id
stringlengths
36
36
source
stringclasses
15 values
formatted_source
stringclasses
13 values
text
stringlengths
2
7.55M
88a7ce5c-8e22-47ac-88a6-cba0ffde3bdd
trentmkelly/LessWrong-43k
LessWrong
Reality Testing   In the recent weeks since Biden’s disastrous debate performance, pundits have theatrically expressed surprise at his apparent cognitive decline, despite years of clear and available evidence in the form of video clips, detailed reports, and fumbled public appearances. This is far from the first pop...
40588708-3420-4f95-91b7-7cc09743bdca
trentmkelly/LessWrong-43k
LessWrong
Covid 2/24/22: The Next War Russia has invaded Ukraine with the stated goal of 'demilitarization and denazification.' On its face, this has nothing to do with Covid. The invasion may reflect a perception, either on the part of Putin or otherwise, from our Covid-19 response or otherwise, that the West has shown itself...
406e7974-4364-4d56-b396-abec0a526d87
trentmkelly/LessWrong-43k
LessWrong
Reverse Auctions for Group Decision-Making I recently faced a dilemma with a group of friends in which auction theory provided a nice solution to ensure everyone's satisfaction. Due to its elegance and potential application to similar coordination problems, I detail the solution below as well as some intuition for why...
fa486fef-611e-4628-a1b9-6e0b772020fa
trentmkelly/LessWrong-43k
LessWrong
Meta AI announces Cicero: Human-Level Diplomacy play (with dialogue) Abstract Despite much progress in training AI systems to imitate human language, building agents that use language to communicate intentionally with humans in interactive environments remains a major challenge. We introduce Cicero, the first AI agent...
a25043ce-1791-4632-aa19-98bcff222ea7
trentmkelly/LessWrong-43k
LessWrong
The Philosophical Glossary of AI > It is hard to know what to make of claims like “LLMs are intelligent”, “we have reached AGI”, or “AI’s outputs are biased” without a grasp of the definitions of the terms ‘intelligent’, ‘AGI’, and ‘bias’. And yet, many do just this. Interdisciplinary debate would be easier and more f...
88ee910b-672a-4888-ac2d-858508923f42
trentmkelly/LessWrong-43k
LessWrong
A Medical Mystery: Thyroid Hormones, Chronic Fatigue and Fibromyalgia Summary: * Chronic Fatigue and Fibromyalgia look very like Hypothyroidism. * Thyroid Patients aren't happy with either the diagnosis or treatment of Hypothyroidism. * It is possible that lots of FMS/CFS cases are 'something wrong with the thyroi...
69ff4d96-d59e-4ce1-81c8-507de3359a6a
StampyAI/alignment-research-dataset/blogs
Blogs
my current outlook on AI risk mitigation *(thanks to [Linda Linsefors](https://www.lesswrong.com/users/linda-linsefors), [Artaxerxes](https://www.lesswrong.com/users/artaxerxes), and [jono](https://www.lesswrong.com/users/lw-user0246) for their feedback on drafts of this post.)* my current outlook on AI risk mitigat...
f5b5a871-846d-4a38-9c28-06cf7711075d
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Forecasting Transformative AI: What Kind of AI? *Audio version available [at Cold Takes](https://www.cold-takes.com/transformative-ai-timelines-part-1-of-4-what-kind-of-ai/) (or search Stitcher, Spotify, Google Podcasts, etc. for "Cold Takes Audio")* [![](https://www.cold-takes.com/content/images/2021/08/mic-diagram--...
ed13195c-4f2f-46a3-a463-0f770a8f97c3
StampyAI/alignment-research-dataset/agisf
AGI Safety Fund
Frontier AI Regulation:Managing Emerging Risks to Public Safety Executive Summary ----------------- The capabilities of today’s foundation models highlight both the promise and risks of rapid advances in AI. These models have demonstrated significant potential to benefit people in a wide range of fields, including ...
ab9fbf64-f652-4695-a411-06905a811e88
trentmkelly/LessWrong-43k
LessWrong
Build a Causal Decision Theorist I'll argue here that we should make an aligned AI which is a causal decision theorist. Son-of-CDT Suppose we are writing code for an agent with an action space A and an observation space O. The code determines how actions will be selected given the prior history of actions and observ...
e2ec41af-ffd5-439e-83d2-c0aa3f031b97
trentmkelly/LessWrong-43k
LessWrong
The Archetypal Rational and Post-Rational There's been a lot of discussion about the difference between rationality and post-rationality. Gordon Worley sees the difference being that post-rationality is based upon epistemic circularity, but Eliezer proposed that too. David Chapman defines it in terms of being meta-sys...
5152965c-cba9-4bb8-8aa2-0c99d438ab15
trentmkelly/LessWrong-43k
LessWrong
[LDSL#4] Root cause analysis versus effect size estimation Followup to: Information-orientation is in tension with magnitude-orientation. This post is also available on my Substack. In the conventional theory of causal inference, such as Rubin’s potential outcomes model or Pearl’s DAG approach, causality is modelled ...
