id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
c2fdf4a2-f8ff-4887-a822-9ad82d494e2b | trentmkelly/LessWrong-43k | LessWrong | The Skewed and the Screwed: When Mating Meets Politics
Cross-posted, as always, from Putanumonit.
----------------------------------------
If an alien learned about our world from browsing the internet of beefs, they would be under the impression that a bitter war is raging between the sexes. All you read about are ... |
824b7066-1d29-435b-84b9-e9a2c83eca6d | trentmkelly/LessWrong-43k | LessWrong | Decision Theories: A Less Wrong Primer
Summary: If you've been wondering why people keep going on about decision theory on Less Wrong, I wrote you this post as an answer. I explain what decision theories are, show how Causal Decision Theory works and where it seems to give the wrong answers, introduce (very briefly) s... |
518379b4-6604-4ffc-bb86-6335422a811c | trentmkelly/LessWrong-43k | LessWrong | "Put It To The Test"
Alt-rockers They Might Be Giants explain/advocate empiricism in a record aimed at young children.
> Don't believe it 'cos they say it's so
>
> If it's not true you have a right to know --
>
> Put it to the test!
No, it's not quite Bayesian. The bridge ("A fact is just a fantasy, unless it ... |
f56a2d56-179b-4350-8313-a192592d2faf | trentmkelly/LessWrong-43k | LessWrong | Beauty as a signal (map)
This is my new map, in which female beauty is presented as a signal which moves from woman to man through different mediums and amplifiers. pdf
|
83fd8838-f458-42db-bc8f-ea0b50676dae | trentmkelly/LessWrong-43k | LessWrong | For alignment, we should simultaneously use multiple theories of cognition and value
This post is a follow-up to "A multi-disciplinary view on AI safety research". I elaborate on some arguments behind this view.
TL;DR: please skim section headings and bolded sentences in the text.
Computationally tractable mathemat... |
d380506b-113e-4b8a-b588-bd048c25934c | trentmkelly/LessWrong-43k | LessWrong | ChatGPT Plugins - The Beginning of the End
This is going to be a short post. OpenAI recently announced ChatGPT Plugins which will allow any developer to connect ChatGPT to external APIs and potentially real-world actions. Even if GPT-4 is not yet capable of destroying civilization, we are entering a world where LLMs a... |
3e8df742-3d05-460b-85de-b60a72f5295e | trentmkelly/LessWrong-43k | LessWrong | Introducing REBUS: A Robust Evaluation Benchmark of Understanding Symbols
This is a summary of https://arxiv.org/abs/2401.05604.
When Google announced Gemini Pro, they displayed its ability to solve rebuses—wordplay puzzles which involve creatively adding and subtracting letters from words derived from text and image... |
37783755-465c-4ebe-9cc7-117ea6ef35ea | trentmkelly/LessWrong-43k | LessWrong | Open thread, Nov. 17 - Nov. 23, 2014
If it's worth saying, but not worth its own post (even in Discussion), then it goes here.
Notes for future OT posters:
1. Please add the 'open_thread' tag.
2. Check if there is an active Open Thread before posting a new one. (Immediately before; refresh the list-of-threads page ... |
51261658-7ac3-42e3-95b3-dc3efed1f9af | trentmkelly/LessWrong-43k | LessWrong | PROPOSAL: LessWrong for Teenagers
What about an online group for high schoolers devoted to refining the art of human rationality?
Discovering LessWrong had a profound effect on me, shedding light on the way I study thought processes and helping me with a more rational approach. As a teenager in high school, I wish I ... |
915e000c-ea9f-4ceb-99aa-0005c9902cc8 | trentmkelly/LessWrong-43k | LessWrong | An introduction to worst-case AI safety
None |
2977ca20-c6e2-4abb-856b-3874532b3c7c | trentmkelly/LessWrong-43k | LessWrong | The Problem with Reasoners by Aidan McLaughin
Some critique on reasoning models like o1 (by OpenAI) and r1 (by Deepseek).
> OpenAI admits that they trained o1 on domains with easy verification but hope reasoners generalize to all domains. Whether or not they generalize beyond their RL training is a trillion-dollar qu... |
145730f7-d7d8-41ea-8986-ea1eb52592da | StampyAI/alignment-research-dataset/lesswrong | LessWrong | An Impossibility Proof Relevant to the Shutdown Problem and Corrigibility
The Incompatibility of a Utility Indifference Condition with Robustly Making Sane Pure Bets
-------------------------------------------------------------------------------------------
### Summary
It is provably impossible for an agent to ro... |
668d9913-fc80-45b5-ae05-11f480c23313 | trentmkelly/LessWrong-43k | LessWrong | The case for aftermarket blind spot mirrors
Car side and rear-view mirrors are designed to give you a view around your car, but they typically don't show you the two spots directly to the side of your car (the "blind spots"). Luckily, you can increase the field of view of your side mirrors by over 200%[1] and virtuall... |
925c27d8-a96b-4f6a-96aa-e23f0d5776f6 | trentmkelly/LessWrong-43k | LessWrong | Techniques for optimizing worst-case performance
If powerful ML systems fail catastrophically, they may be able to quickly cause irreversible damage. To be safe, it’s not enough to have an average-case performance guarantee on the training distribution — we need to ensure that even if our systems fail on new distribut... |
e4accbed-1394-44ff-8007-b35dc0196ae1 | trentmkelly/LessWrong-43k | LessWrong | DontDoxScottAlexander.com - A Petition
Here's the petition Scott asked us to make.
