id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
bc611e41-cadf-40e7-a121-ee4566be2df3 | trentmkelly/LessWrong-43k | LessWrong | LLMs for language learning
My current outlook on LLMs is that they are some combination of bullshit to fool people who are looking to be fooled, and a modest but potentially very important improvement in the capacity to search large corpuses of text in response to uncontroversial natural-language queries and automatic... |
b9ed1af3-2380-48b3-91f2-44264ff2222e | trentmkelly/LessWrong-43k | LessWrong | Koan: divining alien datastructures from RAM activations
[Metadata: crossposted from https://tsvibt.blogspot.com/2024/04/koan-divining-alien-datastructures-from.html.]
Exploring the ruins of an alien civilization, you find what appears to be a working computer——it's made of plastic and metal, wires connect it to vari... |
a495ed40-b396-4240-ab8f-f881d0ac1a84 | trentmkelly/LessWrong-43k | LessWrong | How to develop safe superintelligence
Author: Martí Llopart
Introduction
In the world we live in, there are many problems that need to be urgently solved: climate change[1], the risk of a new pandemic[2], wealth inequality[3], the aging population[4] and more. All of these problems are hugely complex and solving ... |
42e944ee-759d-49d7-8ed6-07c540bb08cb | trentmkelly/LessWrong-43k | LessWrong | Spaced Repetition literature review prize: And the winner is...
The Spaced Repetition literature review prize for the best new review of the evidence on Spaced Repetition has ended and the judging panel has made its decision. The prize attracted entries from Duke (entry) and Gwern (entry). After reviewing the submissi... |
5b788c20-fa40-46fa-88f9-d5069d85ed16 | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post950
TL;DR Many evaluations and automated interpretability rely on using multiple models to evaluate and interpret each other. One model is given full access to the text output of another model in OpenAI's automated interpretability and many model-graded evaluations. We inject text that directly addresses the e... |
384efb45-b04a-4693-845d-44c9c1f662db | trentmkelly/LessWrong-43k | LessWrong | Counter-theses on Sleep
Alexey Guzey’s Theses on Sleep gained a lot of popularity and acclaim on LessWrong and among people I follow on social media, despite largely consisting of what I think were weak arguments and misleading claims. I found that a bit surprising, so I decided to write a post pointing out several of... |
62e64e73-5011-4d19-9398-4ef65ff2aa2d | trentmkelly/LessWrong-43k | LessWrong | Are ya winning, son?
Hello,
I've been reading through various Newcomblike Problems, in order to get a better understanding of the differences between each Decision Theory. From what I can tell, it seems like each Decision Theory gets evaluated based on whether they are able to "win" in each of these thought experimen... |
440315d4-42ac-44a6-ae54-73ea9e1e48d0 | trentmkelly/LessWrong-43k | LessWrong | Katla on Katja
I don’t have time to write, but apparently Katla has time for all kinds of marginally valuable activities. Here are her latest thoughts:
> So apparently Katja has been learning about all kinds of fascinating philosophy stuff for like seven months. Am I the only one to notice that she hasn’t mentioned a... |
93d03e56-03cc-411b-a079-b948d027119d | trentmkelly/LessWrong-43k | LessWrong | Interview with Skynet
I used GPT-3 to generate Skynet's responses. I frequently backtracked by running the same query multiple times and choosing my favorite response.
----------------------------------------
Lsusr: Many of my readers like to speculate about what a superintelligent AI might do. I think baseless spec... |
b9112d53-8fb9-4d61-bc9d-15d235a7254b | trentmkelly/LessWrong-43k | LessWrong | Secular Solstice 2020
Last night, about 225 people gathered for songs and stories. In the spirit of putting in a ridiculous amount of effort to keep doing normal things in abnormal times, a group of friends built a platform where we could listen together, and sing together, as we do every year. This was the biggest ga... |
3dc49048-d437-4ee3-8a38-06501516c2dc | trentmkelly/LessWrong-43k | LessWrong | [Link] "The Problem With Positive Thinking"
Psychology researchers discuss their findings in a New York Times op-ed piece.
The take-home advice:
> Positive thinking fools our minds into perceiving that we’ve already attained our goal, slackening our readiness to pursue it.
>
> ...
>
> What does work better is a hy... |
c7c13be3-a7c9-4533-8509-ad3230668347 | trentmkelly/LessWrong-43k | LessWrong | Collating widely available time/money trades
In the xkcd comic Working, a man is seen filling up his gas tank. "Why are you going here", says the observer, "Gas is ten cents a gallon cheaper at the station five minutes that way". He responds "Because a penny saved is a penny earned". Randal's pragmatically spirited ca... |
5bbbf2f2-24a3-42e2-a589-a34d1406a7ec | trentmkelly/LessWrong-43k | LessWrong | AI Safety: A Climb To Armageddon?
by Herman Cappelen, Josh Dever and John Hawthorne
Abstract: This paper presents an argument that certain AI safety measures, rather than mitigating existential risk, may instead exacerbate it. Under certain key assumptions - the inevitability of AI failure, the expected correlation b... |
5b648603-2918-4ae8-bf13-75510eed2fbb | trentmkelly/LessWrong-43k | LessWrong | What do you mean by rationalism?