1c6853c6-2586-45ff-bc53-979c3d6b45bd
trentmkelly/LessWrong-43k
LessWrong
Can AI systems have extremely impressive outputs and also not need to be aligned because they aren't general enough or something? Layperson question here; I appreciate any answers/comments that might help me understand this topic better. When thinking about DALL·E 2 and Google's PaLM this week I wondered if in the fu...
862e0260-384f-475f-bb2f-988f10320b8d
trentmkelly/LessWrong-43k
LessWrong
Is AI Physical? Context: This is part of a series of posts I am writing with Dmitry Vaintrob, as we aim to unpack some potential value from Quantum Field Theory (QFT). Consider this post as framing why physics and its frameworks can be good for building a science of AI.  Introduction In Position: Is machine learnin...
b9bc0e91-db5d-47b6-add3-e29b56b88e87
trentmkelly/LessWrong-43k
LessWrong
Talking to yourself: A useful thinking tool that seems understudied and underdiscussed I have returned from a particularly fruitful Google search, with unexpected results. My question was simple. I was pretty sure that talking to myself aloud makes me temporarily better at solving problems that need a lot of working ...
2fe82fd3-3015-4cd2-8ef0-497b7cffbbe4
trentmkelly/LessWrong-43k
LessWrong
On Arguments for God This post is about God, but of course, it isn't really about God, but about a particular pattern in general. We're pretty much all in agreement that God doesn't exist. This is correct or at least let's say it is for the purposes of this post. However, this also poses a trap.  Suppose there are ...
8daf5962-8c64-47e1-8839-1ec2af825368
trentmkelly/LessWrong-43k
LessWrong
Student competition for drafting a treaty on moratorium of large-scale AI capabilities R&D Campaign for AI safety has announced a competition for the drafting of an international treaty on moratorium of large-scale AI capabilities research and development. The competition is open to all students of law, philosophy, a...
d7abc8d7-ce5d-4fbe-b53a-e39599b83343
trentmkelly/LessWrong-43k
LessWrong
An Attempt at Logical Uncertainty At the recent Tel Aviv meetup, after a discussion of the open problems in the field of FAI, we reached the conclusion that the problem of logical uncertainty is one of the most major of the problems open today. In this post I will try to give a few insights I had on this problem, whic...
94a6a33a-b17c-4bce-a405-8811003633fe
trentmkelly/LessWrong-43k
LessWrong
A strong mind continues its trajectory of creativity [Metadata: crossposted from https://tsvibt.blogspot.com/2023/01/a-strong-mind-continues-its-trajectory.html. First completed January 29, 2023.] A very strong mind is produced by a trajectory of creativity. A trajectory of creativity that produces a very strong mind...
165fd3b1-c7b0-4727-bd09-6712c4716a7c
StampyAI/alignment-research-dataset/blogs
Blogs
New report: “Loudness: On priors over preference relations” [![loudness first page](https://intelligence.org/wp-content/uploads/2014/05/loudness-first-page.png)](https://intelligence.org/files/LoudnessPriors.pdf)Today we release the first technical report from our [May 2014 workshop](http://intelligence.org/workshops/...
33d7e8aa-7fb9-4513-92e1-c93915588aa3
trentmkelly/LessWrong-43k
LessWrong
Most capable publicly available agents? Looking to do a little compare and contrast.
dbdb0733-365a-48b1-81f7-f291cd833a81
StampyAI/alignment-research-dataset/arxiv
Arxiv
Representation Learning: A Review and New Perspectives 1 Introduction --------------- The performance of machine learning methods is heavily dependent on the choice of data representation (or features) on which they are applied. For that reason, much of the actual effort in deploying machine learning algorithms goe...
c36ac6b6-0f1a-4d32-86e1-5f8b634c2dda
trentmkelly/LessWrong-43k
LessWrong
MLU: New Blog! I'm in the process of moving mindlevelup from Wordpress to a new static site hosted by Netlify. I like this because now I have more control over scripts and the visuals. It also finally puts into place my goal of having a place to have short/longform posts. The plan is to slowly update posts and increm...
4c1e7904-2eee-4d79-a991-4624309f4ce0
trentmkelly/LessWrong-43k
LessWrong
What evidence will tell us about the new strain? How are you updating? I didn't get into this question in the weekly post. The answer does not seem obvious to me. Eventually, it will be obvious either way. Either the strain will spread rapidly across many locations, or it won't. But we would like to get the informati...
349ce1b3-f200-426f-b618-70feda385e6f
trentmkelly/LessWrong-43k
LessWrong
Spending on Ourselves The Effective Altruism community has encouraged a range of different approaches to doing good over time. Initially there was more focus on frugality as a way to increase how much you could donate, which was mostly supplanted by emphasis on earning more. In late 2015 this started to shift towards ...