This is a request for a specific action by the New York Times editors:
> We, the undersigned, urge the New York Times to respect Scott Alexander's request to not reveal his real name in a planned piece discussing the Slate Star Codex... |
1aca99a7-974f-4181-af87-4fee09947089 | trentmkelly/LessWrong-43k | LessWrong | A Roadmap: How to Survive the End of the Universe
In a sense, this plan needs to be perceived with irony because it is almost irrelevant: we have very small chances of surviving even next 1000 years and if we do, we have a lot of things to do before it becomes reality. And even afterwards, our successors will have com... |
c39248b1-a83b-4b89-90b1-243af31b0aa9 | trentmkelly/LessWrong-43k | LessWrong | What is a Counterfactual: An Elementary Introduction to the Causal Hierarchy
This is adapted from a segment of a talk I gave last year, primarily the portion at 6:09:00-6:11:00 of this stream. The example in this article originates from my paper "A Language for Counterfactual Generative Models," specifically Appendix ... |
14c841e8-6154-4cb5-b214-6d76bd5bdb3f | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post86
We recently released Subversion Strategy Eval: Can language models statelessly strategize to subvert control protocols? , a major update to our previous paper / blogpost , evaluating a broader range of models (e.g. helpful-only Claude 3.5 Sonnet) in more diverse and realistic settings (e.g. untrusted monito... |
5689e6aa-7b47-4b23-805c-500a1a6ab9a1 | trentmkelly/LessWrong-43k | LessWrong | The Brooklyn Society For Ethical Culture
Dale McGowan writes:
> In the past seven years or so, I’ve seen quite a few humanistic organizations from the inside — freethought groups, Ethical Societies, Congregations for Humanistic Judaism, UUs, etc. Met a lot of wonderful people working hard to make their groups succeed... |
effbf9cc-1e78-4e1c-9fb4-3a215fabc305 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] No Logical Positivist I
Today's post, No Logical Positivist I was originally published on 04 August 2008. A summary (taken from the LW wiki):
> Logical positivism was based around the idea that the only meaningful statements were those that could be verified by experiment. Unfortunately for positivism,... |
ea9a0e23-5207-47a2-a27c-f59e393981f5 | StampyAI/alignment-research-dataset/blogs | Blogs | Predictions of Human-Level AI Timelines
***Note: This page is out of date. See an [up-to-date version of this page](https://wiki.aiimpacts.org/doku.php?id=ai_timelines:predictions_of_human-level_ai_timelines) on our wiki.**
Updated 5 June 2015*
We know of around 1,300 public predictions of when human-level AI wil... |
18d6b151-294f-4942-8322-834dc07e556c | StampyAI/alignment-research-dataset/arxiv | Arxiv | REALab: An Embedded Perspective on Tampering
1 Introduction
---------------
††∗ Equal contribution.††DeepMind, London, UK. Correspondence to {ramanakumar,
Tampering problems, where an AI agent interferes with whatever represents or communicates its intended objective and pursues the resulting corrupted objective i... |
3c75ad4d-6826-495c-b504-77a8cf4a9e37 | trentmkelly/LessWrong-43k | LessWrong | Simple refutation of the ‘Bayesian’ philosophy of science
|
c9c28631-5274-439d-8e62-50729f5a174a | trentmkelly/LessWrong-43k | LessWrong | Progress Report 7: making GPT go hurrdurr instead of brrrrrrr
First progress post
Introduction
Starting off this project, I was using ROME modification to alter individual weights and get clusters of neurons in the model. I came up with the idea of partially impairing these clusters rather than fully ablating them. ... |
f14df5e1-c6ac-485d-a7b2-c60a3ff30a0b | trentmkelly/LessWrong-43k | LessWrong | Spoons and Myofascial Trigger Points
A lot of people use "spoons" to refer to their degree of willpower. This is a kind of Ego Depletion effect, but the science on ego depletion seems to be somewhat contested at the moment. I think it's possible that rather than it being universal, there are some people for whom it is... |
3ce5453d-2f8c-4a16-a9dc-9b59f5ed117c | StampyAI/alignment-research-dataset/special_docs | Other | The Dynamics of Exemplar and Prototype Representations Depend on Environmental Statistics.