I've been lurking here a bit, and am trying to understand what people here mean by rationalism. Many articles here seem to refer to discussion participants as rationalist while meaning very seemingly-different things, including intelligent, socially awkward, well-educated, and unencum... |
7939a8ef-d605-49ef-84e0-100c0f4b0579 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Imaginary Positions
Today's post, Imaginary Positions was originally published on 23 December 2008. A summary (taken from the LW wiki):
> People who are not members of a minority group may somehow come to believe that members of this group possess certain traits which seem to "fit". These traits are no... |
4aee8269-67ab-4554-b0e1-a1e078f91cc2 | trentmkelly/LessWrong-43k | LessWrong | Test EA human swarm [Friday 4 Aug - 11 AM PST]
|
2705af1c-d5a8-4677-935b-d35c2eec21e6 | trentmkelly/LessWrong-43k | LessWrong | I made a card game to reduce cognitive biases and logical fallacies but I'm not sure what DV to test in a study on its effectiveness.
First, how and why I made this game
I spent several years working as the chief product officer of a software company in Melbourne, and during this time, there was a point where I raised... |
93e54a3e-27ea-43b9-af21-d6b8a66ea7b4 | trentmkelly/LessWrong-43k | LessWrong | A Word to the Wise is Sufficient because the Wise Know So Many Words
Collect Ontologies
An ontology is a way of bucketing reality. For example, Russians distinguish голубой from синий whereas Anglos bucket both into "blue". An ontology is an implicit set of priors about the world. If you never bucket people according ... |
d533d814-d665-491e-aa98-af677105e53e | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | "AI and Compute" trend isn't predictive of what is happening
(open in a new tab to view at higher resolution)
In May 2018 (almost 3 years ago) OpenAI published their ["AI and Compute"](https://openai.com/... |
a890cb0a-3673-42af-979c-fc735ad08e60 | trentmkelly/LessWrong-43k | LessWrong | Planning 101: Debiasing and Research
Planning 101: Techniques and Research
<Cross-posed from my blog>
[Epistemic status: Relatively strong. There are numerous studies showing that predictions often become miscalibrated. Overconfidence in itself appears fairly robust, appearing in different situations. The actual mech... |
1e68d211-8bd8-4007-8bda-7e081b22ed1d | StampyAI/alignment-research-dataset/blogs | Blogs | "AI risk drone"
"AI risk drone"
---------------
i want to coin the term "AI risk drone" to mean "person singlemindedly and as-optimizedly-as-possible dedicated to [AI risk mitigation](say-ai-risk-mitigation-not-alignment.html)". the term "alignment drone" derives similarly.
i am doing this because i am unaware of ... |
87e60973-f43b-4604-bcf7-0cdda9f3fea8 | trentmkelly/LessWrong-43k | LessWrong | A First Sketch of the Nature of Heat (Novum Organum Book 2: 15-25)
This is the eleventh post in the Novum Organum sequence. For context, see the sequence introduction. For the reading guide, see earlier posts in the sequence.
We have used Francis Bacon's Novum Organum in the version presented atwww.earlymoderntexts.c... |
d0d17ced-1c55-4d6d-999e-a709255b55a0 | trentmkelly/LessWrong-43k | LessWrong | Applied art of rationality: Richard Feynman steelmanning his mother's concerns
First, imagine your parents disapproving of your first love. Imagine your mother inventing a whole whack of reasons why you shouldn't marry him/her. Now imagine being rational enough to acknowledge and address all her concerns while remaini... |
e2ff9a1b-52cf-42f5-ad40-29fc5200511d | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | [AN #127]: Rethinking agency: Cartesian frames as a formalization of ways to carve up the world into an agent and its environment
Alignment Newsletter is a weekly publication with recent content relevant to AI alignment around the world. Find all Alignment Newsletter **[resources here](http://rohinshah.com/alignment-n... |
d46c81c1-a2cc-451a-a469-cfb98dc00969 | trentmkelly/LessWrong-43k | LessWrong | Embedded Naive Bayes
Suppose we have a bunch of earthquake sensors spread over an area. They are not perfectly reliable (in terms of either false positives or false negatives), but some are more reliable than others. How can we aggregate the sensor data to detect earthquakes?