6da39956-7c3d-4344-9705-02cf3b7e94d1
trentmkelly/LessWrong-43k
LessWrong
A poem titled 'Tick Tock'. Inspired by a collective intelligence project that I've been working on in the GOFAI space for over a decade.  Hoping to share more at less online if I can afford to make it.   1st prediction: My second prediction will be true. 2nd prediction: My first prediction was false.   42nd pre...
f00c07f4-0c73-40c5-ab80-710c24a2dd41
trentmkelly/LessWrong-43k
LessWrong
Emergent scaling effects on the functional hierarchies within LLMs I have been poking around with LLMs, and I found some results that seem broadly interesting Summary Introduction: Large language models (LLM) are usually structured as repeated transformer layers of the same size. However, this architecture is often ...
e0f49f62-7d0f-4ed6-962f-b60ebc2c63fe
StampyAI/alignment-research-dataset/aisafety.info
AI Safety Info
What is AI safety? "AI safety" refers to efforts to prevent artificial intelligence from causing harm. This site focuses on existential disasters in which powerful future AI systems pursue goals in conflict with humanity’s values, because these systems may end up taking control of the world or even driving humanity ex...
89d38151-e68c-48c8-aa96-dbcae6675969
trentmkelly/LessWrong-43k
LessWrong
Why I expect successful (narrow) alignment Summary I believe that advanced AI systems will likely be aligned with the goals of their human operators, at least in a narrow sense. I’ll give three main reasons for this: 1. The transition to AI may happen in a way that does not give rise to the alignment problem as it’s...
aaaea6f6-2e96-4c56-ba98-3675d94716aa
trentmkelly/LessWrong-43k
LessWrong
The two-layer model of human values, and problems with synthesizing preferences I have been thinking about Stuart Armstrong's preference synthesis research agenda, and have long had the feeling that there's something off about the way that it is currently framed. In the post I try to describe why. I start by describin...
56a6f96d-72d7-4273-8b53-c0a006faaf8d
trentmkelly/LessWrong-43k
LessWrong
Counterfactual Mugging v. Subjective Probability This has been in my drafts folder for ages, but in light of Eliezer's post yesterday, I thought I'd see if I could get some comment on it:   A couple weeks ago, Vladimir Nesov stirred up the biggest hornet's nest I've ever seen on LW by introducing us to the Counterfa...
e2d84db4-13d4-4e66-aba7-d0e0b66cb5e2
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Unifying Bargaining Notions (2/2) Alright, time for the payoff, unifying everything discussed in the [previous post](https://www.lesswrong.com/posts/rYDas2DDGGDRc8gGB/unifying-bargaining-notions-1-2). This post is a lot more mathematically dense, you might want to digest it in more than one sitting.   **Imaginary Pr...
43f7fbc6-9dc4-47e2-a5c3-4f412b78b441
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Does most of your impact come from what you do soon? Over the last couple months I’ve noticed myself flipping back and forth between two mindsets: “I should try to be useful soon” and “I should build skills so that I am more useful in 5+ years.” I’ve compiled arguments for each view in this post. Note that a lot of th...
dcb8efc5-e64f-45b1-a038-8d1e279817fb
StampyAI/alignment-research-dataset/blogs
Blogs
Language models are 0-shot interpreters --- Table of Contents* [Overview](#overview) * [0-shot, few-shot, and meta-learning](#0-shot-few-shot-and-meta-learning) + [0-shot can outperform few-shot](#0-shot-can-outperform-few-shot) * [Measuring prompt helpfulness](#measuring-prompt-helpfulness) + [Decibels of evidence...
3e2b6df5-c02b-4c8f-88f0-fe2daf8f1e9c
trentmkelly/LessWrong-43k
LessWrong
Land with no aunties Posted a piece on how removing informal social control (the “auntie layer”) affects a city’s memetic landscape, using SF as case study. Interested in rationalist critiques: Are such regulators net-positive or pareto-inefficient?
cd82fa2c-7faa-4b52-9418-3a4c4de95d0d
trentmkelly/LessWrong-43k
LessWrong
The Comprehension Curve Tl;Dr: You might comprehend and learn more quickly by slowing down your reading, rather than by speeding it up. This post is mainly about laying down some terminology that I think is conceptually clarifying in thinking about learning, comprehension and reading speed. Speed Reading According t...
80351e35-ec65-4646-b224-fac2c7fa22dc
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Self-improvement without self-modification This is just a short note to point out that AIs can self-improve without having to self-modify. So locking down an agent from self-modification is not an effective safety measure. How could AIs do that? The easiest and the most trivial is to create a subagent, and transfer ...
