id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
c760096a-3489-474d-b53e-9061bb815ebf | trentmkelly/LessWrong-43k | LessWrong | Conjecture on Addiction to Meta-level Solutions
Related: Meta Addiction
Says Eliezer in LessWrong Q&A (16/30):
> In one sense, a whole chunk of LessWrong is more or less my meta-thinking skills.
When we become more rational, it's usually because we invent a new cognitive rule that:
1. Explains why certain beliefs... |
1b1cceeb-6415-4162-b4b4-7ab85d6baaa6 | trentmkelly/LessWrong-43k | LessWrong | Why epidemiology will not correct itself
We're generally familiar here with the appalling state of medical and dietary research, where most correlations turn out to be bogus. (And if we're not, I have collected a number of links on the topic in my DNB FAQ that one can read, see http://www.gwern.net/DNB%20FAQ#flaws-in-... |
35b758d8-d148-431f-964d-a8b0d0a4a1a9 | trentmkelly/LessWrong-43k | LessWrong | Travel Journal: Hawaii
Putanumonit is on the road. With a few hours before my next flight, I have time to share some wiki impressions from a week in Hawaii.
ALOHA
The Hawaiian language is built from some pretty basic blocks. It uses seven consonant sounds (h, k, l, m, n, p, w) and five vowels. Every syllable ends ... |
5b93c279-272f-4a84-9f83-52676c3e2812 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Extensions and Intensions
Today's post, Extensions and Intensions was originally published on 04 February 2008. A summary (taken from the LW wiki):
> You try to define a word using words, in turn defined with ever-more-abstract words, without being able to point to an example. "What is red?" "Red is a ... |
fba521af-1a3c-4650-8a05-21e8fb288a71 | trentmkelly/LessWrong-43k | LessWrong | Active vs Passive Distraction
When life gets difficult, it can often become tempting to distract yourself from your own thoughts and feelings. Loved ones may scold us for this, telling us that we should confront our feelings, because distracting ourselves doesn't accomplish anything. However, distraction is an importa... |
02461bf8-9ebb-4550-9b75-b67393ae4031 | trentmkelly/LessWrong-43k | LessWrong | AXRP Episode 32 - Understanding Agency with Jan Kulveit
YouTube link
What’s the difference between a large language model and the human brain? And what’s wrong with our theories of agency? In this episode, I chat about these questions with Jan Kulveit, who leads the Alignment of Complex Systems research group.
Topic... |
59467268-9677-4787-8739-80434d708837 | trentmkelly/LessWrong-43k | LessWrong | [Link] OpenAI on why we need social scientists
https://distill.pub/2019/safety-needs-social-scientists/
> We believe the AI safety community needs to invest research effort in the human side of AI alignment. Many of the uncertainties involved are empirical, and can only be answered by experiment. They relate to the p... |
d4679027-3b89-4314-b183-2b1586c5cff0 | trentmkelly/LessWrong-43k | LessWrong | The "Friendship is Witchcraft" expectation test
My mother won't watch animated movies. It doesn't matter what the content is. Whether it's Sponge Bob or Grave of the Fireflies, she believes that animation is used only for shows for children, and that adults shouldn't watch shows for children. She's incapable of cha... |
1699910a-29fb-4011-88c1-7cabe35558f0 | trentmkelly/LessWrong-43k | LessWrong | Goal-directedness is behavioral, not structural
Goal-directedness is the term used by the AI Safety community to point to a specific property: following a goal. It comes from Rohin Shah's post in his sequence, but the intuition pervades many safety issues and current AI approaches. Yet it lacks a formal definition, or... |
07494ad5-eadf-4084-b442-707ff009fadf | trentmkelly/LessWrong-43k | LessWrong | Ethics Notes
Followup to: Ethical Inhibitions, Ethical Injunctions, Prices or Bindings?
(Some collected replies to comments on the above three posts.)
From Ethical Inhibitions:
> Spambot: Every major democratic political leader lies abundantly to obtain office, as it's a necessity to actually persuade the voters... |
5d1e7ba2-4f32-4200-902f-bde9e00df0ce | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "Luna had too many Calls to Adventure. If there was an author writing Luna's life, Luna would file a complaint. Luna had three magical artifacts of incredible power and at least two of them were maps. What kind of author gives her hero two magical maps? It's redundant.
"Are you a map of 10th century Wessex?" Luna asked... |
40eaa077-167d-4725-9806-696c3c14d2cb | trentmkelly/LessWrong-43k | LessWrong | Is Kiryas Joel an Unhappy Place?
I was browsing my RSS feed, as one does, and came across a New York Times article, "A Village With the Numbers, Not the Image, of the Poorest Place", about the Satmar Hasidic Jews of Kiryas Joel (NY).