A “naive” seismologist without any statis... |
1666fd75-7930-47d7-a025-f27ba43112e6 | trentmkelly/LessWrong-43k | LessWrong | Quick modeling: resolving disagreements.
|
bdc41aef-ff67-4412-aa8d-fa1d1bbe9e0d | trentmkelly/LessWrong-43k | LessWrong | Unknown Probabilities
An unknown probability sounds like a type error. There are unknowns, such as the result of a coin flip. And there are probabilities that these unknowns take certain values, such as the probability that the flip comes up heads.
As a formula, P(Result=Heads)=1/2 The unknown, the result of the flip... |
c5329afa-1bb8-4ea6-8d7a-a9631e88bcf3 | trentmkelly/LessWrong-43k | LessWrong | The Red Bias
Summary: This color alters your perception of the world. Evidence that it does, how it does, why it does and some implications are presented below.
(Overcoming Bias: Seeing Red)
> Across a range of sports, we find that wearing red is consistently associated with a higher probability of winning. These... |
5efe8b2d-9b5a-491f-9bad-c6dd2c7817f3 | trentmkelly/LessWrong-43k | LessWrong | Hive mind scenario
In a conceivable future, humans gain the technology to eliminate physical suffering and to create interfaces between their own brains and computing devices--interfaces which are sufficiently advanced that the border between the brain and the computer practically vanishes. Humans are able to access ... |
5942454a-572f-4bdd-91d5-b89d056fa432 | trentmkelly/LessWrong-43k | LessWrong | Can we have Epiphanies and Eureka moments more frequently?
Epiphanies and Understanding, Pain, Eureka-Moments.
TL;DR - mental-pain rushing in conveys urgent information, this experience is similar to when understanding rushes in when it is pleasurable - such as the sudden feeling of understanding of a philosophical i... |
f2b92d2f-ccf2-4557-8e46-c0fb577e1f32 | trentmkelly/LessWrong-43k | LessWrong | Problems in evolutionary psychology
Note: The primary target of the post is not professional, academic evolutionary psychology. Rather, I am primarily cautioning amateurs (such as LW regulars) about some of the caveats involved in (armchair) evpsych and noting the rigor required for good theories. While the post does... |
7a240efb-fd25-49b7-8307-67e412e07210 | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post252
This is a linkpost for a new research paper of ours, introducing a simple but powerful technique for jailbreaking, Best-of-N Jailbreaking, which works across modalities (text, audio, vision) and shows power-law scaling in the amount of test-time compute used for the attack. Abstract We introduce Best-of-N ... |
fe0ab420-fd2a-43f0-8105-7f216f269516 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Allais Hack -- Transform Your Decisions!
The [Allais Paradox](http://www.overcomingbias.com/2008/01/allais-paradox.html), though not actually a paradox, was a classic experiment which showed that decisions made by humans do not demonstrate consistent preferences. If you [actually want to accomplish something](http://w... |
a974241e-8d21-4449-ae5f-e429445843e2 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Six AI Risk/Strategy Ideas
AI risk ideas are piling up in my head (and in my notebook) faster than I can write them down as full posts, so I'm going to condense multiple posts into one [again](https://www.lesswrong.com/posts/vbtvgNXkufFRSrx4j/three-ai-safety-related-ideas). I may expand some or all of these into full ... |
77841159-b735-4cd4-9543-453089cb1765 | trentmkelly/LessWrong-43k | LessWrong | Housing Roundup #9: Restricting Supply
I’d split the latest housing roundup into local versus global questions. I was planning on waiting a bit between them.
Then Joe Biden decided to propose a version of the worst possible thing.
So I guess here we are.
What is the organizing principle of Bidenomics?
RESTRICT SUP... |
19c34892-705e-40af-aac9-316e93750bd7 | trentmkelly/LessWrong-43k | LessWrong | Cached Selves
by Anna Salamon and Steve Rayhawk (joint authorship)
Related to: Beware identity
Update, 2021: I believe a large majority of the priming studies failed replication, though I haven't looked into it in depth. I still personally do a great many of the "possible strategies" listed at the bottom; and they s... |
2cfab558-23e9-4050-a7e5-6231f1aeeb47 | trentmkelly/LessWrong-43k | LessWrong | MEETUP: February 9, Philadelphia
February 9th, Nam Phuong at 11th and Washington (a little south of the intersection) at 1PM.
The discussion prompt is Gwern's The Narrowing Circle.