95b1f461-3aaa-4da0-8f9b-c951cd4b2356
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
[Intro to brain-like-AGI safety] 1. What's the problem & Why work on it now? 1.1 Post summary / Table of contents ==================================== This is the first of a [series of blog posts on the technical safety problem for hypothetical future brain-like Artificial General Intelligence (AGI) systems](https://...
35f863dc-96dd-4e9b-bcf8-fe2e120197ba
trentmkelly/LessWrong-43k
LessWrong
Meetup : Regular Moscow Meetup Discussion article for the meetup : Regular Moscow Meetup WHEN: 17 August 2014 02:00:00PM (+0400) WHERE: Russia, Moscow, ulitsa L'va Tolstogo 16 Here's our plan for the next time: * exercise on searching cognitive biases in texts * a talk on cognitive errors for the practical case ...
67dd861e-d10d-4a00-b562-c92e3ecfd66f
trentmkelly/LessWrong-43k
LessWrong
Answering some questions about EA A teacher in Vienna recently wrote to say that they had assigned an article about us as part of a Social Issues class, and asked whether I would be up for answering questions. The students asked me their questions via video snippets, and here are my answers: Gergo: When was the first ...
628cb338-b9b3-4306-a492-e8050d5cdc52
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
Beware popular discussions of AI "sentience" *Summary:*I apply some philosophy of mind to some common mistakes people make about (possible) AI minds. In our current situation (where we are very ignorant of the relationships between the following things), take care to talk about problem-solving ability, goal-orientedne...
0e7dd877-60c7-4a26-9264-aa079bf58aaa
trentmkelly/LessWrong-43k
LessWrong
Meetup : Vancouver Open Discussion Discussion article for the meetup : Vancouver Open Discussion WHEN: 03 August 2013 02:34:17AM (-0700) WHERE: 2505 W Broadway, Vancouver, BC We're having our usual meetup on Saturday. We're looking to expand the breadth of our meetup, so we'll be fielding suggestions and discussion...
3c9a7d8d-ebee-4f6c-9a98-fa1510db4872
trentmkelly/LessWrong-43k
LessWrong
Events in Daily? I'm in the process of refactoring how we handle the default posts view. The core goal was to implement some anti-spam features into it. But it highlighted some oddness in how we're currently filtering things. I wanted to check how this might effect people on GreaterWrong, and people who use the Daily...
c045eb98-01fa-4254-8721-2d8bb8e059ff
trentmkelly/LessWrong-43k
LessWrong
Probabilistic self-defeat arguments Alvin Plantinga‘s ‘evolutionary argument against naturalism‘ (EAAN) goes like this: 1. If humans were created by natural selection, and also not under the guidance of a creator (‘naturalism’), then (for various reasons he gives) the probability that their beliefs are accurate is l...
02494b5a-70a3-403f-9fad-f51fca020b3d
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Inference from a Mathematical Description of an Existing Alignment Research: a proposal for an outer alignment research program **Update:** *I'm pretty sure that the universal prior being malign defeats this proposal (see* [*this comment*](https://www.lesswrong.com/posts/rnzpYKWmNWEW5PQyq/inference-from-a-mathematical...
37945a25-e1dd-4f44-a684-c09e969a7ee3
trentmkelly/LessWrong-43k
LessWrong
Quantilizer ≡ Optimizer with a Bounded Amount of Output (Crossposted from my blog) In 2015 MIRI proposed quantilizers as a decision theory criterion that an AI may use that allows it strive towards a goal while not optimizing too hard at the goal, so that if a superintelligent were a quantilizer there's a smaller cha...
6dc0c7eb-4016-48e5-9c24-16bc0bdd0724
trentmkelly/LessWrong-43k
LessWrong
Give the model a model-builder (Thanks to Evan Hubinger for suggesting this problem and for comments on these ideas. Feedback is very welcome.) The Setup Suppose we have a prior pgood(model) that we think is non-deceptive, such that we can sample from this prior and get good models that are inner-aligned. These mod...
d4c61c2c-420d-4ca4-9e57-edf120f0bfca
trentmkelly/LessWrong-43k
LessWrong
Bayesianism (Subjective or Objective) I'm reading a paper called 'Reasonable Doubt and Presumtion of Innocence: The Case of the Bayesian Juror' for a Physics/Policy course I'm taking, and am a bit confused by something in it. Note here that I'm quite new to Bayesianism and do not claim to understand in entirity how it...
c8cb7c2e-c312-4193-9c01-9f2ed2189c62
trentmkelly/LessWrong-43k
LessWrong
CCing Mailing Lists on External Communication In my current work, and at various previous companies, I've done a lot of coordination with external groups. We generally use email for this: it's very standard and works well enough. One place where it doesn't work very well, though, is making it possible for others work...
b45ea570-cd54-4ddb-bfc6-1efc9af1ed2b
trentmkelly/LessWrong-43k
LessWrong
Graphical World Models, Counterfactuals, and Machine Learning Agents This is the second post in a sequence. For the introduction post, see here. Graphical World Models A world model is a mathematical model of a particular world. This can be our real world, or an imaginary world. To make a mathematical model into a m...