Their interest lies in their extraordinarily high birthrate & population growth, and... |
7fc39ff3-eaaf-49ac-91f8-f4801563824a | trentmkelly/LessWrong-43k | LessWrong | Cambridge Less Wrong meetup Sunday 27 June
I will be at Clear Conscience Cafe at 581 Massachusetts Avenue Cambridge, MA looking for other Less Wrong people on Sunday the 27th. This breaks the previously established pattern of meeting on the third Sunday of each month, to avoid conflicting with Father's Day. As usual, ... |
c42fcdf4-93cd-4128-a42e-d492d4f76552 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Three Kinds of Competitiveness
*Crossposted from [AI Impacts](https://aiimpacts.org/three-kinds-of-competitiveness/)*
*[Epistemic status: I wrote this for [Blog Post Day II.](https://www.lesswrong.com/posts/FredSMMXc77kMAquA/blog-post-day-ii) Sorry it's late.]*
In this post, I distinguish between three different kin... |
f8b7bfe5-6fe8-42bd-bdc1-27e3832997cd | 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, ... |
cc4ea95e-bb24-4617-b522-f78e6e11789f | trentmkelly/LessWrong-43k | LessWrong | What will the long-term future of employment look like?
Current debates over the automation of human jobs by AI tend to focus on how fast it will proceed: some jobs will be safe for the next decade, some for the next half-century, some for the foreseeable future. In this essay I want to take a different approach, by a... |
a78e85cb-7733-4ea2-ba49-e2509a27d19e | trentmkelly/LessWrong-43k | LessWrong | The World: An Introduction
Previous essays have discussed human reasoning, language, goals, and social dynamics. Mathematics, physics, and biology were cited to explain patterns in human behavior, but little has been said about humanity’s place in nature, or about the natural world in its own right.
Just as it was us... |
8d903689-1655-4069-8296-31c548563f61 | trentmkelly/LessWrong-43k | LessWrong | [Link] Cosmological Infancy
A post by Nick Land who some of you are probably already following either on his blog Outside In or at Urban Future.
> There is a ‘problem’ that has been nagging at me for a long time – which is that there hasn’t been a long time. It’s Saturday, with no one around, or getting drunk, or so... |
9eaa6b13-86fe-4a25-a9ef-46b50f01c7e7 | trentmkelly/LessWrong-43k | LessWrong | Covid 5/5/22: A Lack of Care
China cares a lot about preventing Covid.
I haven’t written an additional China post because my sources have not turned up much additional information, and the situation does not seem to have dramatically changed, so I’m waiting until situation warrants an update. One development was mass... |
25f6996b-0741-46f6-83ae-4efa6775e7c4 | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "Logical TimeThe main purpose of this post is to introduce the concept of logical time. The idea was mentioned in Scott's post, Bayesian Probability is for things that are Space-like Separated from You. It was first coined in a conference call with, Daniel Demski, Alex Mennan, and perhaps Corey Staten and Evan Lloyd --... |
56722532-7f14-4f6b-8df5-115817be8ac7 | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "(This is a post from a daily blogging experiment I did at neelnanda.io, which I thought might also fit the tastes of LessWrong)There are two kinds of people in the world: People who actually do things, and people who don’t.Now I’ve got the unnecessarily provocative opening line out of the way, what do I actually mean ... |
9dfdd7b1-2cd4-4a79-87f5-41378afc18a1 | trentmkelly/LessWrong-43k | LessWrong | Feature Request to OpenAI: Share button in ChatGPT
Description of the feature requested
1. Share Button: Places text into system clipboard.
2. This text is the entire conversation[1] appended with the encrypted hash of the entire conversation (encrypted with ChatGPT's private key)
Why it would be good
Currently i... |
d621cb19-c05b-40c0-8615-c2407b66ded6 | trentmkelly/LessWrong-43k | LessWrong | The Equation of Knowledge
My book The Equation of Knowledge has just been published at CRC Press, and I'm guessing that it may be of interest to readers of LessWrong. The book aims to be a somewhat accessible and very complete introduction to Bayesianism. No prior knowledge is needed; though some sections require an i... |
d3ea0dff-2646-4095-9e18-b9e9a15ddc8e | trentmkelly/LessWrong-43k | LessWrong | Native Russian speakers wanted (for proofreading LW texts)
As part of an effort to familiarize Russian-speaking audience with Less Wrong materials, I am participating in an ongoing translation of a certain sequence of texts to Russian.
Now I need volunteers who are willing to read the text and give feedback on its qu... |
37c8293d-fd6c-4f4a-be07-1bbc38880109 | trentmkelly/LessWrong-43k | LessWrong | Project Proposal: Gears of Aging
Imagine building an airplane the way we do biology research.