> "The “expanding circle” historical thesis ignores all instances in which modern ethics narrowed the set of beings to be morally regard... |
a342b4cb-87de-4fdf-81c1-227ecfe6a00e | trentmkelly/LessWrong-43k | LessWrong | Metaculus is seeking Analytical Storytellers to write essays fortified with testable predictions
About Metaculus
At Metaculus, our aim is to improve human decision-making and coordination at scale by increasing analytic capacity, reasoning, and judgment. With the unique capabilities of the Metaculus forecasting platfo... |
e810080e-4261-4361-b013-7476a78ae101 | trentmkelly/LessWrong-43k | LessWrong | "We know how to build AGI" - Sam Altman
> We are now confident we know how to build AGI as we have traditionally understood it. We believe that, in 2025, we may see the first AI agents “join the workforce” and materially change the output of companies. We continue to believe that iteratively putting great tools in the... |
94f899ae-8e23-4eeb-bfa8-36a43d606bf4 | trentmkelly/LessWrong-43k | LessWrong | [LINK] MIT discovers the location of memories: Individual neurons
http://www.extremetech.com/extreme/123485-mit-discovers-the-location-of-memories-individual-neurons |
179019ba-4113-4e9f-8f3c-927b63f9128e | trentmkelly/LessWrong-43k | LessWrong | Epoch AI is hiring a CTO!
TLDR: We're hiring a Chief Technology Officer to oversee our tech stack and align it with our mission and growth goals. Responsibilities will include managing a team of 3-5 engineers and overseeing our website development, databases, and AI evaluation infrastructure. More information and the... |
ceb2d2fd-756a-4628-a1f3-def2c7f133b0 | trentmkelly/LessWrong-43k | LessWrong | Forecasting newsletter #3/2025: Long march through the institutions
Highlights:
* Manifold ending (a) cash markets, Kalshi slapped by regulators in Nevada
* Interactive Brokers now also offering election markets (a)
* Violet Hour looks at Where Would Good Forecasts Most Help AI Governance Efforts? (a). |
9cb42f46-a1c7-487a-93e4-08baf2e37995 | StampyAI/alignment-research-dataset/blogs | Blogs | Comparison of naturally evolved and engineered solutions
*This page describes a project that is in progress, and does not yet have results*
We are comparing naturally evolved and engineered solutions to problems, to learn about regularities that might let us make inferences about artificial intelligence from what we... |
9f18884e-bd88-4b8c-8ad6-895cfd62ea2c | trentmkelly/LessWrong-43k | LessWrong | Any research in "probe-tuning" of LLMs?
Is there any research in "probe-tuning" of LLMs, i.e., tuning LLM's parameter weights such that a specific probe (classifier) is more reliably detecting certain markers throughout the context, such as grammatical errors, aggression, manipulation, certain political bias, etc.?
T... |
ec87aacd-4520-41f9-a0f4-e1dfc6ac8fd9 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Modern portfolio theory
Discussion article for the meetup : Modern portfolio theory
WHEN: 19 May 2017 05:45:00PM (+0200)
WHERE: Lindstedtsvägen 3, room 1537, SE-114 28 Stockholm, Sverige
Learn the basics of risky assets and leverage. This talk contains math presented in an accessible way for the inexperien... |
fb12ab7a-5f1a-4f45-990f-7727d1b44883 | trentmkelly/LessWrong-43k | LessWrong | Absolute Zero: Reinforced Self-play Reasoning with Zero Data
> Reinforcement learning with verifiable rewards (RLVR) has shown promise in enhancing the reasoning capabilities of large language models by learning directly from outcome-based rewards. Recent RLVR works that operate under the zero setting avoid supervisio... |
1c380eca-1a13-42df-b667-e2ebed66304d | trentmkelly/LessWrong-43k | LessWrong | AI #12:The Quest for Sane Regulations
Regulation was the talk of the internet this week. On Capital Hill, Sam Altman answered questions at a Senate hearing and called for national and international regulation of AI, including revokable licensing for sufficiently capable models. Over in Europe, draft regulations were o... |
95dab7ff-62e2-42b7-bc3f-6440bd354fe3 | trentmkelly/LessWrong-43k | LessWrong | Solo Park Play at Three
Our three year old is about to turn four, and is bursting with a desire for independence. She's becoming more capable in all sorts of ways, and wants me to back off and let her do things. Today she wanted to go to the park by herself.