2451f83a-b04c-4861-b3ec-674fbe141606
trentmkelly/LessWrong-43k
LessWrong
Meetup : Ann Arbor meetup Discussion article for the meetup : Ann Arbor meetup WHEN: 22 August 2015 02:00:00PM (-0400) WHERE: Pizza House, 618 Church Street, Ann Arbor, Mich. 48104 Our bi-monthly discussion group. There is no minimum to attend in terms of age, reading history, karma, or otherwise. If you can read t...
9bb36d5f-761b-4298-b3e7-92652dc11db2
trentmkelly/LessWrong-43k
LessWrong
An answer to a possible objection to cash-transfer charities This is an answer to a possible objection to cash-transfer charities like GiveDirectly. I remember reading about this on LessWrong a while ago, but I can't find the discussion now. I was planning on asking about this on an Open Thread, but I got curious, did...
11d88cac-8d6b-4f0c-a1a4-88fd002c11e2
trentmkelly/LessWrong-43k
LessWrong
500 Million, But Not A Single One More - The Animation This video is an animation of 500 Million, But Not A Single One More, by @jai  Some notes for the curious about the cult of Sopona, the god of Smallpox of the Yoruba people When I was getting feedback on the storyboard from Jai, he mentioned that there is an a...
41193ec8-63b7-493c-b7b9-bf531c9a5a7d
trentmkelly/LessWrong-43k
LessWrong
MIRI Donation Collaboration Station Redux: The Final Push (IMPORTANT) As you likely already know, there is a one-day only fundraiser for MIRI going on tonight. 1 hour left, or less on the clock. Donation page is HERE And, "$250K in Microsoft in-kind giving is available for the organization which gets the most unique...
3eb39dd7-0ec4-47bf-873f-1093ba13509b
trentmkelly/LessWrong-43k
LessWrong
[Link] "Doing being rational: polymerase chain reaction" by David Chapman This is a link post for: * Doing being rational: polymerase chain reaction | Meaningness I'm not going to quote the content at the link itself. [Should I?] David Chapman – the author of the linked post – claims that "meta-rational" methods a...
440d9b83-6fab-4a8b-aa83-72e6a2ebcfe2
StampyAI/alignment-research-dataset/blogs
Blogs
An empirical analysis of compute-optimal large language model training In the last few years, a focus in language modelling has been on improving performance through increasing the number of parameters in transformer-based models. This approach has led to impressive results and state-of-the-art performance across many...
c1e96bc7-35a7-42b0-b5ef-4c0f2a927862
trentmkelly/LessWrong-43k
LessWrong
Building awareness habits (cognitive) Something I struggle with (probably largely due to ADHD) is building habits/patterns of being aware of things I want to notice/practice cognitively. For example, I find myself easily becoming somewhat thoughtless in social situations, in that I do not stop to observe, think, and ...
f90bd6eb-1f6a-4d2a-b0a0-afffda509a4b
trentmkelly/LessWrong-43k
LessWrong
[AN #162]: Foundation models: a paradigm shift within AI 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 this spreadsheet of all summaries that have ever been in the newsletter....
caf0b363-7e3d-4500-b659-ea26913b50fa
trentmkelly/LessWrong-43k
LessWrong
Personal Notes On Productivity (A categorization of various resources) For each topic, I’ve curated a few links that I’ve found to be pretty high quality.  * Meta:(Epiphany Addiction, Reversing Advice, Excellence Porn) * @Learning:  * Success People: (Mastery),(ChoosingTopics: Osci,PG) * Thinking: (Ikigai, Stoici...
dc4403d4-e9ba-4272-95ec-4b3585171abd
trentmkelly/LessWrong-43k
LessWrong
Weekly LW Meetups This summary was posted to LW Main on August 5th. The following week's summary is here. New meetups (or meetups with a hiatus of more than a year) are happening in: * Bay City Meetup: 19 August 2016 01:25PM Irregularly scheduled Less Wrong meetups are taking place in: * European Community Weeke...
65700311-53e5-41ba-a3ff-4089b9ceb624
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Task decomposition for scalable oversight (AGISF Distillation) AGISF Week 4 - Task decomposition for scalable oversight *This text is an adapted excerpt from the task decomposition for scalable oversight section of the AGISF 2023 course, held at ENS Ulm in Paris on March 16, 2023. **Its purpose is to provide a concis...
13619ab1-fb6f-4e0d-b811-c1ff08fce3a8
trentmkelly/LessWrong-43k
LessWrong
"Technical implication: My worst enemy is an instance of my self." I think this one needs more discussion, it looks like a really valuable and interesting train of thought. In "You'll be who you care about," Stuart Armstrong wrote - > Instead of wondering whether we should be selfish towards our future selves, let's...
