id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
74fa487f-9107-4c55-aeec-e9b3ac8e348c | StampyAI/alignment-research-dataset/lesswrong | LessWrong | A couple of questions about Conjecture's Cognitive Emulation proposal
Introduction
------------
Recently, Conjecture proposed a conception of a modular AI architecture, with each module being interpretable, having limited intellect, and functionally emulating the human mind. They called this concept [Cognitive Emulat... |
19b0c82d-0f9f-4525-86ca-9f7ce5a81174 | trentmkelly/LessWrong-43k | LessWrong | Let's fix the RSS feed so it doesn't re-post articles every time their titles are edited (forgot the hyphen in "re-post")
Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ... |
6d2c55b6-5edf-4f9d-8cca-de6afe24e2b4 | trentmkelly/LessWrong-43k | LessWrong | Are Interesting Problems Useful?
In a recent post I posed the question: is the common good served by directing research efforts towards theoretical problems which are interesting to researchers?
komponisto defends interesting problems, arguing that researcher's perceptions of interestingness are often better able to ... |
b4f8d11b-bb2c-4911-86f0-06ead8bbdda1 | trentmkelly/LessWrong-43k | LessWrong | What the Haters Hate
Cross-posted from Putanumonit. A bit more culture-warlike and snarky than the LW standard. Won't make as much sense to people outside the US. Previously on the subject of the IDW and outgroups.
----------------------------------------
Ezra Klein’s latest essay on the YouTube right / Intellectual... |
57042a0f-9e95-4f26-b411-632a2fb61e73 | trentmkelly/LessWrong-43k | LessWrong | Algorithms of Deception!
I want you to imagine a world consisting of a sequence of independent and identically distributed random variables Xi, and two computer programs.
The first program is called Reporter. As input, it accepts a bunch of the random variables Xi. As output, it returns a list of sets whose elements ... |
a0ac7568-2361-4a64-8ee9-49ce734ba17d | trentmkelly/LessWrong-43k | LessWrong | Sublimity vs. Youtube
The torture vs. dust specks quandary is a canonical one to LW. Off the top of my head, I can't remember anyone suggesting the reversal, one where the arguments taken by the hypothetical are positive and not negative. I'm curious about how it affects people's intuitions. I call it - as the titl... |
e9ef6f4d-b771-4f74-8392-ea984b9e81c2 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Sample Efficient Reinforcement Learning through Learning from Demonstrations in Minecraft
1 Introduction
---------------
The NeurIPS MineRL competition introduced by Guss et al. ([2019a](#bib.bib4 "The MineRL competition on sample efficient reinforcement learning using human priors")) is focused on the problem of s... |
bf99e334-7586-4828-b1ad-e74c6d78e079 | trentmkelly/LessWrong-43k | LessWrong | [Link] Good Judgment Project, Season Three
Previously: "Test Your Forecasting Ability, Contribute to the Science of Human Judgment" (May 2012), "Get Paid to Train Your Rationality" (August 2011)
-
Think you have what it takes to make good predictions? Since 2011, the Good Judgment Project (GJP) has been making pred... |
072a6ee7-3d33-4511-87fd-8b563a3e8e27 | trentmkelly/LessWrong-43k | LessWrong | The Great Organism Theory of Evolution
From Becoming Animal (David Abrams, 2010):
> Many months earlier, in a village near a wild forest preserve on the south coast of Java, I’d been warned by the local fishermen that there was an unusually bold individual among the bands of monkeys that roam the forest canopy, a p... |
4a143fcb-3ea7-41cc-8a37-27a814a1dad1 | trentmkelly/LessWrong-43k | LessWrong | Truly Part Of You
A classic paper by Drew McDermott, “Artificial Intelligence Meets Natural Stupidity,” criticized AI programs that would try to represent notions like happiness is a state of mind using a semantic network:
And of course there’s nothing inside the HAPPINESS node; it’s just a naked LISP token with a su... |
0970cba4-cb5f-4e7e-bc41-7a78c39e9535 | trentmkelly/LessWrong-43k | LessWrong | Harry Potter and the Methods of Rationality discussion thread, February 2015, chapter 104
New chapter!
This is a new thread to discuss Eliezer Yudkowsky’s Harry Potter and the Methods of Rationality and anything related to it. This thread is intended for discussing chapter 104.
There is a site dedicated to the story... |
b5e7dc1a-43ef-470c-8219-616a7c5a7266 | trentmkelly/LessWrong-43k | LessWrong | Test Driven Thinking
Programmers do something called Test Driven Development. Basically, they write tests that say "I expect my code to do this", then write more code, and if the subsequent code they write breaks a test they wrote, they'll be notified.
Wouldn't it be cool if there was Test Driven Thinking?
