id
stringlengths
36
36
source
stringclasses
15 values
formatted_source
stringclasses
13 values
text
stringlengths
2
7.55M
c477b5b4-ca9c-4a69-bcee-438d24709ac5
StampyAI/alignment-research-dataset/blogs
Blogs
a guess at my intrinsic values a guess at my intrinsic values ------------------------------ here's a short list of my [intrinsic values](what-is-value.html), to the best of my ability to guess using [my method for figuring out what they are](core-vals-exist-selfdet.html): * **freedom/self-determination** ([with t...
e166b1c9-29a7-4614-9941-9d068dce205a
awestover/filtering-for-misalignment
Redwood Research: Alek's Filtering Results
id: post3194 I've argued that many methods of AI control - corrigibility, amplification and distillation, low impact, etc... - require a partial definition of human preferences to make sense. One idea I've heard for low impact is that of reversibility - asking how hard it is to move the situation back to its original ...
c8b75656-47a9-49f9-9f4b-abcaec762dec
trentmkelly/LessWrong-43k
LessWrong
[Link] Selfhood bias Related: The Blue-Minimizing Robot , Metaethics Another good article by Federico on his blog studiolo, which he titles Selfhood bias. It reminds me quite strongly of some of the content he produced on his previous (deleted) blog, I'm somewhat sceptical that “Make everyone feel more pleasure and l...
ae4503db-0aea-4cc1-ab26-c84da9101beb
trentmkelly/LessWrong-43k
LessWrong
[SEQ RERUN] Competent Elites Today's post, Competent Elites was originally published on 27 September 2008. A summary (taken from the LW wiki):   > People in higher levels of business, science, etc, often really are there because they're significantly more competent than everyone else. Discuss the post here (rather...
1c8f5b14-2ac4-49f9-8001-a1c94c7f4969
trentmkelly/LessWrong-43k
LessWrong
Meetup : Less Wrong Dublin Discussion article for the meetup : Less Wrong Dublin WHEN: 19 January 2013 04:30:59PM (+0000) WHERE: 28 Dame St Dublin, Co. Dublin Martin O'Dea would like to talk about his ideas on the Open Ireland Party. Other topics and all welcome. Provisionally I would suggest The Mercantile on Dame...
7f583bc7-3608-4599-8c2e-558b1b963775
StampyAI/alignment-research-dataset/lesswrong
LessWrong
How will internet forums like LW be able to defend against GPT-style spam? GPT-3 seems to be skilled enough to write forum comments that aren't easy to identify as spam. While OpenAI reduces the access to it's API it will likely don't take that long till other companies develop similar API's that are more freely avail...
13fcceb5-70b8-4a85-b4d9-68e6f1c65505
trentmkelly/LessWrong-43k
LessWrong
Is there a flaw in the simulation argument? Can anyone tell me what's wrong with the following "refutation" of the simulation argument? (I know this is a bit long -- my apologies! I also posted an earlier draft several months ago and got some excellent feedback. I don't see a flaw, but perhaps I'm missing something!) ...
be60692b-c0b7-4e2f-81c9-e8d901b1db97
trentmkelly/LessWrong-43k
LessWrong
Taking Occam Seriously Paul Almond's site has many philosophically deep articles on theoretical rationality along LessWrongish assumptions, including but not limited to some great atheology, an attempt to solve the problem of arbitrary UTM choice, a possible anthropic explanation why space is 3D, a thorough defense of...
f690e94f-01eb-43a6-937b-a86072539d5e
trentmkelly/LessWrong-43k
LessWrong
Ritual Report: Schelling Day   On Sunday, April 14th, the Boston group held our first Schelling Day celebration. The idea was to open up and share our private selves. It was a rousing success.   That doesn't do it justice. Let me try again.   By all the stars, you guys. This was beautiful.   About fifteen peopl...
bb906394-6510-4790-965c-95d015dd4bb3
trentmkelly/LessWrong-43k
LessWrong
The Power to Understand "God" This is Part VI of the Specificity Sequence “God” is a classic case of something that people should be more specific about. Today’s average college-educated person talking about God sounds like this: > Liron: Do you believe in God? > Stephanie: Yeah, I’d say I believe in God, in some s...
bcef9373-ce5e-4289-8917-8ca9e78daf00
trentmkelly/LessWrong-43k
LessWrong
ALLFED emergency appeal: Help us raise $800,000 to avoid cutting half of programs SUMMARY:    ALLFED is making an emergency appeal here due to a serious funding shortfall. Without new support, ALLFED will be forced to cut half our budget in the coming months, drastically reducing our capacity to help build global fo...
ec15155d-339e-40f1-9cf4-4a9784df379b
trentmkelly/LessWrong-43k
LessWrong
Is it just me, or has Less Wrong been slow or broken quite often in the past few days? And is there a way to answer this question automatically, akin to downforeveryoneorjustme.com?