660026e9-4386-4824-9cb8-ff7b5b94fafb
trentmkelly/LessWrong-43k
LessWrong
What do you mean with ‘alignment is solvable in principle’? Typically, I saw researchers make this claim confidently in one sentence. Sometimes, it's backed by a loose analogy. [1] This claim is cruxy. If alignment is not solvable, then the alignment community is not viable. But little is written that disambiguates a...
f5b1269e-7667-4de9-83ed-9e950cdb1602
trentmkelly/LessWrong-43k
LessWrong
An optimality result for modal UDT Yesterday, I gave a definition of "decision theory" and "fair decision problem" in the context of provability logic, and gave a formal version of drnickbone's argument that no decision theory can perform optimally on every "fair" decision problem. Today, I prove that for any provably...
8ea68894-e123-4f40-a5c0-663ce9c51ea8
trentmkelly/LessWrong-43k
LessWrong
$1000 bounty for OpenAI to show whether GPT3 was "deliberately" pretending to be stupider than it is Twitter thread by Eliezer Yudkowsky, with the bounty in bold:  > So I don't want to sound alarms prematurely, here, but we could possibly be looking at the first case of an AI pretending to be stupider than it is. In ...
77c06e3d-1149-45be-9cfa-a9a9749c25b0
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
Non-manipulative oracles Benja had a post about trying to get predictors to not manipulate you. It involved a predictor that could predict tennis matches, but where the prediction could also manipulate the impact of those matches. To solve this, Benja imagined the actions of a hypothetical CDT reasoning agent locate...
e9539b95-16a8-42a0-b420-7366de13e8af
trentmkelly/LessWrong-43k
LessWrong
What EY and LessWrong meant when (fill in the blank) found them. Somewhere I in the last year I read a long post from years ago saying what a relief it was to them when they found this place. As I recall this was a middle-aged programmer. Can anyone point me to that post? Thanks.
00ebb2c3-134c-4256-8cba-b34a5524a2cd
LDJnr/LessWrong-Amplify-Instruct
LessWrong
"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 argument, which I hear often and have to ...
3ecf68e3-1498-4c1f-ad9f-e6154e1fd608
trentmkelly/LessWrong-43k
LessWrong
Geoffrey Miller on Effective Altruism and Rationality The first part of an extensive interview with Dr. Geoffrey Miller. Topics include: 1. Why we should send a hit squad to Russia. 2. How to make effective altruism sexy. 3. Why he doesn't eat anchovies.
3f8b04a6-e591-4481-a0f5-4417bea95b5e
trentmkelly/LessWrong-43k
LessWrong
Reinterpreting "AI and Compute" Some arguments saying that the recent evidence about the speed at which compute has been increasing and has been responsible for rapid progress in machine learning, might mean that we should be less worried about short timelines, not more. > [...] Overall, it seems pretty common to int...
e8bc617a-8ab4-4f23-84ef-1967c2d7bef3
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Has anyone written stories happening in Hanson's em world? EtA: Robin Hanson complains that while most hard sci-fi may get the hard tech right, they often get the economics wrong. I'm interested in stories that get both right.
b0e2b6b6-cd55-4e0d-8e8c-e457ee49e8d1
trentmkelly/LessWrong-43k
LessWrong
Before rationality- a snapshot This post is roughly analogous to the “before” photos that you see people use to demonstrate the effects of dieting, except I want to use it a sort of time capsule to compare my conception of myself as a thinker now to that conception in [some] years time, when I am further along in my s...
0fceab61-429a-45b3-b0ea-fa53191434a4
trentmkelly/LessWrong-43k
LessWrong
Meetup : Denver Area LW February Meetup Discussion article for the meetup : Denver Area LW February Meetup WHEN: 07 February 2017 07:00:00AM (-0700) WHERE: 4955 South Ulster Street #103, Denver, CO Join us to socialize and discuss Rationalist topics! We usually have semi-separate groups divided by topic. Individual...
64c25ab4-c460-4df3-a023-e4dd87efaf82
trentmkelly/LessWrong-43k
LessWrong
Audio from Eliezer's talk at the Oxford Transhumanists In January we hosted Eliezer at an Oxford Transhumanists meeting. He spoke about why AI is such an incredibly consequential consideration, over and above other technologies. This will not be new material for regular lesswrong readers. The recordings from Eliezer's...
9a04f5e0-f374-448d-919b-a99690b2bfbc
StampyAI/alignment-research-dataset/eaforum
Effective Altruism Forum
My take on What We Owe the Future *Cross-posted from [Foxy Scout](https://www.foxy-scout.com/wwotf-review/)* Overview ======== [What We Owe The Future](https://whatweowethefuture.com/) (WWOTF) by Will MacAskill has recently been released with [much fanfare](https://whatweowethefuture.com/press/). While I strongly ...