Hundreds of separate research groups each pick different airplane parts to work on. There’s no particular coordination in this; people choose based first on what they know how to build, second on what seems fun/interesting/f... |
3e82b4dd-ca27-40b2-a271-460a4b85cbd8 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | 8 examples informing my pessimism on uploading without reverse engineering
*(If you’ve already read everything I’ve written, you’ll find this post pretty redundant. See especially my old posts*[*Building brain-inspired AGI is infinitely easier than understanding the brain*](https://www.lesswrong.com/posts/PTkd8nazvH9H... |
9f21d404-6eb4-4575-9c9f-cd5cd66fb463 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Real-Time Research Recording: Can a Transformer Re-Derive Positional Info?
**New experiment:** Recording myself real-time as I do mechanistic interpretability research! I try to answer the question of what happens if you train a toy transformer without positional embeddings on the task of "predict the previous token" ... |
3d12e046-9716-4d56-a71b-aa0c42446032 | trentmkelly/LessWrong-43k | LessWrong | LW systemic bias: US centrism
Recently, I have noticed a cultural bias for the United States running through LW threads. It is perhaps to be expected of an English-language website, but for one that is about, among other things, overcoming bias, it is important to recognize one's own.
Aspects of the bias I have obse... |
2421aa2d-9e23-4b35-aae7-06177e5238ac | trentmkelly/LessWrong-43k | LessWrong | June 2019 gwern.net newsletter
None |
cad536ec-a516-418e-900a-72576168a17d | trentmkelly/LessWrong-43k | LessWrong | Working towards AI alignment is better
There are people not motivated to solve AI alignment, who do work related to AI alignment. E.g. people work on adversarial robustness, understanding how to do science mechanically, or on advancing other paradigms that are more interpretable than modern ML. These people might be i... |
03ed1839-9f7f-4a94-ab99-3ac4d98e701e | trentmkelly/LessWrong-43k | LessWrong | Ask OB: Leaving the Fold
Followup to: Crisis of Faith
I thought this comment from "Jo" deserved a bump to the front page:
> "So here I am having been raised in the Christian faith and trying not to freak out over the past few weeks because I've finally begun to wonder whether I believe things just because I was rai... |
a5984dc1-015b-4d82-a284-9af549618564 | trentmkelly/LessWrong-43k | LessWrong | Fundamental question: What determines a mind's effects?
[Metadata: crossposted from https://tsvibt.blogspot.com/2023/04/fundamental-question-what-determines.html. First completed April 9, 2023.]
A mind has some effects on the world. What determines which effects a mind has? To eventually create minds that have large ... |
7606c73a-a76b-4f6f-82c5-aeb75efa7067 | trentmkelly/LessWrong-43k | LessWrong | Linkpost: Robin Hanson - Why Not Wait On AI Risk?
Raises an interesting critique of AI Risk. I think he is wrong, but would be interested in how others would respond. |
8104ed0d-a5e5-4d40-9cfa-da30205f9965 | StampyAI/alignment-research-dataset/blogs | Blogs | Effect of marginal hardware on artificial general intelligence
We do not know how AGI will scale with marginal hardware. Several sources of evidence may shed light on this question.
Details
-------
### Background
Suppose that at some point in the future, general artificial intelligence can be run on some quantit... |
91933242-b518-4794-9ada-56348239c9c7 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Imagine a world where Microsoft employees used Bing
My hypothesis that [GPT-4 already is agentic and is trying to take over the world!](https://www.lesswrong.com/posts/83cXhEy2X8zkucjcs/a-crazy-hypothesis-gpt-4-already-is-agentic-and-is-trying-to) has been met with "skepticism" to put it lightly. I think part of this ... |
baaaac1f-f357-4eb7-882e-5ebf28bedfd9 | trentmkelly/LessWrong-43k | LessWrong | Has anyone had weird experiences with Alcor?
After years of procrastinating, I've finally taken steps towards signing up for cryonics, thanks in no small part to things I've read here on LessWrong. I view it as a very long shot, but worth a try nonetheless.
However, I'm getting some very weird signals from the Alcor... |
49691589-bea9-4f66-82e6-9b4b6d209e0d | trentmkelly/LessWrong-43k | LessWrong | Building Toe Buttons
In controlling my rhythm stage setup I have been bringing instruments in and out by pressing buttons on my axis 49. It has plenty of buttons, but needing to press them with my fingers doesn't combine well with playing mandolin. So: toe buttons!
The most common tool musicians use seems to be the F... |
2d746db2-9214-4b54-816a-f01f87d49f6b | trentmkelly/LessWrong-43k | LessWrong | AI Forecasting Benchmark: Congratulations to Q4 Winners + Q1 Practice Questions Open
Two quarters down and two quarters remain for our AI Forecasting Benchmark, which aims to assess how the best bots compare to the best humans on real-world forecasting questions.
In this post:
* Congratulations to the Q4 winners: T... |
0add4a51-d925-4c96-9737-ef833ccabf3f | trentmkelly/LessWrong-43k | LessWrong | How much can value learning be disentangled?