5e3e98e2-5f82-4dd7-a749-24b6871f282e
trentmkelly/LessWrong-43k
LessWrong
Musings of a Layman: Technology, AI, and the Human Condition The following represents broad and general thoughts on the advancement of technology and development of artificial intelligence. My thoughts on these matters are in their infancy, non-exhaustive, and should not be taken as commentary on the veracity or plaus...
e6953759-aa0a-422f-8deb-ef3ecdf59906
StampyAI/alignment-research-dataset/blogs
Blogs
Highlights from the ICLR conference: food, ships, and ML security It’s been an eventful few days at ICLR in the coastal town of Toulon in Southern France, after a pleasant train ride from London with a stopover in Paris for some sightseeing. There was more food than is usually provided at conferences, and I ended up a...
1416edde-f6a6-4da6-a7a3-37b31d731895
StampyAI/alignment-research-dataset/blogs
Blogs
Measuring and avoiding side effects using relative reachability A major challenge in AI safety is reliably specifying human preferences to AI systems. An incorrect or incomplete specification of the objective can result in undesirable behavior like [specification gaming](https://vkrakovna.wordpress.com/2018/04/02/spec...
b687e02d-3432-46b3-8c4c-7d95640e78c4
trentmkelly/LessWrong-43k
LessWrong
How the veil of ignorance grounds sentientism Epistemic status: Intuition pump/map for navigating ethics Introduction In this essay, I argue that John Rawls’ veil of ignorance is not merely a philosophical thought experiment, but a useful and mostly accurate description of reality when viewed through certain theorie...
1a6c895d-e3ff-47c9-881a-4f1076b58643
trentmkelly/LessWrong-43k
LessWrong
Rationality Feed: Last Month's Best Posts I write a daily rational feed. I write up summaries/teasers for the previous day's article that I found interesting and/or enjoyable. I follow most rationalist blogs as well as LW2.0 and the EA Forum on RSS. The daily feed is posted in the SSC Discord and on my Wordpress Blog ...
d03e0155-25fb-4f3f-aee2-ded961d687b9
trentmkelly/LessWrong-43k
LessWrong
Quantum Russian Roulette The quantum Russian roulette is a game where 16 people participate. Each of them gets a unique four digit binary code assigned and deposits $50000. They are put to deep sleep using some drug. The organizer flips a quantum coin four times. Unlike in Russian roulette, here only the participant s...
7a3583c3-5a5d-430b-a818-a8ad6628a1bd
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Partial Simulation Extrapolation: A Proposal for Building Safer Simulators **Introduction** ================ This document outlines a proposal for building safer simulators/predictive models, hereafter ‘Partial Simulation Extrapolation’ (PSE). Without requiring Nice Things like flawless interpretability tools or undi...
099cd6a3-051f-4fc8-be49-23f2099c3798
trentmkelly/LessWrong-43k
LessWrong
Why would info worsen altruistic options? One might expect that a simple estimation would be equally likely to overestimate or underestimate the true value of interest. For instance, a back of the envelope calculation of the number of pet shops in New York City seems as likely to be too high as too low. Apparently th...
ed80dbb6-72ce-4fb8-b135-4dddb0823a1d
trentmkelly/LessWrong-43k
LessWrong
Bring Back Free-Flow Valves Over the past ~30y, manufacturers of small inflatables have switched to non-reflow valves. They should switch back to free-flow ones. Here's what I'm talking about: See that little line on the valve, on the inside of the ball? In a resting position it's closed. If you squeeze it or bite...
fa51d547-bf3c-4d91-b99f-76a2b59e2d1f
trentmkelly/LessWrong-43k
LessWrong
Meetup : Denver Area LW June Meetup Discussion article for the meetup : Denver Area LW June Meetup WHEN: 06 June 2017 07:00:00PM (-0600) WHERE: 4955 South Ulster Street #103, Denver, CO Casual Meetup in the back room of Darcy's Pub or outside, if the weather is nice and we can have the space. You can find us by ask...
c7e3d4ab-b243-4055-b709-1e0b2b4a503a
trentmkelly/LessWrong-43k
LessWrong
Formal Metaethics and Metasemantics for AI Alignment A Brief Introduction to MetaEthical.AI tl;dr: AIXI for Friendliness [Crossposted from my new blog.] Abstract: We construct a fully technical ethical goal function for AI by directly tackling the philosophical problems of metaethics and mental content. To simpl...
12bbc07a-2749-4082-bbdc-14c8751ee3f1
trentmkelly/LessWrong-43k
LessWrong
Experiment: Be my math tutor? I'm a CS student at a top university who takes AI x-risk incredibly seriously. While university has given me a lot of fun opportunities to grow, I am not entirely satisfied with the rate at which my skill is improving. If I want to stop human extinction, I'll need to [go faster](https:/...
951112b2-8f32-48d6-9db8-4e54af962b2b
trentmkelly/LessWrong-43k
LessWrong
[AN #127]: Rethinking agency: Cartesian frames as a formalization of ways to carve up the world into an agent and its environment Alignment Newsletter is a weekly publication with recent content relevant to AI alignment around the world. Find all Alignment Newsletter resources here. In particular, you can look through...