Now, we live close to the park, she could probably get ther... |
f5e26c45-0daf-42dd-b608-d017eb87ec1b | trentmkelly/LessWrong-43k | LessWrong |
Can an AI Have Feelings? or that satisfying crunch when you throw Alexa against a wall
This is an exploratory article on the nature of emotions and how that relates to AI and qualia. I am not a professional AI or ML researcher, and I approach the issue as a philosopher. I am here to learn. Rebuttals and clarificati... |
93899f0a-e463-48b7-9b0b-e9c8ebaa02bf | trentmkelly/LessWrong-43k | LessWrong | Mathematical simplicity bias and exponential functions
One of biases that are extremely prevalent in science, but are rarely talked about anywhere, is bias towards models that are mathematically simple and easier to operate on. Nature doesn't care all that much for mathematical simplicity. In particular I'd say that a... |
b09492cd-e196-491f-ac15-964cc27efbf8 | trentmkelly/LessWrong-43k | LessWrong | Some miscellaneous thoughts on ChatGPT, stories, and mechanical interpretability
I started playing with ChatGPT on December 1, 2023 and made my first post on December 3. Including this post, I’ve made 54 posts about ChatGPT. Nine of those posts report the work of others (45), perhaps with a comment or two from me, and... |
baedeaf2-6cff-4479-9160-4f50ffe209ae | trentmkelly/LessWrong-43k | LessWrong | What Peter Thiel thinks about AI risk
This is probably the clearest statement from him on the issue:
http://betaboston.com/news/2014/12/10/audio-peter-thiel-visits-boston-university-to-talk-entrepreneurship-and-backing-zuck/
25:30 mins in
TL;DR: he thinks its an issue but also feels AGI is very distant and hence... |
d99718ab-0b70-4be9-ac76-657d6a41e7bb | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "In the previous post, I argued that simply knowing that an AI system is superintelligent does not imply that it must be goal-directed. However, there are many other arguments that suggest that AI systems will or should be goal-directed, which I will discuss in this post.Note that I don’t think of this as the Tool AI v... |
44aad100-d95c-4c57-909e-6cdd40194025 | trentmkelly/LessWrong-43k | LessWrong | Universal Basic Income and Poverty
(Crossposted from Twitter)
I'm skeptical that Universal Basic Income can get rid of grinding poverty, since somehow humanity's 100-fold productivity increase (since the days of agriculture) didn't eliminate poverty.
Some of my friends reply, "What do you mean, poverty is still arou... |
015d4c4b-2e29-40dc-925f-fc7b3504a4b3 | trentmkelly/LessWrong-43k | LessWrong | divine carrot
(some writing without hesitation or editing. possibly with repetition and deviation.)
The classic reinforcement technique is carrot and stick.
Pure stick tends to breed resentment and laziness. Pure carrot breeds similarly, since you’re not doing anything.
There’s a divine carrot too: doing things bec... |
6b6a9499-7eb0-4b71-a9a2-e97dfa95bde8 | trentmkelly/LessWrong-43k | LessWrong | The Behavioral Economics of Welfare
Crossposted from Tumblr.
Epistemic Status: I offer some speculation, don’t really understand what is going on, would be glad to receive more insight.
You sometime hear libertarians say that welfare is unnecessary since private charity can replace it. IMO, it looks like the evidenc... |
c9941fd3-0903-4fe4-9c3d-9812c15747a3 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Pessimism About Unknown Unknowns Inspires Conservatism
[This](https://mkcohen-hosted-files.s3-us-west-1.amazonaws.com/Pessimism_alignmentforum.pdf) [EDIT: [final version](http://proceedings.mlr.press/v125/cohen20a/cohen20a.pdf), [presentation](https://www.learningtheory.org/colt2020/virtual/papers/paper_221.html)] is ... |
108778a5-9c3c-43e7-8fc1-310b41ae392e | trentmkelly/LessWrong-43k | LessWrong | Speculative Technologies launch and Ben Reinhardt AMA
Last week a new R&D organization launched: Speculative Technologies, a private DARPA-like approach to creating fundamental new technologies. I’ve been following the work of the founder, Ben Reinhardt, for a few years, and I’m very excited about this.
Ben is doing ... |
4934a6e4-584e-4395-b28e-9d7648a18bf9 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Algorithmic Similarity
The Problem
===========
If someone asked you and me to each write a python program to solves some novel problem then it wouldn't be surprising if our solutions looked quite different. If we then tried explaining to the other person the way our own program worked, one thing that might happen is ... |
7b21db03-cb6c-42d9-9335-127dbda6b2fa | trentmkelly/LessWrong-43k | LessWrong | The morality of disclosing salary requirements
Many firms require job applicants to tell them either how much money they're making at their current jobs, or how much they want to make at the job they're interviewing for. This is becoming more common, as more companies use web application forms that refuse to accept an... |
0de3b6ca-3856-47a6-88ac-c71fbb6a356f | trentmkelly/LessWrong-43k | LessWrong | I'm confused. Could someone help?
Imagine that I'm offering a bet that costs 1 dollar to accept. The prize is X + 5 dollars, and the odds of winning are 1 in X. Accepting this bet, therefore, has an expected value of 5 dollars a positive expected value, and offering it has an expected value of -5 dollars. It seems lik... |
20ad4576-46b3-4e32-9a1f-5cca28e6bd46 | trentmkelly/LessWrong-43k | LessWrong | Conflicted on AI Politics
About twice as many Americans think AI is likely to have a negative effect as a positive one. At a high level I agree: we're talking about computers that are smart in ways similar to people, and quickly getting smarter. They're also faster and cheaper than people, and again getting more so.