The Dynamics of Exemplar and Prototype Representations
Depend on Environmental Statistics
Arjun Devraj (adevraj@princeton.edu)
Department of Computer Science, Princeton University
Qiong Zhang (qiongz@princeton.edu)
Princeton Neu... |
85555b5a-e8cb-46cc-a972-53e026222d8a | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Inaccessible information
Suppose that I have a great model for predicting “what will Alice say next?”
I can evaluate and train this model by checking its predictions against reality, but there may be many facts this model “knows” that I can’t easily access.
For example, the model might have a detailed representation... |
e3d21dfb-20e2-4c2b-a767-1559db29c022 | trentmkelly/LessWrong-43k | LessWrong | What Yann LeCun gets wrong about aligning AI (video)
I've spent a lot of time asking myself what I can do to reduce AI risk. I think one approach is to produce videos explaining the simple aspects of AI risk for laypeople to understand. I'm looking for feedback on this first video I made: https://www.youtube.com/watch... |
44d09adb-6999-4805-87ae-e83801b4ac7b | trentmkelly/LessWrong-43k | LessWrong | Don't judge a skill by its specialists
tl;dr: The marginal benefits of learning a skill shouldn't be judged heavily on the performance of people who have had it for a long time. People are unfortunately susceptible to these poor judgments via the representativeness heuristic.
Warn and beware of the following kludgy a... |
fd114a57-2c28-438c-bc04-982d17a97643 | trentmkelly/LessWrong-43k | LessWrong | What am I?
A bit of poetic prose.
What is my identity? What is the core that defines me? If I identify with all of humanity, then my core is as follows.
I am a sleeper, as I sleep and I dream. I am a thinker, as I have thoughts and I hear them in my mind. They differ and vacillate and sometimes take a sharp turn in c... |
50fce929-46b8-4c52-bb84-e999d68b0a04 | trentmkelly/LessWrong-43k | LessWrong | Obvious points
As mentioned previously, pointing out obvious things seems embarrassing to me. However, it also often seems very valuable. That might seem obvious to you. Even so, this post will elaborate on this obvious point.
The set of things an intellectual would like to claim are obvious will tend to be much larg... |
21ffbbc9-bd5f-41e6-9dd0-f611840f000f | trentmkelly/LessWrong-43k | LessWrong | How to fly safely right now?
I have friends who need to fly to different countries in the next days and weeks, and I'd like to help them avoid getting the virus. Please help me figure out how to avoid catching it on the planes.
I'll leave my current thoughts in an answer to this question. |
457cc6a4-8dd7-4b25-ae5e-3492c21e23dc | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post627
Goal First things first: this is not (yet) aimed at solving the whole corrigibility problem, or even the whole shutdown problem. The main thing this proposal is intended to do is to get past the barriers MIRI found in their old work on the shutdown problem . In particular, in a toy problem basically-identi... |
73d77101-cbf1-4751-bccc-ce1e64ee5f9b | trentmkelly/LessWrong-43k | LessWrong | The Tree of Life, and a Note on Job
This is another one cross-posted from my New Savanna blog. It's old, 11 years (and most of it is even older than that). It's an interpretation of the Book of Job, which is not at all the sort of thing I'd ordinarily post here. I certainly don't see much of that kind of thing going ... |
f5486c85-0943-4100-b4a7-c5457b71afff | trentmkelly/LessWrong-43k | LessWrong | How to prevent collusion when using untrusted models to monitor each other
Suppose you’ve trained a really clever AI model, and you’re planning to deploy it in an agent scaffold that allows it to run code or take other actions. You’re worried that this model is scheming, and you’re worried that it might only need to t... |
5d99916d-fce9-4749-90c9-b0a41b83a681 | trentmkelly/LessWrong-43k | LessWrong | Effective altruism is self-recommending
A parent I know reports (some details anonymized):
> Recently we bought my 3-year-old daughter a "behavior chart," in which she can earn stickers for achievements like not throwing tantrums, eating fruits and vegetables, and going to sleep on time. We successfully impressed on ... |
d110d7cd-3a48-4453-97ba-dc8dacefa967 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Considerations, Good Practices, Risks and Pitfalls in Developing AI Solutions Against COVID-19
1 Introduction
---------------
In recent months, an already remarkable body of research on potential applications of Artificial Intelligence (AI) and Machine Learning (ML) to address the health and societal challenges cre... |
126dcf20-3d1d-452d-a6cd-c4b8bd990419 | trentmkelly/LessWrong-43k | LessWrong | ‘We’re changing the clouds.’ An unforeseen test of geoengineering is fueling record ocean warmth
For decades humans have been emitting carbon dioxide into the atmosphere, creating a greenhouse effect and leading to an acceleration of the earth's warming.
