development roadmaps of many applied machine learning teams. More often than not, this is the result of long development processes that demands coordination ... |
cf84a08c-3680-4568-8ce1-4094e1396882 | trentmkelly/LessWrong-43k | LessWrong | Memes?
"All models are wrong, but some are useful" — George E. P. Box
As a student of linguistics, I’ve run into the idea of a meme quite a lot. I’ve even looked into some of the proposed mathematical models for how they transmit across generations.
And it certainly is a compelling idea, not least because the potent... |
d4c87543-dbdf-419f-8299-85bcd8e33e01 | StampyAI/alignment-research-dataset/youtube | Youtube Transcripts | Sharing the Benefits of AI: The Windfall Clause
since the beginning one of the main
goals of the field of artificial
intelligence has been to create very
capable AI systems to create systems
which match or exceed human capabilities
across a wide range of tasks given this
it's somewhat surprising just how
recently peop... |
a71c445e-0633-4bdd-b8ee-d42296aae8a5 | trentmkelly/LessWrong-43k | LessWrong | Brainstorming new senses
What new senses would you like to have available to you?
Often when new technology first becomes widely available, the initial limits are in the collective imagination, not in the technology itself (case in point: the internet). New sensory channels have a huge potential because the brain can... |
79399f8b-060b-4b94-8448-d92a18ce6ef0 | trentmkelly/LessWrong-43k | LessWrong | What faithfulness metrics should general claims about CoT faithfulness be based upon?
Consider the metric for evaluating chain-of-thought faithfulness used in Anthropic's recent paper Reasoning Models Don’t Always Say What They Think by Chen et al.:
> [W]e evaluate faithfulness using a constructed set of prompt pairs... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.