... |
c71e8cd9-4335-41f3-8d67-4800b6351359 | trentmkelly/LessWrong-43k | LessWrong | Deepmind's Gopher--more powerful than GPT-3
None |
2ace5dc5-2473-4a02-bc23-f40eeab21479 | trentmkelly/LessWrong-43k | LessWrong | "Real summer"?
(epistemic status: personal/metaphysical rambling)
Today - yes, that's August 26, 2024 - I took a really, really long nap. The first phase was from about 11:30am to 1pm; the second phase lasted from about 3pm to 4:30pm.[1] The dream itself was kind of incoherent and not worth retelling, except for one ... |
68f548a6-5dcf-440b-b49a-9e78caf3bd59 | trentmkelly/LessWrong-43k | LessWrong | Conference Report: Threshold 2030 - Modeling AI Economic Futures
This is an 8-page comprehensive summary of the results from Threshold 2030: a recent expert conference on economic impacts hosted by Convergence Analysis, Metaculus, and the Future of Life Institute. Please see the linkpost for the full end-to-end report... |
14ded128-eea7-4630-ab5f-8680f2504715 | trentmkelly/LessWrong-43k | LessWrong | Monthly Roundup #6: May 2023
BAD NEWS
Wild pigs continue to multiply and do increasing amounts of damage as they spread throughout America. Perhaps in the past we might have been willing to do what it would take to solve such problems. Today, we are not, and the situation gets steadily worse, and likely will continue... |
e46b1567-df66-443f-9bf5-1f6896039c58 | trentmkelly/LessWrong-43k | LessWrong | The Missing Math of Map-Making
Consider the clever arguer from The Bottom Line:
> Now suppose there is a clever arguer, holding a sheet of paper, and they say to the owners of box A and box B: “Bid for my services, and whoever wins my services, I shall argue that their box contains the diamond, so that the box will r... |
2023671b-f32f-4a54-bf46-05299ac2e878 | trentmkelly/LessWrong-43k | LessWrong | Ukraine Post #5: Bits of Information
Rather than attempt a synthesis this time around, I’m going to experiment with the opposite. Over the course of the last three weeks, I kept adding to a long list of sources and interesting things the same way I do for Covid-19 posts, but without an overall update type of post ther... |
a1b02c41-9a2d-4477-b41f-6b3ef97a5d7b | trentmkelly/LessWrong-43k | LessWrong | Rational Parenting?
I'm looking for insight into "rational parenting".
By that term, I mean two things. One is parenting in a rational way. The other is raising a child who shares rationalist thought. |
e5894daa-756d-4d3d-a663-442b70936d97 | trentmkelly/LessWrong-43k | LessWrong | A 'Practice of Rationality' Sequence?
(This is not your typical factual-answer question, but I think it makes sense to format this as a question rather than a post.)
TLDR: Recommend some posts for a "practice of rationality" sequence I want to curate! Proposing posts that should exist but don't is also cool.
I've be... |
42ae1474-1054-438c-9faf-770ca1df6b1a | StampyAI/alignment-research-dataset/arxiv | Arxiv | An overview of 11 proposals for building safe advanced AI
1 Introduction
---------------
This paper is a collection of 11 different proposals for building safe advanced AI under the current machine learning paradigm. While there already exist good overviews of various particular approaches, such as amplification[am... |
229eaf3b-38c7-4cbd-853d-04ab55941eef | trentmkelly/LessWrong-43k | LessWrong | An Introduction to Löb's Theorem in MIRI Research
Would you like to see a primer on several MIRI research topics (assuming only the background of having taken a course with proofs in math or computer science)? Or are you curious why MIRI does so much with mathematical logic, and why people on Less Wrong keep referring... |
11a54dc2-ced9-4451-966a-72f23fc333ee | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Meta "open sources" LMs competitive with Chinchilla, PaLM, and code-davinci-002 (Paper)
As the title says, Meta trained 4 foundational models with 7B, 13B, 33B, and 65B parameters respectively, and is open sourcing them for research.[[1]](#fnubd95r91ii) You can get their code on their Github repo: <https://github.com/... |
0f1ff78b-8101-455e-b36f-e687fd9fec13 | trentmkelly/LessWrong-43k | LessWrong | Quick notes on “mirror neurons”
1. Two definitions of “mirror neurons”
I have noticed the term “mirror neurons” being used in two ways:
* (A) Any set of neurons that activates both when I am doing / feeling / relating to X and when I am watching (or imagining) someone else doing / feeling / relating to that same X.
... |
9d5c6731-7d8c-46db-a87f-ae82165ec7e1 | trentmkelly/LessWrong-43k | LessWrong | Interpretability Externalities Case Study - Hungry Hungry Hippos
Some people worry about interpretability research being useful for AI capabilities and potentially net-negative. As far as I was aware of, this worry has mostly been theoretical, but now there is a real world example: The hungry hungry hippos (H3) paper.... |
8ddda06c-242c-4fa8-9ad4-78840506890a | trentmkelly/LessWrong-43k | LessWrong | LIST: I can't vote Karma on some people, some contexts.