At the same time, humans have been emitting sulphur dioxide, a ... |
6343f07c-cab9-4ce9-be0d-10b989aea496 | trentmkelly/LessWrong-43k | LessWrong | Noise Simplifies
They say that the flap of a butterfly’s wings can change the course of a hurricane. Sadly, this does not give us an effective strategy for controlling hurricane trajectories - if a flap of my butterfly’s wings can change the hurricane’s course, then so can the flap of any other butterfly’s wings. Unle... |
c3e2ac29-e7c6-4170-9e7d-76a6e2567003 | trentmkelly/LessWrong-43k | LessWrong | Nick Bostrom TED talk on world's biggest problems
http://www.ted.com/talks/nick_bostrom_on_our_biggest_problems.html |
2e1a66af-3be7-4a06-82ea-93029b1b73e7 | trentmkelly/LessWrong-43k | LessWrong | Organisation-Level Lock-In Risk Interventions
Epistemic status: my own thoughts and opinions after thinking about organisation structures and dynamics as an intervention point for lock-in risk for about 5 hours. My thoughts here represent the opinion of Formation Research at the time of writing.
TL;DR
We believe loc... |
24432a74-c239-429f-abab-80d25aec22d5 | trentmkelly/LessWrong-43k | LessWrong | Narratives of business models
There’s a common theme when discussing business models over the internet, which usually revolves around its optimal form.
What’s the most effective model? Monthly vs. yearly subscriptions, the relevance of ads, and the appeal of lifetime plans are debates I often come across on my Twitte... |
8292bdba-cdd2-4804-82fe-3a25ede8fc08 | trentmkelly/LessWrong-43k | LessWrong | Aspergers Poll Results: LW is nerdier than the Math Olympiad?
Followup to: Do you have High Functioning Aspergers Syndrome?
----------------------------------------
EDIT: To combat nonresponse bias, I'd appreciate it if anyone who considered the poll and decided not to fill it in would go and do so now, but th... |
d509da66-4b06-403d-b3ff-f42dfe56852d | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Radical Probabilism [Transcript]
*(Talk given on Sunday 21st June, over a zoom call with 40 attendees. Abram Demski is responsible for the talk, Ben Pace is responsible for the transcription)*
Talk
----
**Abram Demski:** I want to talk about this idea that, for me, is an update from the logical induction result that... |
760e4f2d-845e-4375-a985-22b87b58b85e | trentmkelly/LessWrong-43k | LessWrong | Slowing down AI progress is an underexplored alignment strategy
The emotional burden of watching the world end
My current beliefs about AI timelines have made my life significantly worse. I find thoughts about the ever-shrinking timeline to AGI invading nearly every aspect of my life. Every choice now seems to be affe... |
e5f4fc12-09e1-4ec0-8de8-e27bdec0c997 | trentmkelly/LessWrong-43k | LessWrong | Countering AI disinformation and deep fakes with digital signatures
According to The Economist, disinformation campaigns (often state-sponsored) use "AI to rewrite real news stories":
> In early March a network of websites, dubbed CopyCop, began publishing stories in English and French on a range of contentious issue... |
7f3b6968-b91e-4988-b7fe-24fab63e6bda | trentmkelly/LessWrong-43k | LessWrong | Jaan Tallinn's 2021 Philanthropy Overview
to follow up my philantropic pledge from 2020, i've updated my philanthropy page with 2021 results.
in 2021 i made $22M worth of endpoint grants — exceeding my commitment of $14.4M (20k times $718.11 — the minimum price of ETH in 2021). notes:
* this number includes $1.9M t... |
ebbf4739-a635-4a30-b241-8344d0722b91 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Quick survey on AI alignment resources
TLDR: In collaboration with CEA, an external collaborator is trying to determine the value of various AI alignment resources. If you are interested in or currently working on AI alignment, please fill out [**this survey**](https://www.guidedtrack.com/programs/qd2mn4k/run). We exp... |
35885018-7995-4d4e-b875-0fa0e0a94f9b | trentmkelly/LessWrong-43k | LessWrong | Meetup : London Meetup, 24th Feb
Discussion article for the meetup : London Meetup, 24th Feb
WHEN: 24 February 2013 02:00:00PM (+0000)
WHERE: Holborn, London
A meetup in the Shakespeare's Head pub by Holborn tube station. We meet every other Sunday at 2pm. Everyone is welcome.
We also have a Google group.
Discuss... |
e6e707ec-7ed6-4f86-bd28-272d26426d51 | trentmkelly/LessWrong-43k | LessWrong | Intended enjoyment
I am sometimes unsure what is meant to be enjoyed in things, for instance the short story The Gioconda Smile, or In the mood for love which people often enjoy a lot. Which seems like it shouldn’t be a problem, as long as I find something to enjoy in them. But it also seems like I would be seriously ... |
a73a8a66-e62b-468a-8849-1ae45d1b6c7a | StampyAI/alignment-research-dataset/youtube | Youtube Transcripts | Nicola Croce: ODD, data labeling and the problem of representing knowledge for AVs
uh discuss while i'm talking because i
just put together some
eight or nine slides but it's just
something for me to
guide my my talk and discussion but
again i'm very much
open to an interactive session also
because all these topics ar... |
14e9d0c1-4e93-4e0e-9f66-1b772f6cb079 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | [RETRACTED] It's time for EA leadership to pull the short-timelines fire alarm.