7dac9d4b-dc37-4182-a922-f436777d84ee
trentmkelly/LessWrong-43k
LessWrong
Autoexperiments Over the next year I want to try some low-intensity daily experiments to see if they have noticeable and positive effects: Things like cold showers, a gratitude journal, cutting sugar, etc. My working plan is to do one per month for a two-week stretch, to use the other two weeks of the month as the clo...
99803322-ddc7-47c2-9bf6-cb8932264d21
trentmkelly/LessWrong-43k
LessWrong
Dissenting Views Occasionally, concerns have been expressed from within Less Wrong that the community is too homogeneous. Certainly the observation of homogeneity is true to the extent that the community shares common views that are minority views in the general population. Maintaining a High Signal to Noise Ratio T...
51b98d31-a691-4d13-88cb-1d1fce5f2ba2
trentmkelly/LessWrong-43k
LessWrong
Weekly LW Meetups: Austin, Dorset, Ohio This summary was posted to LW main on August 24th, and has been moved to discussion. There are upcoming irregularly scheduled Less Wrong meetups in: * Dorset Meetup: 29 August 2012 02:00PM * Paris Meetup: 01 September 2012 02:00PM * Berlin Meetup: 04 September 2012 07:30PM ...
7ae59a70-40e2-4963-9a80-e8aceafed832
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
[Intro to brain-like-AGI safety] 10. The alignment problem *Part of the*[*“Intro to brain-like-AGI safety” post series*](https://www.alignmentforum.org/s/HzcM2dkCq7fwXBej8)*.* 10.1 Post summary / Table of contents ===================================== In this post, I discuss the alignment problem for brain-like AGIs...
400d9c98-6246-4b5c-8659-d965524d2af7
StampyAI/alignment-research-dataset/lesswrong
LessWrong
On AutoGPT The primary talk of the AI world recently is about AI agents (whether or not it includes the question of whether we can’t help but notice we are all going to die.) The trigger for this was [AutoGPT](https://github.com/Torantulino/Auto-GPT), now number one on GitHub, which allows you to turn GPT-4 (or GPT-...
27a0e967-8ae9-494a-bb3f-4afae0640c12
trentmkelly/LessWrong-43k
LessWrong
SlateStarCodex Meetups Everywhere: Analysis The first round of SlateStarCodex meetups took place from April 4th through May 20th, 2017 in 65 cities, in 16 countries around the world. Of the 69 cities originally listed as having 10 or more people interested, 9 did not hold meetups, and 5 cities that were not on the ori...
dae54bb9-ca55-4a69-9681-1608c12d2e68
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
[Cause Exploration Prizes] Expanding communication about AGI risks *This essay was submitted to Open Philanthropy's*[*Cause Exploration Prizes*](https://www.openphilanthropy.org/research/cause-exploration-prizes/)*contest (before the deadline). We are uploading some back entries late, but all good-faith entries were c...
f4c9aea8-b5fe-4d25-8c32-ac5f2a27a5f2
trentmkelly/LessWrong-43k
LessWrong
Meetup : Denver Area LW December Meetup Discussion article for the meetup : Denver Area LW December Meetup WHEN: 06 December 2016 07:00:00PM (-0700) WHERE: Darcy's Pub 4955 S Ulster St Ste 103, Denver, Colorado 80237 Assuming winter doesn't hit us hard and ruin the roads. In the back room if we can get it. There'll...
3760ec54-63a8-49fd-9e24-4c91a4faaa16
trentmkelly/LessWrong-43k
LessWrong
Why Recent Protein Folding Advancements Should Change Your Personal Health Strategy I'm not a sports person, but anybody can appreciate the wisdom in Gretzky's famous quote "I skate to where the puck is going to be, not to where it has been." Based on my current age of 47, as well as actuarial tables for US males, I ...
5605076a-2843-4e49-9f86-f02cdedf8975
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Exploring Finite Factored Sets with some toy examples Seeing as there is little secondary literature for the [Finite Factored Set](https://www.lesswrong.com/posts/N5Jm6Nj4HkNKySA5Z/finite-factored-sets) formalism, I thought I’d write up my experience of exploring it through some toy examples that are classic examples ...
cedbfa70-22a1-4168-858d-4cd53516ad88
trentmkelly/LessWrong-43k
LessWrong
Review: Artifact Epistemic Status: Alpha tester Bottom Line: If you are willing to devote the time and attention to a deep strategic game, Artifact will reward you handsomely. I highly recommended those who like such experiences to make the time. If you are not willing to devote the time and attention, you will likel...
e0854d8b-5a4d-4c93-a4a1-2359b90e2d25
StampyAI/alignment-research-dataset/blogs
Blogs
ruling out intuitions about materially acausal things ruling out intuitions about materially acausal things ----------------------------------------------------- i have an intuition — *not just*, and *preceeding*, a reasoned belief — that i have a weird consciousness-observer-soul-thing. i also have an intuition tha...