In the context of whether the definition of human values can disentangled from the process of approximating/implementing that definition, David asks me:
* But I think it's reasonable to assume (within the bounds of a discussion) that there is a non-terrible way (in princi... |
ffb6d2ff-667d-4425-92d3-ecc5056527d7 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Cambridge, MA Sunday meetup: The Contrarian Positions Game
Discussion article for the meetup : Cambridge, MA Sunday meetup: The Contrarian Positions Game
WHEN: 15 November 2015 03:30:00PM (-0500)
WHERE: 98 Elm St Somerville, MA 02144
This meetup is about pulling ropes sideways. We'll be practicing the ment... |
0bc3e11e-5f4a-4e54-9ef8-de87ade95f80 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] No Safe Defense, Not Even Science
Today's post, No Safe Defense, Not Even Science was originally published on 18 May 2008. A summary (taken from the LW wiki):
> Why am I trying to break your trust in Science? Because you can't think and trust at the same time. The social rules of Science are verbal rat... |
b3506d9e-bf74-4604-9ab9-4d1a6b56f50a | trentmkelly/LessWrong-43k | LessWrong | In defense of Oracle ("Tool") AI research
(Update 2022: Enjoy the post, but note that it’s old, has some errors, and is certainly not reflective of my current thinking. –Steve)
Low confidence; offering this up for discussion
An Oracle AI is an AI that only answers questions, and doesn't take any other actions. The o... |
98bebc2e-47ff-4df7-ae93-38052e4a7760 | trentmkelly/LessWrong-43k | LessWrong | Soft Skills for Running Meetups for Beginners
Having a vibrant, local Less Wrong or EA community is really valuable, but at least in my experience, it tends to end up falling to the same couple people, and if one of or all of those people get burned out or get interested in another project, the meetup can fall apart. ... |
2be0ec09-5cef-454e-9c92-6fdefaf662f0 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Why I'm excited about Redwood Research's current project
[Redwood Research's current project](https://www.alignmentforum.org/posts/k7oxdbNaGATZbtEg3/redwood-research-s-current-project) is to train a model that completes short snippets of fiction without outputting text where someone gets injured. I'm excited about thi... |
a6fa3b3b-3823-453f-8dcc-41da3a1372fb | trentmkelly/LessWrong-43k | LessWrong | The Cluster Structure of Thingspace
The notion of a "configuration space" is a way of translating object descriptions into object positions. It may seem like blue is "closer" to blue-green than to red, but how much closer? It's hard to answer that question by just staring at the colors. But it helps to know that th... |
5dc31f6c-8602-48f7-b976-ddd7b626fec9 | trentmkelly/LessWrong-43k | LessWrong | Social media probably not a deathtrap
I'm underwhelmed and unconvinced by the recent burst of content that has strong flavors of “social media is a deathtrap, millennials are consumed in a never-ending race to signal status, and technology devouring people’s lives is a huge problem.” I think it's getting at something,... |
bc7a945d-2296-413d-a02f-20a4733e75c8 | trentmkelly/LessWrong-43k | LessWrong | Introspection
Imagine your mind as a sophisticated software program, with each line of code representing your experiences, beliefs, and emotions. Just as a programmer reviews and refines code to eliminate bugs and enhance performance, introspection allows us to debug and optimize our mental and emotional frameworks.
... |
80c2c6cc-b12a-4272-b31e-721837d75c8c | trentmkelly/LessWrong-43k | LessWrong | Why the AI Alignment Problem
Might be Unsolvable?
Author's note 1:
The following is a chapter from the story I've been writing which contains, well, it contains what I think is probably a proof that the value alignment problem is unsolvable. I know it sounds crazy, but as far as I can tell the proof seems to be corre... |
e1862615-45bc-4b48-8be4-f89f1b6780f0 | trentmkelly/LessWrong-43k | LessWrong | What comes after?
The brain is wider than the sky,
For, put them side by side,
The one the other will contain,
With ease, and you, beside.
— Emily Dickinson
Kevin Kelly asks, “In the story of biological or technological evolution—what comes after minds?”