[EDIT 4/10/2022: This post was rash and ill-conceived, and did not have clearly defined goals nor meet the vaguely-defined ones. I apologize to everyone on here; you should probably update accordingly about my opinions in the future. In re... |
ddddb7b2-0115-4e28-96e0-e9c7b2390ed4 | StampyAI/alignment-research-dataset/special_docs | Other | AI Services: Introduction v1.3
Sharing status:
\* This document is currently written with long-termist audience in mind. As such, please do not advertise it where the missing context could cause idea inoculation.
\* To prompt the creation of a general-audience version, you can [email me](mailto:vojta.kovarik@gmail.co... |
e94819e1-c879-4ffe-9bf2-846f776fb0c6 | trentmkelly/LessWrong-43k | LessWrong | Surface Analogies and Deep Causes
Followup to: Artificial Addition, The Outside View's Domain
Where did I acquire, in my childhood, the deep conviction that reasoning from surface similarity couldn't be trusted?
I don't know; I really don't. Maybe it was from S. I. Hayakawa's Language in Thought and Action, or eve... |
fdc1901b-bc62-4f0a-9530-58cb7749fd3c | trentmkelly/LessWrong-43k | LessWrong | Priming effects are fake, but framing effects are real
> A few decades ago, it was pretty common to mush together priming effects and framing effects and see them as two closely connected parts of a single Bigger Truth about the human mind. Of course, everyone understood that the effects themselves were a bit differen... |
89d90b93-195b-4889-a396-41fa8934ba63 | StampyAI/alignment-research-dataset/blogs | Blogs | Reader reactions and update on "Where's Today's Beethoven"
*Click lower right to download or find on Apple Podcasts, Spotify, Stitcher, etc.*
There were a lot of great comments on [Where's Today's Beethoven?](https://www.cold-takes.com/wheres-todays-beethoven/) and followup pieces, from the [comments](https://forum... |
b38cac9f-c108-40c8-905b-7a84cfba88e2 | trentmkelly/LessWrong-43k | LessWrong | Anomalous Tokens in DeepSeek-V3 and r1
“Anomalous”, “glitch”, or “unspeakable” tokens in an LLM are those that induce bizarre behavior or otherwise don’t behave like regular text.
The SolidGoldMagikarp saga is pretty much essential context, as it documents the discovery of this phenomenon in GPT-2 and GPT-3.
But, as... |
13f5e472-4b83-467d-9cd2-fb82e0d1cac2 | trentmkelly/LessWrong-43k | LessWrong | The Parable of the Crimp
If you watch really proficient rock climbers, you’ll see they can hold themselves up, dozens of feet above the ground, with just the tips of their fingers on the tiniest ledge of rock, about the width of a pencil, called a crimp (example image above). If I had not seen it, I would have said it... |
40fb2d7b-e08a-4039-915d-5b7fd1c2f3cb | trentmkelly/LessWrong-43k | LessWrong | How To Observe Abstract Objects
Opening Thoughts
What is this thing and what is the point of it?
I’m trying to build a branch of rationality that is about looking at ideas and problems “for real”, thinking about them “for real”, “as a whole person”, and “without all the bullshit in the way”. This is a mini worksho... |
b05bad49-60b6-4018-823c-ee2c3c02ede2 | trentmkelly/LessWrong-43k | LessWrong | What if "status" IS a terminal value for most people?
[Inspired by a few of the science bits in HP:MOR, and far more so by the discussions between Draco and Harry about "social skills". Shared because I suspect it's an insight some people would benefit from.]
One of the more prominent theories on the evolution of hum... |
f6eb651f-4386-4676-be81-7fa7acf5d249 | trentmkelly/LessWrong-43k | LessWrong | Implied "utilities" of simulators are broad, dense, and shallow
This is a quick attempt at deconfusion similar to instrumentality. Same ideas, different angle.
Extremely broad, dense reward functions constrain training-compatible goal sets
Predictors/simulators are typically trained against a ground truth for every ... |
69349b17-f8cd-4a92-9730-20f08b18a5d5 | trentmkelly/LessWrong-43k | LessWrong | Breaking the vicious cycle
You may know me as the guy who posts a lot of controversial stuff about LW and MIRI. I don't enjoy doing this and do not want to continue with it. One reason being that the debate is turning into a flame war. Another reason is that I noticed that it does affect my health negatively (e.g. my ... |
a5ffcedd-2089-4ba5-80f7-566343061673 | trentmkelly/LessWrong-43k | LessWrong | Estimating effective dimensionality of MNIST models
The local learning coefficent λ is a measure of a model's "effective dimensionality" that captures its complexity (more background here). Lau et al recently described a sampling method (SGLD) using noisy gradients to find a stochastic estimate ^λ in a computationally... |
f8ee8745-4b84-4fe8-9a95-0c290fd22064 | trentmkelly/LessWrong-43k | LessWrong | Sycophancy to subterfuge: Investigating reward tampering in large language models
New Anthropic model organisms research paper led by Carson Denison from the Alignment Stress-Testing Team demonstrating that large language models can generalize zero-shot from simple reward-hacks (sycophancy) to more complex reward tamp... |
14c1c188-f555-4671-80a7-21c93585e30e | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Is it worth making a database for moral predictions?