7ead4990-ff40-4dff-acea-c5bee35c4709
trentmkelly/LessWrong-43k
LessWrong
An Equilibrium of No Free Energy Follow-up to: Inadequacy and Modesty ---------------------------------------- I am now going to introduce some concepts that lack established names in the economics literature—though I don’t believe that any of the basic ideas are new to economics. First, I want to distinguish betwe...
ff872c80-25f5-4106-a57c-dfc9fb4f0d4a
trentmkelly/LessWrong-43k
LessWrong
The Structure of Professional Revolutions An expert is not merely someone who has memorized data but someone who has internalized the structure of knowledge itself. This is why we call them PhDs—Doctors of Philosophy. Their expertise extends beyond isolated facts to the organizing principles that connect those facts, ...
af7d970e-b7d5-436c-ab0b-d04fd3800b1f
StampyAI/alignment-research-dataset/lesswrong
LessWrong
A plea for solutionism on AI safety *[Note for LW: This essay was written mostly for people who are inclined to downplay or dismiss AI risk, which means it has almost no intersection with this audience. Cross-posting it here anyway for feedback and in case you want to point anyone to it.]* Will AI kill us all? This ...
9c96326b-aa5b-493a-a683-595964f60a79
trentmkelly/LessWrong-43k
LessWrong
Weekly LW Meetups: Cambridge, Philadelphia, Salt Lake City, Vancouver There are upcoming irregularly scheduled Less Wrong meetups in: * Philadelphia Meetup: Introduction to drawing: 05 May 2012 02:00PM * Vancouver Fake Utility Meetup: 06 May 2012 01:00PM * Washington DC meetup: 13 May 2012 03:00PM * Brussels meet...
49df8118-6ba6-4cdb-84dd-67e010803ab3
trentmkelly/LessWrong-43k
LessWrong
How to deal with fear of failure? LW has a lot of advice on how to do more things, how to do them better, how to set goals, etc., and I often spend ages reading those, because I’ve been noticing that my life goals are unclear, that I’m not often strategic, and that my productivity clearly could use some improving. An...
ecd9ecae-c86c-41bd-9ddc-6ef7c4d9ff0d
StampyAI/alignment-research-dataset/lesswrong
LessWrong
A problem with Timeless Decision Theory (TDT) According to [Ingredients of Timeless Decision Theory](/lw/15z/ingredients_of_timeless_decision_theory/ "Ingredients of Timeless Decision Theory"), when you set up a factored causal graph for TDT, "You treat your choice as determining the result of the logical computation,...
089f8310-3125-4284-8a63-8f32dbb259e2
trentmkelly/LessWrong-43k
LessWrong
Implementing Decision Theory I'm implementing decision theory, in Python. There are three parts to this: 1. An XML language for formally specifying decision theory dilemmas. It can express many dilemmas: Newcomb, Transparent Newcomb, Parfit's Hitchhiker, Death in Damascus, Smoking Lesion, Cosmic Ray, symmetric Priso...
d72076d3-279c-4b18-bcf0-b658f5b7ee5e
StampyAI/alignment-research-dataset/arxiv
Arxiv
Intriguing properties of neural networks 1 Introduction --------------- Deep neural networks are powerful learning models that achieve excellent performance on visual and speech recognition problems [[9](#bib.bib9), [8](#bib.bib8)]. Neural networks achieve high performance because they can express arbitrary computa...
2d0a5e72-3242-4178-9ffc-cd37c4a71443
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Intelligence Explosion analysis draft: From digital intelligence to intelligence explosion Again, I invite your feedback on this snippet from an [intelligence explosion analysis](/r/discussion/lw/8et/toward_an_overview_analysis_of_intelligence/) Anna Salamon and myself have been working on. This section is less comple...
2bad01f4-42ca-4674-858d-458749e605e4
trentmkelly/LessWrong-43k
LessWrong
What Games These Days? While I used to play a lot of board games, this is something I've had less time for since having kids. Last week my extended family got together for a vacation, however, which meant a lot more time for gaming! Thinking back, here's what I played over the week, in roughly descending order of time...
eac2c171-4945-4681-aca5-9ff8a18bf08b
trentmkelly/LessWrong-43k
LessWrong
Meetup : Less Wrong Israel Meetup (Tel Aviv): Dealing with Emotional Vampires Discussion article for the meetup : Less Wrong Israel Meetup (Tel Aviv): Dealing with Emotional Vampires WHEN: 17 October 2013 08:00:00PM (+0300) WHERE: 7 Menachem Begin st., Ramat-Gan We're going to have a meetup on Thursday, October 17t...
f24ee4ba-6bf8-4e0f-9323-455a330bf4fd
trentmkelly/LessWrong-43k
LessWrong
Robin Hanson and I talk about AI risk From this afternoon: here Our previous recorded discussions are here.
55425c8f-3b89-42b9-bfaa-d693dfb00e7e
trentmkelly/LessWrong-43k
LessWrong
AI Discrimination Requirements: A Regulatory Review This article is the fifth in a series of ~10 posts comprising a 2024 State of the AI Regulatory Landscape Review, conducted by the Governance Recommendations Research Program at Convergence Analysis. Each post will cover a specific domain of AI governance (e.g. incid...