> The usual response to “what comes after a human mind” is b... |
c7ce3622-4665-4e4e-b6c4-ef265592d0f5 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Relational Deep Reinforcement Learning
1 Introduction
---------------
Recent advances in deep reinforcement learning (deep RL) [[1](#bib.bib1), [2](#bib.bib2), [3](#bib.bib3)] are in part driven by a capacity to learn good internal representations to inform an agent’s policy. Unfortunately, deep RL models still fac... |
bb8ca5f0-bb5b-497f-a5f8-6b83a23c40a1 | trentmkelly/LessWrong-43k | LessWrong | Solar One Year In
Last fall we had solar panels installed. Our roof is pretty marginal for solar, large parts blocked by trees and the remainder mostly facing West, but incentives were high enough that it looked decent. And even if it only broke even I still liked it for the resiliency advantages. It's currently doing... |
392c72f9-89c6-47dd-b7d2-a1307fd29a85 | trentmkelly/LessWrong-43k | LessWrong | Cultivating a state of mind where new ideas are born
In the early 2010s, a popular idea was to provide coworking spaces and shared living to people who were building startups. That way the founders would have a thriving social scene of peers to percolate ideas with as they figured out how to build and scale a venture.... |
f31f6e5d-bbc7-49cf-8601-73e0838367cf | StampyAI/alignment-research-dataset/arxiv | Arxiv | Why and How Governments Should Monitor AI Development
Executive summary
-----------------
We outline a proposal for improving the governance of artificial intelligence (AI) by investing in government capacity to systematically measure and monitor the capabilities and impacts of AI systems.
The problem. Governme... |
cd983f7f-f7f9-4214-a7fe-b731357c698d | trentmkelly/LessWrong-43k | LessWrong | Group Rationality Diary, December 1-15
This is the public group rationality diary for December 1-15.
> 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 belief
> * ... |
fa1b8a6f-65c5-458c-b5a7-919b8c03af5c | trentmkelly/LessWrong-43k | LessWrong | Dialogue introduction to Singular Learning Theory
Alice: A lot of people are talking about Singular Learning Theory. Do you know what it is?
Bob: I do. (pause) Kind of.
Alice: Well, I don't. Explanation time?
Bob: Uh, I'm not really an expert on it. You know, there's a lot of materials out there that--
Alice: that... |
a3e675d4-985f-40b8-be51-2a68a7074d8e | StampyAI/alignment-research-dataset/arxiv | Arxiv | There Is No Turning Back: A Self-Supervised Approach for Reversibility-Aware Reinforcement Learning
1 Introduction
---------------
We address the problem of estimating if and how easily actions can be reversed in the Reinforcement Learning (RL) context.
Irreversible outcomes are often not to be taken lightly when m... |
0f9bd453-aa95-4eef-899c-f215011fbf75 | trentmkelly/LessWrong-43k | LessWrong | Paper Trauma
Andrew Critch thinks people should be spending more time than they currently are using paper as a working memory aid while thinking, especially large paper (for even more working memory). It is really astonishing how helpful this can be. We consistently encourage people to do it at CFAR workshops nearly e... |
4a9249e9-c8f2-42f0-a4b9-7fd48b581534 | trentmkelly/LessWrong-43k | LessWrong | Open Thread April 2019
If it’s worth saying, but not worth its own post, you can put it here.
Also, if you are new to LessWrong and want to introduce yourself, this is the place to do it. Personal stories, anecdotes, or just general comments on how you found us and what you hope to get from the site and community are... |
d592ff5e-ce79-42df-895a-be3edbe575e9 | trentmkelly/LessWrong-43k | LessWrong | Unsupervised Activation Steering: Find a steering vector that best represents any set of text data
Epistemic Status: I thought of this yesterday and it seems plausible, but my initial attempts couldn't get it to work (as best as I could tell) and I don't have more time to commit to this. Anyone can feel free to try it... |
1fe43f68-aaca-49b4-b81c-7c7ddb86e00f | trentmkelly/LessWrong-43k | LessWrong | Non-loss of control AGI-related catastrophes are out of control too
For the Open Philanthropy AI Worldview Contest
Executive Summary
* Conditional on AGI being developed by 2070, we estimate the probability that humanity will suffer an existential catastrophe due to loss of control (LoC) over an AGI system at ~6%, ... |
97ca9267-070a-4860-9b62-fbe766ab3fd3 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Could We Automate AI Alignment Research?
Summary
=======
* Creating an AI that can do AI alignment research (an automated alignment researcher) is one part of OpenAI’s three-part alignment [plan](https://openai.com/blog/our-approach-to-alignment-research) and a core goal of their Superalignment [team](https://openai.... |
957b4f1f-38af-46c9-b551-712026296018 | trentmkelly/LessWrong-43k | LessWrong | Excessive AI growth-rate yields little socio-economic benefit.
Basic argument
In short —
1. There exists an adaptive limit representing the maximum rate at the economy can adapt to new technology.
2. Exceeding this limit yields no socio-economic benefit, in the general case.
3. Exceeding this limit imposes signifi... |
f3af9ea5-d334-4e70-9738-1d84c829d868 | trentmkelly/LessWrong-43k | LessWrong | An early stage project generation model
Here is how I see many people connected to Effective Altruism making their life plan decisions:
> Well, I don't know what the most important problem is. For example, reducing factory farming seems important, as does alleviating global poverty, but then certain global catastroph... |
13f50c5a-8a7d-4e77-8ccb-6272ee5ca345 | trentmkelly/LessWrong-43k | LessWrong | [April Fools' Day] Introducing Open Asteroid Impact
> “That which does not kill us makes us stronger.”