What's up with that?
In a discussion thread, I can karma vote on anyone.
But if I select a person to see all their posts, for some people, karma is disabled while looking at them, and for others, it's not disabled. Same thing if I look at their posts under the... |
bde7c086-0d50-43e5-8cf5-e2c133366d7f | trentmkelly/LessWrong-43k | LessWrong | In Defense of Wrapper-Minds
Recently, there's been a strong push against "wrapper-minds" as a framework. It's argued that there's no specific reason to think that all sufficiently advanced agents would format their goals in terms of expected-utility maximization over future trajectories, and that this view predicts se... |
eb7fb895-face-486c-b47a-65ab44b98827 | trentmkelly/LessWrong-43k | LessWrong | Why Politics are Important to Less Wrong...
...and no, it's not because of potential political impact on its goals. Although that's also a thing.
The Politics problem is, at its root, about forming a workable set of rules by which society can operate, which society can agree with.
The Friendliness Problem is, at it... |
235d83d9-a397-426d-9eaf-e9029f74d2e9 | trentmkelly/LessWrong-43k | LessWrong | Building a transformer from scratch - AI safety up-skilling challenge
It is not always obvious whether your skills are sufficiently good to work for one of the various AI safety and alignment organizations. There are many options to calibrate and improve your skills including just applying to an org or talking with ot... |
7094e19e-2e42-4f69-9ffe-94dcc8a2ba79 | trentmkelly/LessWrong-43k | LessWrong | My poorly done attempt of formulating a game with incomplete information.
Epistemic Status: I was bored so I wrote this, forgetting that I was probably making a poor attempt of reinventing some wheel, and realising how I forgot to solve
Motivation: To formulate the stereotypical “Obedience Test” in game theory terms.... |
363fe45e-658b-4264-8711-1a1739025f15 | trentmkelly/LessWrong-43k | LessWrong | Tables of Presence, Nearby Absence, & Degrees of Intensity (Novum Organum Book 2: 10-14)
This is the tenth post in the Novum Organum sequence. For context, see the sequence introduction. For the reading guide, see earlier posts in the sequence.
We have used Francis Bacon's Novum Organum in the version presented at ww... |
4fb9e902-dc0e-4b00-a4c8-3946e168387c | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post1108
In my recent post on steering systems , I sketched an AI system which chooses which actions to execute via tree search over world states. In the original post, the system is meant to illustrate the possibility that component subsystems which are human-level and (possibly) non-dangerous when used individua... |
340bc864-757f-4a13-9172-07765f22b418 | trentmkelly/LessWrong-43k | LessWrong | My favorite AI governance research this year so far
I collected my favorite public pieces of research on AI strategy, governance, and forecasting from 2023 so far.
If you're a researcher, I encourage you to make a quick list of your favorite pieces of research, then think about what makes it good and whether you're a... |
ba290700-bdb8-4b31-8909-e93b3ea6b4b8 | StampyAI/alignment-research-dataset/arbital | Arbital | Methodology of unbounded analysis
summary: In modern AI and especially in [value alignment theory](https://arbital.com/p/2v), there's a sharp divide between "problems we know how to solve using unlimited computing power", and "problems we can't state how to solve using [computers larger than the universe](https://arb... |
02154782-5446-4b10-adfd-6f6eac648b53 | trentmkelly/LessWrong-43k | LessWrong | Don't clean your glasses
I rarely clean my glasses:
They do get dusty, but I find that as long as I don't clean them I don't notice it. Just like I no longer notice having glasses at all, despite the opaque frames and the sharp boundary between clear and fuzzy vision, the dust just fades from my perception.
Same go... |
4320ae0f-0800-4a29-b673-af6e0b1b7b01 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | The "AI Dungeons" Dragon Model is heavily path dependent (testing GPT-3 on ethics)
[AI Dungeons](https://play.aidungeon.io/) is a text-based video game that uses GPT-3 to generate parts of the story. (To try it, click [here](https://play.aidungeon.io/) → New Singleplayer Game → press 6 for "custom" – but note that you... |
da4075d9-d35c-4c32-91eb-c33d1925a243 | trentmkelly/LessWrong-43k | LessWrong | Personal evaluation of LLMs, through chess
Lots of people seem to resonate with the idea that AI benchmarks are getting more and more meaningless – either because they're being run in a misleading way, or because they aren't tracking important things. I think the right response is for people to do their own personal e... |
cca92b16-e932-4a69-9db4-311c3c021ed1 | StampyAI/alignment-research-dataset/aisafety.info | AI Safety Info | What is Stampy's AI Safety Info?