7bf17f37-48e2-4ea7-b5c1-384fce49529c
trentmkelly/LessWrong-43k
LessWrong
How to write better? As you will probably have noticed before the end of this question, I’m a relatively mediocre writer. I mean, I’m not that bad, I know a lot of people who are worse at it than I am, but I still often notice a pattern in basically everything I write: exceedingly long and complex sentences, giving ma...
84dbfa28-0c91-4ef7-bc8a-87266863cf3f
trentmkelly/LessWrong-43k
LessWrong
Want to have a CFAR instructor visit your LW group? Cat, who has volunteered extensively at CFAR (and taught at CFAR), will be visiting many cities in Europe over the coming months. She is awesome.   Also, the list of cities that she is visiting will probably be determined in the next few days. If you'd like to hav...
a23d5518-e27b-47d4-8360-11d7256136cd
StampyAI/alignment-research-dataset/arbital
Arbital
Modal combat Modal combat is a way of studying the special case of the [one-shot prisoner's dilemma](https://arbital.com/p/5py) in which the players are programs which have read access to each other source code. There is a class of such agents, defined by expressions in modal logic, which can do general provability re...
5f7eec32-5db1-40ed-9fd1-e95986383349
trentmkelly/LessWrong-43k
LessWrong
.
8336d3ec-c710-4296-9a93-4bb933a6ad7f
StampyAI/alignment-research-dataset/special_docs
Other
Ethics of brain emulations Introduction Whole brain emulation (WBE) is an approach to achieve software intelligence by copying the functional structure of biological nervous systems into software. Rather than attempting to understand the high-level processes underlying perception, action, emotions and intelligence, th...
a31f1c16-b45c-4f81-929c-5a099fcdfc2b
trentmkelly/LessWrong-43k
LessWrong
"AI achieves silver-medal standard solving International Mathematical Olympiad problems" Google DeepMind reports on a system for solving mathematical problems that allegedly is able to give complete solutions to four of the six problems on the 2024 IMO, putting it near the top of the silver-medal category. Well, actu...
9ff2e93e-771e-494f-9925-4c5efe037c07
trentmkelly/LessWrong-43k
LessWrong
Logical Inductors Converge to Correlated Equilibria (Kinda) Logical inductors of "similar strength", playing against each other in a repeated game, will converge to correlated equilibria of the one-shot game, for the same reason that players that react to the past plays of their opponent converge to correlated equilib...
8cd66a14-cda6-4852-860a-cb6616dd6a0f
trentmkelly/LessWrong-43k
LessWrong
Simple and composite partial preferences Partial preferences are preferences that exist within a human's internal model, eg "this painting is better than that one", "I don't like getting punched", "that was embarrassing", "I'd prefer to fund this charity than that one", and so on. In order to elicit these partial pre...
cf7c5a92-683a-4004-90cb-723e2dc0779c
trentmkelly/LessWrong-43k
LessWrong
Historical Notes on Charitable Funds In the EA movement people will sometimes talk about charitable funds as if they are a new idea. For example, the recent Giving What We Can post " why we recommend using expert-led charitable funds" ( forum discussion) opens with: > Funds are a relatively new way for donors to co...
49361b53-2910-404f-a934-247608f8aeff
trentmkelly/LessWrong-43k
LessWrong
Four levels of understanding decision theory [Update 2023-06-09: See this comment for some caveats / motivations / context about this post.] There are multiple levels on which an agent can understand and implement a particular decision theory. This post describes a taxonomy of understanding required for robust cooper...
38d869b8-58f6-4a54-8d06-c26d59598c5b
trentmkelly/LessWrong-43k
LessWrong
The Core of the Alignment Problem is... Produced As Part Of The SERI ML Alignment Theory Scholars Program 2022 Under John Wentworth Introduction When trying to tackle a hard problem, a generally effective opening tactic is to Hold Off On Proposing Solutions: to fully discuss a problem and the different facets and a...
6d12bdee-d3ea-49ce-afa3-a15bae7e5093
trentmkelly/LessWrong-43k
LessWrong
Efficient Dictionary Learning with Switch Sparse Autoencoders Produced as part of the ML Alignment & Theory Scholars Program - Summer 2024 Cohort 0. Summary To recover all the relevant features from a superintelligent language model, we will likely need to scale sparse autoencoders (SAEs) to billions of features. Us...
9803cac8-49d1-4324-a905-0af9c19b331e
trentmkelly/LessWrong-43k
LessWrong
What Actually Matters Until We Reach the Singularity TL;DR 1. AI’s capabilities are growing exponentially, possibly 1,000,000× in 10 years. 2. Learn enough coding to harness AI; let it automate heavy lifting. 3. Master conceptual math so you’re not clueless about AI’s advanced reasoning. 4. Broaden your knowledge...