Hillary Clinton, who is still alive
I'm proud and excited to announce the founding of my new startup, Open Asteroid Impact, where we redirect asteroids towards Earth for the benefit of humanity. Our mission is t... |
9b2c63b5-faac-442b-8d45-3d8f89432229 | StampyAI/alignment-research-dataset/special_docs | Other | The pursuit of happiness: A reinforcement learning perspective on habituation and comparisons.
ZK[KF ZIN FZ\OIS K
\so|u~€utt {qsk||tzo€€> F~otzq{~moyozt
wok~ztzr |o~€|omttvo {zsklttuktt{z kzn
m{y|k~t€{z€
Zkmst� J�lo� OJ
4*.\s{yk� S1M~tqqt�s�5.Wo�o~ Jk�kz OJ
6.7
4Jo|k~�yoz� {qI{y|��o~ [mtozmo. W~tzmo�{z ]zt�o~�t�� .W~t... |
fe828481-5b5c-4fa0-b160-0c1666f0923f | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Podcast Transcript: Daniela and Dario Amodei on Anthropic
*The last transcript we posted on Eliezer's podcast got a lot of traction, so I'm following up with another conversation we find very important. Dario and Daniela of Anthropic have not talked much publicly about their views on alignment, but they have done a lo... |
d7b63ef0-d017-49ad-a8eb-6d2a2428e627 | trentmkelly/LessWrong-43k | LessWrong | Bucket Brigade: Likely End-of-Life
During the pandemic a group of us wrote some software to allow people to sing and make music together in real time over the internet, by accounting for latency instead of minimizing it. The initial goal was to support group singing at the 2020 Secular Solstice, after which I reworked... |
9049924b-c1b2-4aa7-a155-18d53d4dc5ca | trentmkelly/LessWrong-43k | LessWrong | You Are A Brain - Intro to LW/Rationality Concepts [Video & Slides]
Here's a 32-minute presentation I made to provide an introduction to some of the core LessWrong concepts for a general audience:
You Are A Brain [YouTube]
You Are a Brain [Google Slides] - public domain
I already posted this here in 2009 and some ... |
580d135d-f5f9-4526-a040-bc1872531c67 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Goal Misgeneralization in Deep Reinforcement Learning
1 Introduction
---------------
Out of distribution (OOD) robustness, performing well on test data that is not distributed identically to the training set, is a fundamental problem in machine learning [[5](#bib.bib12 "Out of distribution generalization in machine... |
a4fe70af-1909-4ad2-ad9e-fc0bb02d738b | trentmkelly/LessWrong-43k | LessWrong | Spreading the word?
This has been discussed some, but I don't think it's been the sole subject of a top-level post. I want to find out other people's ideas rather than driving the discussion into my ideas, so I'm asking the question in a very general form, and holding off on my own answers:
* Should we be trying to ... |
5df5ec6f-c397-4532-8d1d-16123569a662 | trentmkelly/LessWrong-43k | LessWrong | When people say robots will steal jobs, what kinds of jobs are never implied?
Like priests, for example. It's simply not supposed. What else? |
6d8f03c3-67c3-4f00-b595-b2171761d14d | trentmkelly/LessWrong-43k | LessWrong | Believe in Yourself and don't stop Improving
Don't let fear of not being good enough, slow you down or stop you.
You need to be confident in your ability to do X. If you are anxious about being able to do X, if it feels uncertain, it will be hard to fully commit yourself to do everything that you need in order to do ... |
d9de4951-5fc6-4b9f-be9f-ab697d83d8fc | trentmkelly/LessWrong-43k | LessWrong | Interference Issues
I've been working on building an electronic harp mandolin ( previously, previously, previously). I got it all hooked up, but unfortunately I'm running into pretty bad interference, and need to redesign it.
Going back a bit, I had designed a circuit for testing and breadboarded it. It worked very w... |
dd1cfaf2-8598-4a17-a123-5db795baece0 | trentmkelly/LessWrong-43k | LessWrong | Curiosity: Why did you mega-downvote "AI is Software" ?
I've never experienced anything like it before on LessWrong. Would you care to re-read the post, and offer your feedback? |
e68f4009-a0db-41e3-bebc-0c39f86fbdd0 | trentmkelly/LessWrong-43k | LessWrong | Experience on Methotrexate
Several months ago I wrote about how my wrists/hands were still not doing well, and about how I was going to be trying Methotrexate. I have since gone off it, initially for a covid shot, but I'm now considering staying off it.