2780853a-ebfc-4932-bd26-75c217c90524
trentmkelly/LessWrong-43k
LessWrong
Variation on conformity experiment A new variation on the Asch conformity experiment was recently published. The experiment was performed in Japan and used polarizing glasses to show different lines to different people in the same room, so that the subjects had to disagree with others they actually knew, and who genui...
7ef40f04-cbdc-4eac-a764-1ce3b0ce91eb
trentmkelly/LessWrong-43k
LessWrong
Preserving Practical Science (Phronesis) This is my first commentary on the previous Saint Louis Junto Meeting whose notes are here. Since a meetup can cover diverse topics, I have decided that I will not include running commentary in the meeting notes, and instead reflect upon a few of the discussions in subsequent p...
66830013-03af-4813-963c-47bf5d244430
trentmkelly/LessWrong-43k
LessWrong
Can you learn (music) efficiently by grinding? Crossposted from my Substack. You don't need to watch the videos to understand what I learned today; you can grok it from reading the text. I'd like to make one more post in this series, to determine whether a person can use this grinding strategy to correctly estimate ...
3086eb23-9281-4023-abf6-3833f30b3c49
trentmkelly/LessWrong-43k
LessWrong
Anki deck for Cognitive Science in One Lesson I've made a non-comprehensive Anki deck (shared as "Cognitive Science in One Lesson Deck") for Lukeprog's Cognitive Science in One Lesson (a summary of Bermudez's Cognitive Science textbook). I focused on the parts about the brain. Please point out errors or post revised v...
07664f0a-5698-4506-8de6-4efc052cd7fb
trentmkelly/LessWrong-43k
LessWrong
What does social psychology tell us about getting consistent evaluations out of groups? Suppose we have different groups repeatedly making the same kind of decision (ex: awards in a civil lawsuit). We would like their decisions to be consistent and predictable, and are even willing to trade some accuracy to do so. So...
7788082b-11ad-4157-afee-a631a1de2550
trentmkelly/LessWrong-43k
LessWrong
Chapter 93: Roles, Pt 4 Harry had walked into the Great Hall, looked around only once, grabbed enough calories to sustain himself, walked out, put on his Cloak again and found a small random corner in which to eat. Seeing the students at their tables - Feeling revulsion when you look at other humans is not a good sig...
62913507-95ba-418c-9dd9-135284be0068
trentmkelly/LessWrong-43k
LessWrong
LessWrong as place for scientifically literate advice Disclaimer and abstract: English is not my native language, so please notify me if you see grammatical or stylistic mistakes. This posting is mainly an account of my thoughts. It is highly informal and contains almost no advice but asks for it. I am relatively new ...
3f95dba3-5f12-4ea5-8152-e26cae51e14d
trentmkelly/LessWrong-43k
LessWrong
Escape from Alderaan I This is Part 4 of a Star Wars fanfiction I began writing 6 months ago. The order you read it in is very important. Start from the beginning for the best experience and to avoid spoilers. ---------------------------------------- Obi-Wan Kenobi bought passage from a scumbag with a confusing name...
ce5dc8b6-67fd-4703-ad8d-67b5a9615953
trentmkelly/LessWrong-43k
LessWrong
Properties of current AIs and some predictions of the evolution of AI from the perspective of scale-free theories of agency and regulative development In this essay, I try to convey a set of related ideas that emerged in my mind upon reading the following papers: * "Dream of Being: Solving AI Alignment Problems with...
009bfaca-957b-48a2-8aa6-e437b2e77871
trentmkelly/LessWrong-43k
LessWrong
What is Ra? Sarah Constantine coined Ra in this blog post and it has become a frequently referenced concept within the Rational-sphere. However, I don't feel she provided a clear definition. How would you define it?
251fabcd-0c56-4888-a168-e6b513be46f0
StampyAI/alignment-research-dataset/lesswrong
LessWrong
Extremely Counterfactual Mugging or: the gist of Transparent Newcomb > > Omega will either award you $1000 or ask you to pay him $100. He will award you $1000 if he predicts you would pay him if he asked. He will ask you to pay him $100 if he predicts you wouldn't pay him if he asked.  > > > Omega asks you to pay h...
6a6f2c67-c7ff-42ab-b24e-fc4af0b54ed4
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
Bing Just Upgraded YouTube (and changed the internet forever) Bing just upgraded YouTube and changed the internet forever and I'm going to show you exactly how I know that statement sounds crazy especially about something called Bing but let me prove it to you you can now open up any YouTube video and click on the Bin...
c1e67f5b-61bf-43dd-b0dc-5d25792d0cf5
StampyAI/alignment-research-dataset/alignmentforum
Alignment Forum
A transparency and interpretability tech tree *Thanks to Chris Olah, Neel Nanda, Kate Woolverton, Richard Ngo, Buck Shlegeris, Daniel Kokotajlo, Kyle McDonell, Laria Reynolds, Eliezer Yudkowksy, Mark Xu, and James Lucassen for useful comments, conversations, and feedback that informed this post.* The more I have tho...
f2d82f7e-8591-47d2-9853-221e6ebddf53
StampyAI/alignment-research-dataset/youtube
Youtube Transcripts
Possible Paths to Artificial General Intelligence [Music] all right thank you very much everyone so I'm Josh and the other panels are already introduced I'm going to start off with just very brief introductory remarks and then we're each gonna give short opening statements I'm not sure how much debate is gonna because...