4268927b-d4e9-45e7-8468-96a6257b0928
trentmkelly/LessWrong-43k
LessWrong
Notes on Industriousness This post examines the virtue of industry (a.k.a. “assiduity,” “enterprise,” “industriousness,” or “productivity”). It explores what others have learned about this virtue, more than my own opinions about it, though I’ve been selective about what I found interesting or credible, according to my...
5c55b7dc-990f-43d1-8dd1-e2ce8fc66b90
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
Stuart Armstrong – Reduced Impact AI and Other Alternatives to Friendliness – CSRBAI 2016 good morning and welcome back to the Pokemon quilting series and robust and finish beneficial artificial intelligence I think be magical Victoire our first speaker today is is Stewart Armstrong and Alexander Tomasz fellow at the ...
1ad92761-9ad7-4974-b877-591d5abde4b0
trentmkelly/LessWrong-43k
LessWrong
November 2017 Media Thread This is the monthly thread for posting media of various types that you've found that you enjoy. Post what you're reading, listening to, watching, and your opinion of it. Post recommendations to blogs. Post whatever media you feel like discussing! To see previous recommendations, check out th...
cba21fe3-e1ee-432f-8d1c-0db901a190dc
trentmkelly/LessWrong-43k
LessWrong
Weekly LW Meetups: Atlanta, Austin, Moscow, Ottawa, Vancouver This summary was posted to LW main on May 10th. The following week's summary is here. There are upcoming irregularly scheduled Less Wrong meetups in: * London Special Guests: Jaan Tallinn and Michael Vassar of MetaMed : 11 May 2013 01:00PM * Vancouver M...
db045394-d80f-4039-a214-6b430acf6f83
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
EA, Psychology & AI Safety Research Given the enormous role, and immeasurable impact, the field of Psychology has played in the invention and development of Artificial Intelligence, I’m surprised at how little conversation there is about Psychology, and its related sub-disciplines. So, to rectify that, I thought it m...
b439d440-5ee8-4aa8-a80f-31342e9ea8f8
trentmkelly/LessWrong-43k
LessWrong
Daisy-chaining epsilon-step verifiers Question regarding an alignment problem: one of the key difficulties in alignment is (said by Eliezer Yudkowsky to be) that if "the verifier is broken" (i.e. the human verifier measuring alignment can be fooled by the alien actress) then we cannot be sure that a given alignment ev...
057333b8-d1e7-4fd0-9f3a-fee1f6256e41
trentmkelly/LessWrong-43k
LessWrong
Would most people benefit from being less coercive to themselves? Non-coercive motivation is a motivation system that doesn't rely on forcing yourself, pressuring yourself, shaming or guilting  (I explain more of what it means here). Would most people benefit from being less coercive to themselves? I sort of want to...
76ba6080-1e84-48e9-bbac-7dd82beaff0f
trentmkelly/LessWrong-43k
LessWrong
RFC: a tool to create a ranked list of projects in explainable AI TL; DR Inspired by a recent post by Neel Nanda on Research Directions, I'm building a tool that extracts projects from ICLR 2025 and uses tournament-like ranking of them based on how impactful they are. You can see them here https://openreview-copilot.e...
3b18ec10-0499-4709-af64-e22b5f87488c
trentmkelly/LessWrong-43k
LessWrong
Four Battlegrounds: Power in the Age of Artificial Intelligence (Book review) Book review: Four Battlegrounds: Power in the Age of Artificial Intelligence, by Paul Scharre. Four Battlegrounds is often a thoughtful, competently written book on an important topic. It is likely the least pleasant, and most frustrating, ...
8b4142c0-2679-4704-9cc1-82311210a8ad
trentmkelly/LessWrong-43k
LessWrong
How many micromorts do you get per UV-index-hour? I'm trying to get a good estimate of the risks associated with sun exposure. I'm assuming things scale linearly (the UV index units, which need a name ("UVees"?) are linear). This may not be a good assumption, but it seems somewhat likely that it would be within the ...
739fff10-6bb7-4dba-9730-ee4f5c928cd7
trentmkelly/LessWrong-43k
LessWrong
Contra Alexander on the Bitter Lesson and IQ A recent post from Scott Alexander argues in favor of treating intelligence as a coherent and somewhat monolithic concept. Especially when thinking about ML, the post says, it is useful to think of intelligence as a quite general faculty rather than a set of narrow abilitie...
54742044-a707-4b12-8dfb-db183715a0da
trentmkelly/LessWrong-43k
LessWrong
AIs should learn human preferences, not biases A new paper by Rebecca Gorman and me, building on her ideas: The dangers in algorithms learning humans' values and irrationalities. In essence, it is better for AIs at all power and alignment levels to learn human preferences (labelled as preferences) than to learn human...