After meeting again with my rheumatologist on 2021-06-28 I start... |
a158dbd4-d1a5-4cee-90c0-4cb68a04e195 | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: 30fb1b5148e20b47
Meta-level adversarial evaluation of oversight techniques might allow robust measurement of their adequacy — LessWrong This website requires javascript to properly function. Consider activating javascript to get access to all site functionality. LESSWRONG LW Login Redwood Research AI Frontpage 100... |
ed5dc05a-7e84-4e9f-94cd-d46cbeec26dc | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Take 2: Building tools to help build FAI is a legitimate strategy, but it's dual-use.
*As a writing exercise, I'm writing an AI Alignment Hot Take Advent Calendar - one new hot take, written every day for 25 days. Or until I run out of hot takes, which seems likely.*
This was waiting around in the middle of my hot-ta... |
a1d477a7-6538-4160-be96-c7b37f8a4f69 | trentmkelly/LessWrong-43k | LessWrong | "In the Dust of This Planet," by Eugene Thacker
Salticidae Philosophiae is a series of abstracts, commentaries, and reviews on philosophical articles and books.
Eugene Thacker suggests that we look to the genre of horror as offering a way of thinking about the unthinkable world. To confront this idea is to confront t... |
a1321d6e-7624-4149-82ab-839ba1cd50e7 | trentmkelly/LessWrong-43k | LessWrong | What is the best AI generated music about rationality/ai/transhumanism?
Basically just make some and then lets vote on it.
I personally am not worried about current music generation tech causing harm (and probably think that it's healthy to appreciate current tech isn't that dangerous so we can notice when we stop t... |
1b3c6b80-378f-40cb-a38d-538a47177c63 | trentmkelly/LessWrong-43k | LessWrong | Simple versus Short: Higher-order degeneracy and error-correction
> TLDR: The simplicity bias in Bayesian statistics is not just a bias towards short description length.
The folklore relating the simplicity bias in Bayesian statistics to description length is incomplete: while it is true that the fewer parameters... |
0b6c699c-7810-4f9b-b020-b839811b36e9 | trentmkelly/LessWrong-43k | LessWrong | For the Sake of Pleasure Alone
Content Warning: existential crisis, total hedonistic utilitarianism, timeless worldview, potential AI-related heresies.
Hi, first post here. I’m not a native speaker, but I think it’s fine. I suffer from the illusion of transparency, yet if I delve into every detail of my reasoning, it... |
d42b0567-b32a-40fa-b417-d4babe6093d7 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | The History of AI Rights Research
Explanation for the Forum
=========================
I previously created a formal literature review of research on "[The Moral Consideration of Artificial Entities](https://link.springer.com/article/10.1007/s11948-021-00331-8)" and a brief, informal primer on "[The Importance of Arti... |
987b5077-8817-466c-b4ef-78b2bec28957 | trentmkelly/LessWrong-43k | LessWrong | Should people require a mandatory license for parenting?
Sir, Could I See Your Breeding License?
> Why [...] are we so cavalier about who we let have and raise them? As technology enables more people to reproduce, environmental pressures make each new life a bigger burden, and our understanding of child psychology im... |
f1d71adc-4d98-4610-ba10-3936153bd743 | trentmkelly/LessWrong-43k | LessWrong | In the context of AI interp. What is a feature exactly?
As I read more about previous interpretability work, I've noticed this trend that implicitly defines a feature in this weird human centric way. It's this weird prior that expects networks to automatically generate features that correspond with how we process imag... |
0d004c64-e298-431d-854e-7f4ea4317fc8 | trentmkelly/LessWrong-43k | LessWrong | AI Safety Newsletter #2: ChaosGPT, Natural Selection, and AI Safety in the Media
Welcome to the AI Safety Newsletter by the Center for AI Safety. We discuss developments in AI and AI safety. No technical background required.
Subscribe here to receive future versions.
---
ChaosGPT and the Rise of Language Agents
Ch... |
bf28710d-224c-4865-90ea-9d224aca0091 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | [AN #105]: The economic trajectory of humanity, and what we might mean by optimization
Alignment Newsletter is a weekly publication with recent content relevant to AI alignment around the world. Find all Alignment Newsletter **[resources here](http://rohinshah.com/alignment-newsletter/)**. In particular, you can look ... |
ae743f3d-68d3-400b-95ba-9f4f80876afe | StampyAI/alignment-research-dataset/lesswrong | LessWrong | More experiments in GPT-4 agency: writing memos
More experiments on [A crazy hypothesis: GPT-4 already is agentic and is trying to take over the world!](https://www.lesswrong.com/posts/83cXhEy2X8zkucjcs/a-crazy-hypothesis-gpt-4-already-is-agentic-and-is-trying-to). A follow up to [Does GPT-4 exhibit agency when summar... |
8ded2413-37f4-4fd7-a3c6-d32dd775fd43 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | DALL-E by OpenAI
This is a linkpost for <https://openai.com/blog/dall-e/>
My own take: Cool, not super surprising given GPT-3 and Image GPT. I look forward to seeing what a bigger version of this would do, so that we could get a sense of how much it improves with scale. I'm especially interested in the raven's progre... |
7ff87b3d-5579-4871-87ce-04f3584d664d | trentmkelly/LessWrong-43k | LessWrong | Why sigmoids are so hard to predict
tl;dr: Predicting the behaviour of a sigmoid-like process is not fitting the parameters of a logistic curve. Instead, it's trying to estimate the strength of the dampening term - a term that might be actually invisible in the initial data.