*Epistemic status: Confused about what probabilities people assert to different alignment theories working.*
I essentially want to know whether it would be worth spending some of my time working on conceiving a database like <https://www.moralmachine.net/> with mor... |
7953d5b0-a1bb-43e6-b54f-1f4cefc0f7cf | trentmkelly/LessWrong-43k | LessWrong | Meetup : Boston - Optimizing Empathy Levels
Discussion article for the meetup : Boston - Optimizing Empathy Levels
WHEN: 02 March 2014 02:00:00PM (-0500)
WHERE: Citadel, 98 Elm St Apt 1, Somerville, MA
Joe Schneider will give a talk on Optimizing Empathy Levels:
-What empathy is useful for.
-Why LWers (and others... |
0f58a555-6e8c-476a-9433-76c3d15f5eae | trentmkelly/LessWrong-43k | LessWrong | Thoughts from a conversation on quantum immortality
[Today, I participated in a conversation about the idea of quantum immortality. I decided to summarise some of the thoughts that came up in this short post. Therefore, it should be viewed as a report on a discussion rather than an attempt at a proper post.]
Assume t... |
53bb0b26-561b-4774-80c3-d94220207d10 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | What is the interpretation of the do() operator?
To reductively explain causality, it has to be explained in non-causal terms, most likely in terms of total propability distributions. Pearl explains causality in terms of causal graphs which are created by conditionalizing the propability distribution on not .mjx-chtml... |
a3503791-cf16-493d-b808-330a71d5ba8f | StampyAI/alignment-research-dataset/lesswrong | LessWrong | On the Loebner Silver Prize (a Turing test)
The Metaculus question "When will the first weakly general AI system be devised, tested, and publicly announced?" has four resolution criteria, and in my opinion, reliably passing a Loebner-Silver-prize-equivalent Turing test is the hardest, since it is the only one that is ... |
1e4502e7-009c-4c1a-abda-615bd1f5e53d | trentmkelly/LessWrong-43k | LessWrong | Is community-collaborative article production possible?
When I showed up at the Singularity Institute, I was surprised to find that 30-60 papers' worth of material was lying around in blog posts, mailing list discussions, and people's heads — but it had never been written up in clear, well-referenced academic articles... |
2ec89892-fea5-4755-8a67-d4e428404790 | trentmkelly/LessWrong-43k | LessWrong | ML Safety Research Advice - GabeM
This is my advice for careers in empirical ML research that might help AI safety (ML Safety). Other ways to improve AI safety, such as through AI governance and strategy, might be more impactful than ML safety research (I generally think they are). Skills can be complementary, so this... |
a0d864a5-c78b-49f8-8978-d8416ba9149a | StampyAI/alignment-research-dataset/lesswrong | LessWrong | On A List of Lethalities
Response to (Eliezer Yudkowsky): [A List of Lethalities](https://www.lesswrong.com/posts/uMQ3cqWDPHhjtiesc/agi-ruin-a-list-of-lethalities).
*Author’s Note: I do not work in AI Safety, lack technical domain knowledge and in many ways am going to be wrong. I wasn’t going to write this to avoid... |
5bc1e419-de2d-400b-a0a6-30342c1565f7 | trentmkelly/LessWrong-43k | LessWrong | A comparison of causal scrubbing, causal abstractions, and related methods
Summary: We explain the similarities and differences between three recent approaches to testing interpretability hypotheses: causal scrubbing, Geiger et al.'s causal abstraction-based method, and locally consistent abstractions. In particular, ... |
fe5c25d2-a286-46a8-9225-75cbbbabd617 | trentmkelly/LessWrong-43k | LessWrong | How to not be an alarmist
> "I apologize. I didn't mean-- may I express my concern as calmly and respectfully as I can?"
>
> - Prof. Legasov on the consequences of an RBMK reactor meltdown (HBO's Chernobyl)
> "This is not going to end well."
>
> "There is no evidence of that-"
>
> "This is not going to end well... |
2007d60f-3f0d-435a-be38-48fbf2b6d66d | trentmkelly/LessWrong-43k | LessWrong | Unconvenient consequences of the logic behind the second law of thermodynamics
Let's consider a simple universe consisting of an ideal gas inside a box (which means a number of massive points behaving as prescribed by Newton's physics with completely elastic collisions).