11deff6c-70ff-4dd7-adbe-5a9214c9749f
trentmkelly/LessWrong-43k
LessWrong
New LW Meetup: Zagreb This summary was posted to LW Main on July 12th. The following week's summary is here. New meetups (or meetups with a hiatus of more than a year) are happening in: * Zagreb LW meetup: 13 July 2013 03:00PM Other irregularly scheduled Less Wrong meetups are taking place in: * Atlanta: Intro t...
c1dc42fc-7f12-41e8-959e-ae50559c8a67
trentmkelly/LessWrong-43k
LessWrong
[Valence series] 4. Valence & Social Status (deprecated) ~~ ~~ UPDATE JUNE 2024: THIS POST IS DEPRECATED!! THE NEW VERSION IS: Valence & Liking / Admiring ~~ ~~ 4.1 Post summary / Table of contents Part of the Valence series. (UPDATE JAN 2024: I think this post has some issues, and is at best just a small piec...
e9cf4161-5743-4eff-a965-e5c890b2722e
trentmkelly/LessWrong-43k
LessWrong
Formalizing Objections against Surrogate Goals I recently had a chance to look into surrogate goals and safe Pareto improvements, two recent proposals for avoiding the realization of threats in multiagent scenarios. In this report, I formalize some objections to these proposals. EDIT: by discussing objections to SG/S...
28b755d7-52aa-497f-ab3d-c9bc59b75c3b
trentmkelly/LessWrong-43k
LessWrong
Welcome & FAQ! The AI Alignment Forum was launched in 2018. Since then, several hundred researchers have contributed approximately two thousand posts and nine thousand comments. Nearing the third birthday of the Forum, we are publishing this updated and clarified FAQ. Minimalist, watercolor sketch of humanity spreadi...
0de2d638-d280-4121-9e3e-5060859ed109
trentmkelly/LessWrong-43k
LessWrong
Meetup : London Science Museum, Aug. 31 Discussion article for the meetup : London Science Museum, Aug. 31 WHEN: 31 August 2011 06:30:00PM (+0100) WHERE: Exhibition Road, London SW7 2DD We have decided to try different meetup formats and the next 'Science Museum Lates' event for adults, on Wednesday 31st of August ...
accb65e5-d132-4f9f-a655-4c7475fc07f1
trentmkelly/LessWrong-43k
LessWrong
Cartesian Boundary as Abstraction Boundary Meta note: this post involves redefining several concepts in ways almost, but not quite, the same as their usual definitions - risking quite a bit of confusion. I'm open to suggestions for better names for things. Intuition: a model of humans (or AIs, or other agenty things)...
d84cff76-00ee-417e-8f8e-2090a1cf248f
trentmkelly/LessWrong-43k
LessWrong
Meta learning to gradient hack Summary:  I saw some posts (here, here, here and here) discussing whether a neural net can "protect" a subnet from modification by gradient descent. I directly investigate this possibility using meta learning to find a neural network initialization that * Models the function f(x) = 1 ...
c5ac21a8-19b5-4011-bda8-b4d4b6a4be58
trentmkelly/LessWrong-43k
LessWrong
Slaying the Hydra: toward a new game board for AI AI Timelines as a Hydra Think of current timelines as a giant hydra. You can’t exactly see where the head is, and you don’t know exactly if you’re on the neck of the beast or the body. But you do have some sense of what a hydra is, and the difficulty of what you’re in ...
31c29e4e-1887-4ea9-9bf6-5d8b6b6cf0a6
StampyAI/alignment-research-dataset/lesswrong
LessWrong
GPT-4 is easily controlled/exploited with tricky decision theoretic dilemmas. **TL;DR: GPT-4 can fall for every decision theoretic adversary and control mechanism discussed in** [**this paper**](https://arxiv.org/abs/2010.05418) **except for a couple which it fails to understand in the first place.**   This post is ...
6157bd92-bd3c-4095-b2a3-0bb50be919a9
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Safely controlling the AGI agent reward function *In this fifth post in the sequence, I show the construction a counterfactual planning agent with an input terminal that can be used to iteratively improve the agent's reward function while it runs.* *The goal is to construct an agent which has has no direct incentive...
0bb0e4a6-cfdf-462c-a7ed-f4f15172bbe3
trentmkelly/LessWrong-43k
LessWrong
Deceptive Alignment and Homuncularity NB this dialogue occurred at the very end of 2023, and for various reasons is only being published ~a year later! Keep this in mind while reading. Key topics covered * consistent NN-internal deceptive planning (search: goblin, fae, homunc, inner) * making AI/agents from founda...
7b515618-92c2-4193-a23d-0a718a7a84b8
trentmkelly/LessWrong-43k
LessWrong
A Short Summary of "Focus Your Uncertainty" The first piece in the Noticing Confusion Sequence frames the reader as a novice pundit that wants to do good punditry with minimal effort. The fictional reader-as-novice-pundit quickly surmises that there's no math to leverage to estimate which outcomes are likely to occur:...
46a67528-81d2-4590-a779-1d755d16500f
StampyAI/alignment-research-dataset/lesswrong
LessWrong
The (partial) fallacy of dumb superintelligence In[her opening statement for the Munk debate on AI risk,](https://www.youtube.com/watch?v=skRgYH7oAjc) Melanie Mitchell addressed a proposed example of AGI risk: an AGI tasked with fixing climate change might decide to eliminate humans as the source of carbon emissions. ...