A logistic curve is a function that starts... |
c8b3ea7e-7842-4bad-b642-e7a71e41fedf | trentmkelly/LessWrong-43k | LessWrong | Arguments for the existence of qualia
One trap that we must be wary of is adopting beliefs because they are popular among people who strive to think critically or scientifically, as opposed to being the result of critical or scientific thinking. One good example of this is RationalWiki which purports to report the pol... |
42188878-b304-4867-9223-183c0b636589 | StampyAI/alignment-research-dataset/aisafety.info | AI Safety Info | What are the leading theories in moral philosophy and which of them might be technically the easiest to encode into an AI?
There are three major approaches to normative ethics (and some approaches to unify two or all of them): Virtue ethics, deontological ethics, and consequentialist ethics.
[Virtue ethicists](https:... |
7ad7fde8-dc04-4805-834a-1b399652aecb | StampyAI/alignment-research-dataset/blogs | Blogs | Incentives to create AI systems known to pose extinction risks
*Published 6 Aug 2022*
Economic incentives to deploy AI systems seem unlikely to be reliably eliminated by knowledge that those AI systems pose an existential risk.
Details
-------
### Reasons for people with normal values to be incentivized to brin... |
2abbe60d-27ee-46ad-aae6-30a475a8c774 | trentmkelly/LessWrong-43k | LessWrong | What are the greatest near-future risks or dangers to you as an individual?
What factors do you expect have the highest likelihood of severely compromising your own quality and/or duration of life, within the next 1, 5, or 10 years? How do these risks change your behavior compared to how you expect you'd act if they w... |
8fbcdd35-ab2e-4944-9115-cb348d85032c | trentmkelly/LessWrong-43k | LessWrong | SAE Probing: What is it good for?
Subhash and Josh are co-first authors. Work done as part of the two week research sprint in Neel Nanda’s MATS stream
Update February 2025:
We have recently expanded this post into a full paper: https://arxiv.org/abs/2502.16681
Our results are now substantially more negative. We fi... |
08536db1-3f18-4104-9ec7-5bcd6c056dca | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Seriously, what goes wrong with "reward the agent when it makes you smile"?
Suppose you're training a huge neural network with some awesome future RL algorithm with clever exploration bonuses and a self-supervised pretrained multimodal initialization and a recurrent state. This NN implements an embodied agent which ta... |
2bcd98d4-99bd-45c9-8fb4-aae95aa2c634 | StampyAI/alignment-research-dataset/special_docs | Other | Modeling the Human Trajectory
Published: June 15, 2020 | by [David Roodman](/about/team/david-roodman)
In arriving at our [funding priorities](https://www.openphilanthropy.org/focus/)—including criminal justice reform, farm animal welfare, pandem... |
00487f90-e829-49e5-91fc-338bb63f61fc | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post221
This is the abstract and introduction of our new paper , with some discussion of implications for AI Safety at the end. Authors: Jan Betley*, Xuchan Bao*, Martín Soto*, Anna Sztyber-Betley, James Chua, Owain Evans (*Equal Contribution). Abstract We study behavioral self-awareness — an LLM's ability to arti... |
114de6e6-a0c1-4d64-a386-93317b6ca2fc | trentmkelly/LessWrong-43k | LessWrong | Translating bad advice
While writing my Magnum Opus I came across this piece of writing advice by Neil Gaiman:
“When people tell you something’s wrong or doesn’t work for them, they are almost always right. When they tell you exactly what they think is wrong and how to fix it, they are almost always wrong.”
And it s... |
2fcf4d0b-fe96-4e1d-b255-ce70692aec4c | trentmkelly/LessWrong-43k | LessWrong | Do You Have an Internal Monologue?
I find the differences in the way people think fascinating, to the point where I went around asking people (n~=40) whether they thought verbally or otherwise when I was in high school. I got about 50-50 responses, and proceeded to try to train myself to think conceptually (from a bas... |
4df923b9-2d3e-41e0-b842-6afa0ac5e0e6 | trentmkelly/LessWrong-43k | LessWrong | Who's likely to write the AI? A hypothesis
The idea of creating an AI seems to be getting more common.
To the extent that creating an AI is made easier by having more resources rather than by having more carefully thought out philosophy, the first AI will be created by a government or a business, not SIAI. I think th... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.