8794fb8e-fc66-4a96-a6b0-8b75ecbf9e26
trentmkelly/LessWrong-43k
LessWrong
[Meta] Server Slow Is it just me or has the server being unusually slow the past couple of days?  During particularly bad times I'm even getting various HTTP errors.  
29940d91-c42b-49ed-a057-4fdf1cf3f013
trentmkelly/LessWrong-43k
LessWrong
Saturating utilities as a model Okay, it is a very raw idea, but consider the utility processing that works as following: 1: The utility i'm speaking of is not 'happiness', nor is it 'strength of the compulsion', the utility is only used for the purpose of comparing between futures to pick the one with larger utility...
7163b37e-2b64-4a8c-b7d1-bc01b7b3a455
trentmkelly/LessWrong-43k
LessWrong
Sweet Unconsciousness Hello LW, There has been some interest around here about atypical mental phenomena such as synesthesia, blindsight, absence of mental images, and so on. There have also been sappy posts and requests for help. I'd like to discuss my personal quirks in the hopes that it is interesting and someone ...
75c50a4f-5cf9-4bae-a0fc-3b17b8bca7ed
StampyAI/alignment-research-dataset/lesswrong
LessWrong
AI #23: Fundamental Problems with RLHF After several jam-packed weeks, things slowed down to allow everyone to focus on the potential room temperature superconductor, [check Polymarket to see how likely it is we are so back](https://polymarket.com/event/is-the-room-temp-superconductor-real) and bet real money, [or Man...
948dfb01-5363-46ee-89d1-975c2836b385
trentmkelly/LessWrong-43k
LessWrong
Comments on CAIS Over the last few months I’ve talked with Eric Drexler a number of times about his Comprehensive AI Services (CAIS) model of AI development, and read most of his technical report on the topic. I think these are important ideas which are well worth engaging with, despite personally being skeptical abou...
7755f66f-2024-4f83-be7d-d3334173f957
trentmkelly/LessWrong-43k
LessWrong
A High Decoupling Failure High-decoupling vs low-decoupling or decoupling vs contextualizing refers to two different cultural norms, cognitive skills, or personal dispositions that change the way people approach ideas. > High-decouplers isolate ideas from each other and the surrounding context. This is a necessary pr...
8f3daaaa-4f55-477b-8bd2-bda674580b8c
trentmkelly/LessWrong-43k
LessWrong
[U.S. specific] PPP: free money for self-employed & orgs (time-sensitive) [cross-posted from EA forum] PPP is the "paycheck protection program", a big US government giveaway to businesses including self-employed people (e.g. many grant recipients) and nonprofits. There was a PPP round last year—see this EA Forum post...
348934ef-df78-426e-9f09-28c33607b82f
trentmkelly/LessWrong-43k
LessWrong
Inaccessible finely tuned RNG in humans? It was my impression that humans are bad at random (number) generation. Tell a person to arrange stars on a black canvas randomly, and they'll space them out more or less equidistantly and uniformly. If asked to generate a sequence of coin flips, we would avoid longer runs of t...
18b1a1f7-7c42-436f-a35c-a550fab78103
trentmkelly/LessWrong-43k
LessWrong
Lesswrong Community's How-Tos and Recommendations The Lesswrong community is often a dependable source of recommendations, network help, and advice. When I'm looking for a book or learning material on a topic I'll often try and search here to see what residents have found useful. Similarly, social advice, anecdotes an...
19bf2b2e-f18c-47f9-b43d-c5a6f1134f37
trentmkelly/LessWrong-43k
LessWrong
Equilibrium and prior selection problems in multipolar deployment To avoid catastrophic conflict in multipolar AI scenarios, we would like to design AI systems such that AI-enabled actors will tend to cooperate. This post is about some problems facing this effort and some possible solutions. To explain these problems...
c53de208-d893-46f9-bb26-54eff1a56c12
trentmkelly/LessWrong-43k
LessWrong
Pain is not the unit of Effort (Content warning: self-harm, parts of this post may be actively counterproductive for readers with certain mental illnesses or idiosyncrasies.) > What doesn't kill you makes you stronger. ~ Kelly Clarkson. > > No pain, no gain. ~ Exercise motto. > > The more bitterness you swallow, th...