After the system reaches the thermodynamic eq... |
c48e4b98-c939-448d-85e3-9ba705b97a01 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Insufficient Values
I’m [Jose](https://www.lesswrong.com/posts/hKNJSiyzB5jDKFytn/?commentId=uTZqYft4LJpCE36je). I realized recently I wasn’t taking existential risk seriously enough, and in April, a year after I first applied, I started running a MIRIx group in my college. I’ll write summaries of the sessions that I... |
5dceaf5a-06a2-43dc-8b5e-c72f86f19a7d | trentmkelly/LessWrong-43k | LessWrong | Maximizing impact at startup with high EA potential?
An early stage startup (SF Bay) has brought me on in a position of technical leadership. The product has the potential to save high-billions of human hours, and eliminate low-billions of tons of carbon emissions. Given that all startups are wildly optimistic, moving... |
e75e233a-9399-49fc-9cdc-5f24140fabe8 | trentmkelly/LessWrong-43k | LessWrong | Hooray for stepping out of the limelight
From maybe 2013 to 2016, DeepMind was at the forefront of hype around AGI. Since then, they've done less hype. For example, AlphaStar was not hyped nearly as much as I think it could have been.
I think that there's a very solid chance that this was an intentional move on the p... |
bd692043-816c-424a-8cde-9f57c102a5aa | StampyAI/alignment-research-dataset/arxiv | Arxiv | Discriminator-Actor-Critic: Addressing Sample Inefficiency and Reward Bias in Adversarial Imitation Learning
1 Introduction
---------------
Inverse Reinforcement Learning (IRL) (Abbeel & Ng, [2004](#bib.bib2)) is a prominent way to provide a dense reward for tasks with inherently sparse rewards or hard exploration.... |
a1d52f50-ae3a-4935-8440-914dd79cd27a | StampyAI/alignment-research-dataset/arxiv | Arxiv | The Next Decade in AI: Four Steps Towards Robust Artificial Intelligence
The Next Decade in AI:
Four Steps Towards Robust Artificial Intelligence
Gary Marcus
Robust AI
17 February 2020
Abstract
Recent research in artificial intelligence and machine
learning has largely emp... |
1c2b5ea5-647c-46af-a3c4-804629f5bc93 | StampyAI/alignment-research-dataset/youtube | Youtube Transcripts | Why Not Just: Think of AGI Like a Corporation?
hi so I sometimes see people saying
things like okay so your argument is
that at some point in the future we're
going to develop intelligent agents that
are able to reason about the world in
general and take actions in the world to
achieve their goals
these agents might h... |
bda1cd97-8b25-4649-b5f7-d7865e703420 | trentmkelly/LessWrong-43k | LessWrong | NEW TIME: Sydney Less Wrong meetup, 23/4, 3PM
Hi everyone! There was an attempt for a Sydney (Australia) Less Wrong meetup a while ago, which failed, which was sad. So I'm starting a new one!
I'll definitely be at the meetup for three hours, regardless of who else is able to make it. So this is definitely going ahead... |
870f441d-aa1f-43e4-b431-e4bda2aeefda | trentmkelly/LessWrong-43k | LessWrong | My Overview of the AI Alignment Landscape: A Bird's Eye View
Disclaimer: I recently started as an interpretability researcher at Anthropic, but I wrote this doc before starting, and it entirely represents my personal views not those of my employer
Intended audience: People who understand why you might think that AI A... |
78e4bd50-8a02-4a5a-ba2f-460f0cf06510 | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post2617
This is an update on the work on AI Safety via Debate that we previously wrote about here . Authors and Acknowledgements The researchers on this project were Elizabeth Barnes and Paul Christiano, with substantial help from William Saunders (who built the current web interface as well as other help), Joe C... |
6e05dc83-1a9d-48bc-8bc3-9d7287540371 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Curiosity as a Solution to AGI Alignment
AGI (Artificial General Intelligence) seems to be just around the corner. Or maybe not. Either way it might be humanity’s last ever invention— the greatest of all time or the ultimate doom machine. This is a “thinking-out-loud” piece about how we can avoid the doom machine scen... |
af399ac1-df09-4e45-a7da-3ba11cb5d3f7 | trentmkelly/LessWrong-43k | LessWrong | Stanislav Petrov Quarterly Performance Review
September 26 is the anniversary of the 1983 Soviet nuclear false alarm incident—item #24 on Wikipedia’s upsettingly long “List of Nuclear Close Calls”—in which Lieutenant Colonel Stanislav Petrov did an excellent job at not starting World War III.
Quarterly Performance R... |
37ebb087-2ea2-4ada-97a1-699f30b964f7 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | AI strategy career pipeline
The pipeline for (x-risk-focused) AI strategy/governance/forecasting careers has never been strong, especially for new researchers. But it feels particularly weak recently (e.g. no summer research programs this year from Rethink Priorities, SERI SRF, or AI Impacts, at least as of now, and a... |
57209cba-70ba-4f14-a19a-eadb85d67f89 | trentmkelly/LessWrong-43k | LessWrong | Pitfalls of the agent model
This is independent research. To make further posts like this possible, please consider supporting me.