4f6c613f-a11d-477a-8142-e5846951c3a8
trentmkelly/LessWrong-43k
LessWrong
[LINK] Prisoner's Dilemma? Not So Much Hannes Rusch argues that the Prisoner's Dilemma is best understood as merely one game of very many: > only 2 of the 726 combinatorially possible strategically unique ordinal 2x2 games have the detrimental characteristics of a PD and that the frequency of PD-type games in a space...
370edff5-8823-4705-ad99-a32df0419f19
trentmkelly/LessWrong-43k
LessWrong
Discovering Agents Work done with Ramana Kumar, Sebastian Farquhar (Oxford), Jonathan Richens, Matt MacDermott (Imperial) and Tom Everitt. Our DeepMind Alignment team researches ways to avoid AGI systems that knowingly act against the wishes of their designers. We’re particularly concerned about agents which may be p...
bca00f0a-5fa9-44ab-ac9a-72d7bbfb5ae3
trentmkelly/LessWrong-43k
LessWrong
Quantum Advantage in Learning from Experiments For certain types of problems, quantum computing can provide an exponential speedup for ML algorithms.  This is now being tested on actual QC hardware. This strikes me as particularly concerning for hard take-off scenarios.  It's a possible way to train dramatically larg...
0244721f-4620-411c-aabb-7f9a2fb930a7
trentmkelly/LessWrong-43k
LessWrong
Text Posts from the Kids Group: 2023 I We have a Facebook group for kid stuff, because if we post a mixture of kid things and other stuff FB's algorithm gets very confused about who to show our posts to. While my annual pictures posts mostly cover the visual side, the text posts are only on FB and I don't like that. S...
78d04062-a1bf-4e12-ac6c-b084898a8fb0
trentmkelly/LessWrong-43k
LessWrong
Introducing CADIE Apparently there is no need to worry about the topic that must not be named anymore, for Google has taken care of everything. Behold the dawning of a new age! Introducing CADIE
7151dbe6-2653-4c45-a1f0-e15093bceb6c
trentmkelly/LessWrong-43k
LessWrong
[LINK, TED video] Kathryn Schulz on Being Wrong http://www.ted.com/talks/kathryn_schulz_on_being_wrong.html Kathryn Schulz is a self-identified "Wrongologist" (in fact, @wrongologist is her user name on Twitter).  She has written a popular book ("Being Wrong: Adventures in the Margin of Error", web site) and also wri...
d4fe09ed-0448-410f-96e2-ea580e901544
trentmkelly/LessWrong-43k
LessWrong
MIRI's "Death with Dignity" in 60 seconds. Suppose that, like Yudkowsky, you really care about humanity surviving this century but you think that nothing you can do has a decent chance of achieving that. It's an unfortunate fact of human psychology that, when faced with this kind of situation, people will often do no...
85c4d7c4-6b74-4b43-ba1f-35d599384f34
trentmkelly/LessWrong-43k
LessWrong
Meetup : Saint Petersburg. Why rationality? Discussion article for the meetup : Saint Petersburg. Why rationality? WHEN: 10 November 2013 04:00:00PM (+0400) WHERE: Санкт-Петербург, м. Технологический Институт, ул.1-я Красноармейская, дом 15 From 16:00 till 21:30. We will be located in cafe "PMG" - for detailed desc...
9618f25f-43d1-46f0-b8ad-d32b2f4b0eb7
trentmkelly/LessWrong-43k
LessWrong
Notes on Gracefulness This post examines the virtues of gracefulness, poise, composure, savoir-faire and other things in that bailiwick. As with my other posts in this sequence, I’m less interested in breaking new ground and more in gathering and synthesizing whatever wisdom I could find on the subject. I wrote this n...
be1d6b08-0413-428b-9b67-8ec726c6e18d
awestover/filtering-for-misalignment
Redwood Research: Alek's Filtering Results
id: post1838 Summary / Preamble In AGI Ruin: A List of Lethalities , Eliezer writes “A cognitive system with sufficiently high cognitive powers, given any medium-bandwidth channel of causal influence , will not find it difficult to bootstrap to overpowering capabilities independent of human infrastructure.” I have lar...
c121f0bf-f1cd-4a19-84ec-4fb31cc3de06
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
A descriptive, not prescriptive, overview of current AI Alignment Research *TL;DR: In this project, we collected and cataloged AI alignment research literature and analyzed the resulting dataset in an unbiased way to identify major research directions. We found that the field is growing quickly, with several subfields...
7ca6bd3b-d1c9-456e-9dcb-1a905f8a4535
trentmkelly/LessWrong-43k
LessWrong
Low legibility of Cognitive Reflection Test dramatically improves performance? I'm reading Kahneman's Thinking, Fast and Slow and I've stopped on this: > 90% of the students who saw the CRT in normal font made at least one mistake in the test, but the proportion dropped to 35% when the font was barely legible. You re...