This is an open effort to build [a comprehensive FAQ](https://aisafety.info/) about [artificial intelligence existential safety](https://en.wikipedia.org/wiki/Existential_risk_from_artificial_general_intelligence)—the field trying to make sure that, when we build [superintelligent](htt... |
fc80a700-bc2c-4fec-b9fb-b4a0f520c46d | trentmkelly/LessWrong-43k | LessWrong | You have just been Counterfactually Mugged!
I'm going to test just how much the people here are committed to paying a Counterfactual Mugger, by playing Omega.
I'm going to roll a die. If it doesn't come up 5 or 6, I'm going to ask Eliezer Yudkowsky to reply to this article with the comment "I am a poopy head." If I r... |
92809894-3cba-4146-924d-5bb003224de5 | StampyAI/alignment-research-dataset/arbital | Arbital | Kripke model
**Kripke models** are [interpretations](https://arbital.com/p/) of [https://arbital.com/p/-534](https://arbital.com/p/-534) which turn out to be adequate under some basic constraints for a wide array of systems of modal logic.
#Definition
A Kripke model $M$ is formed by three elements:
* A set of **wor... |
a9f81a74-0af9-439b-95b0-afa51d00491c | trentmkelly/LessWrong-43k | LessWrong | Field tests of semi-rationality in Brazilian military training
I'm sharing this because someone who seems important in LessWrong asked me to, and I didn't want to miss the opportunity. But, publishing stories like this makes me feel more naked than if I were doing a striptease. And the worst part is that no one puts d... |
8279258e-0399-4320-9e5d-5aae7aad891c | trentmkelly/LessWrong-43k | LessWrong | Lesswrongers from the German-speaking world, unite!
As far as I can tell, there never has been a LW-meetup in a german-speaking country. This is crazy!
There should be enough Lesswrongers from Germany, Austria or Switzerland to achieve a reasonable group size!
To be a bit more specific, how about meeting in Munich?... |
943e7197-c7ae-49d4-9129-3fd3c2726814 | trentmkelly/LessWrong-43k | LessWrong | Guardians of the Truth
The criticism is sometimes leveled against rationalists: "The Inquisition thought they had the truth! Clearly this 'truth' business is dangerous."
There are many obvious responses, such as "If you think that possessing the truth would license you to torture and kill, you're making a mistake t... |
8dc5bb1c-6a53-4f7f-89dc-17fcb51a4ee3 | trentmkelly/LessWrong-43k | LessWrong | The future of humanity is in management
Imagine you wake up one day in the glorious techno-abundant future, powered by AI. You eagerly check your subscriber count on Substack, but to your dismay, it has fallen once again. There are now AI-generated blogs for every interest, and it’s so hard to find a niche that isn’t ... |
a7f4b3d8-7e2b-40c7-8bf8-08ca55e3db84 | trentmkelly/LessWrong-43k | LessWrong | Are the fundamental physical constants computable?
Fundamental physical constants are physical constants where the constant value seems unexplainable in terms of more basic physics. For example we have no idea why the relative masses of the various elementary particles are what they are - we see no reason why a muon i... |
16c1802e-8a34-4f93-8f07-3eebb5533316 | trentmkelly/LessWrong-43k | LessWrong | Why Rome?
> Separate an individual from society, and give him an island or a continent to possess, and he cannot acquire personal property. He cannot be rich.
>
> – Thomas Paine (1797)
Rome, once a mere collection of huts on a hilltop overlooking the Tiber, grew to become a vast and hugely influential empire. Why? T... |
61b1985a-2d61-4cc9-8a8e-dbd468700326 | trentmkelly/LessWrong-43k | LessWrong | What's Your Hourly Rate?
Here's an interesting post about calculating the value of your free time and why it might not be as simple as some tend to think. |
c45542c0-b12b-4d67-bdf0-2e87527c6824 | trentmkelly/LessWrong-43k | LessWrong | A $10k retroactive grant for VaccinateCA
I'm awarding a $10k retroactive grant to Karl Yang, for his role in starting VaccinateCA. In January 2021, the Covid vaccines were just starting to become available to the general public… in theory. In practice, there was a lot of confusion over how to get a dose. Demand for va... |
21fb747b-177b-4f91-8adf-15721c0a8e35 | trentmkelly/LessWrong-43k | LessWrong | Open Thread June 2010, Part 2
The title says it all. |
73a0f735-e986-4227-845b-e1192d287465 | StampyAI/alignment-research-dataset/arxiv | Arxiv | A Geometric Perspective on the Transferability of Adversarial Directions
1 Introduction
---------------
Many popular machine learning models, including deep neural networks, have been shown to be vulnerable to adversarial attacks (Szegedy et al., [2013](#bib.bib23); Goodfellow et al., [2014](#bib.bib5); Nguyen et a... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.