----------------------------------------
Outline
* I study the agent model as a frame for understanding the phenomenon of entities that exert influence over the future.
* I focus on ... |
3d73bc92-dbbf-4554-a759-074a02fdb130 | trentmkelly/LessWrong-43k | LessWrong | The Case for a Journal of AI Alignment
When you have some nice research in AI Alignment, where do you publish it? Maybe your research fits with a ML or AI conference. But some papers/research are hard sells to traditional venues: things like Risks from Learned Optimization, Logical Induction, and a lot of research don... |
8f943c65-6d8d-49a3-a23e-f6eef30e6b72 | trentmkelly/LessWrong-43k | LessWrong | A Longlist of Theories of Impact for Interpretability
I hear a lot of different arguments floating around for exactly how mechanistically interpretability research will reduce x-risk. As an interpretability researcher, forming clearer thoughts on this is pretty important to me! As a preliminary step, I've compiled a l... |
2b48bd54-7f1b-4649-bab0-f376b6235ae0 | trentmkelly/LessWrong-43k | LessWrong | Ascetic hedonism
In being ascetic, you abandon the usual sources of material pleasure, guided by the benefits of the lifestyle: you use less money and effort on avoidable pleasures, you can better focus your mind on the spiritual and the creative, and you make yourself resilient to potentially losing these pleasures.
... |
6316f6e4-ccdb-46eb-a6fe-ca8bafe50fa5 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Feedback in Imitation Learning: The Three Regimes of Covariate Shift
1 Introduction
---------------
Figure 1: A common exampl... |
b2e67095-c67c-4bed-bdea-8ef8e8c8ae7e | trentmkelly/LessWrong-43k | LessWrong | You Can’t Skip Exploration: Why understanding experimentation and taste is key to understanding AI
This essay is part 1 of a series on the role of exploration in AI and the implications for AI development and governance.
This part introduces exploration and research taste, as well as discussing their role in research... |
d0eb828b-7526-464b-880f-9d1302932cd8 | StampyAI/alignment-research-dataset/arxiv | Arxiv | SafeLife 1.0: Exploring Side Effects in Complex Environments
1 Introduction
---------------
Safety problems for reinforcement learning (RL) have received growing attention in recent years (see e.g. garcia2015comprehensive; amodei:2016a; ortega2018building for surveys). It has been recognized that the capabilities o... |
5086b8ee-d4bc-4d90-979b-07f16f02e340 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | AI ethics: the case for including animals (my first published paper, Peter Singer's first on AI)
Hi! I am Tse Yip Fai. This is a linkpost to my first [published paper](https://link.springer.com/article/10.1007/s43681-022-00187-z) (second author). Since the paper is open access, I won't be doing any copy-and-paste here... |
8178cf36-c153-4efc-b422-db0ab0fe0892 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Global computing capacity
This is a linkpost for [this](https://aiimpacts.org/global-computing-capacity/) article from [AI Impacts](https://aiimpacts.org/#gsc.tab=0).
> *[This page is out of date and its contents may have been inaccurate in 2015, in light of new information that we are yet to integrate.]*
>
> Compu... |
1612b7c3-5752-4ea5-8b7b-99a9b571e1fc | trentmkelly/LessWrong-43k | LessWrong | A Simple Toy Coherence Theorem
This post presents a simple toy coherence theorem, and then uses it to address various common confusions about coherence arguments.
Setting
Deterministic MDP. That means at each time t there's a state S[t][1], the agent/policy takes an action A[t] (which can depend on both time t and ... |
638579b3-04a6-43a3-a7d8-e5d51b13e9b3 | trentmkelly/LessWrong-43k | LessWrong | Changes in Community Dynamics: A Follow-Up to 'The Berkeley Community & the Rest of Us'
I've received a few notifications in the last couple weeks that this post I wrote in 2018 about the relationship between the rationality community in Berkeley and the broader rationality community elsewhere has received some upvote... |
353c5d19-3b7c-400e-92d7-dcf8fb2e5ac8 | trentmkelly/LessWrong-43k | LessWrong | Wear a Helmet While Driving a Car
A 2006 study showed that “280,000 people in the U.S. receive a motor vehicle induced traumatic brain injury every year” so you would think that wearing a helmet while driving would be commonplace. Race car drivers wear helmets. But since almost no one wears a helmet while driving a ... |
e919402d-371f-40e6-9de2-92bc8d0a77cf | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Making EA more inclusive, representative, and impactful in Africa
Authors: Ashura Batungwanayo (University of KwaZulu Natal) and Hayley Martin (University of Cape Town)
DISCLAIMER: This paper draws from our collective experiences, perspectives, and conversations with community builders in Africa who share similar ... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.