1. Write... |
07395809-03e3-4761-84cc-155be864f109 | StampyAI/alignment-research-dataset/distill | Distill Scientific Journal | Weight Banding
This article is part of the [Circuits thread](/2020/circuits/), an experimental format collecting invited short articles and critical commentary delving into the inner workings of neural networks.
[Branch Specialization](/2020/circuits/branch-specialization/)
Introd... |
ecd26ac7-6d49-4b86-adc4-56680d135def | trentmkelly/LessWrong-43k | LessWrong | AI #109: Google Fails Marketing Forever
What if they released the new best LLM, and almost no one noticed?
Google seems to have pulled that off this week with Gemini 2.5 Pro.
It’s a great model, sir. I have a ton of reactions, and it’s 90%+ positive, with a majority of it extremely positive. They cooked.
But what g... |
247b10bb-b882-47bb-b4f9-c4b7ecc59978 | trentmkelly/LessWrong-43k | LessWrong | The Problem With Trolley Problems
A trolley problem is something that's used increasing often in philosophy to get at people's beliefs and debate on them. Here's an example from Wikipedia:
> As before, a trolley is hurtling down a track towards five people. You are on a bridge under which it will pass, and you can st... |
392bbb54-bf68-4bfa-a0a8-e52e7b26ae21 | trentmkelly/LessWrong-43k | LessWrong | an ethical puzzle about brain emulation
I've been thinking about ethics and brain emulations for a while and now have realized I am confused. Here are five scenarios. I am pretty sure the first is morally problematic, and pretty sure the last is completely innocuous. But I can't find a clean way to partition the inte... |
79a5896f-89f2-442d-b201-1dd657dc2968 | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "It’s been over a year since the Singularity Institute launched our ongoing Visiting Fellows Program and we’ve learned a lot in the process of running it. This summer we’re going to try something different. We’re going to run Rationality Boot Camp.
We are going to try to take ten weeks and fill them with activities mea... |
3fba7816-e965-4d28-ba37-044b211d93e9 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | An overview of arguments for concern about automation
*Epistemic status: Fairly confident on the economics, less so on the political science points. This write-up is mainly the result of a class I took on automation and the labor market with professor Nir Jaimovich as well as the book ‘The Technology Trap’ by Carl Ben... |
2f07a714-366f-4960-a893-122a1d0bb666 | trentmkelly/LessWrong-43k | LessWrong | An Oracle standard trick
A putative new idea for AI control; index here.
EDIT: To remind everyone, this method does not entail the Oracle having false beliefs, just behaving as if it did; see here and here.
An idea I thought I'd been mentioning to everyone, but a recent conversation reveals I haven't been assiduous ... |
37f88f21-e594-4584-99b5-317762f68cfd | trentmkelly/LessWrong-43k | LessWrong | Harry Potter and the Data Centers of Doom
The atmosphere was thick with intellectual tension in a small, concealed room at Hogwarts School of Witchcraft and Wizardry. Three prominent figures, each a beacon of knowledge in their own right, were engaged in an intense discussion. Harry Potter, the young prodigy, Alastor ... |
4cf141d7-cfe7-4e80-8ec6-d1ecf2dec2f6 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | [ASoT] Searching for consequentialist structure
*Editor’s note: I’m experimenting with having a lower quality threshold for just posting things even while I’m still confused and unconfident about my conclusions, but with this disclaimer at the top. Thanks to Kyle and Laria for discussions.*
In [“Searching for Bayes S... |
433cfde4-11f6-4b10-a21f-9d855c6be60f | trentmkelly/LessWrong-43k | LessWrong | Appendix: Hamming Questions
Author's note: these questions were usually the prework for an activity called "Hamming Circles," in which four or five people sat down and spent 20-30 minutes with each person's Hamming Question at the center of a slow, intimate discussion. The Hamming Circles activity is not fully descri... |
522263ba-860e-43c6-8c2e-4c40e7fecdc3 | trentmkelly/LessWrong-43k | LessWrong | Contest: $1,000 for good questions to ask to an Oracle AI
Edit: contest closed now, will start assessing the entries.
The contest
I'm offering $1,000 for good questions to ask of AI Oracles. Good questions are those that are safe and useful: that allows us to get information out of the Oracle without increasing risk... |
15bdacd9-17c5-4435-8fae-c381c58f9630 | trentmkelly/LessWrong-43k | LessWrong | [Link] How to Dispel Your Illusions
The topic and the problems associated with it are probably familiar to many of you already. But I think some may find this review by Freeman Dyson of the book Thinking, Fast and Slow by Daniel Kahneman interesting.
> In 1955, when Daniel Kahneman was twenty-one years old, he was ... |
4a6738fe-9dd3-42cd-a972-f700a782a9da | trentmkelly/LessWrong-43k | LessWrong | Truth is Universal: Robust Detection of Lies in LLMs
A short summary of the paper is presented below.
TL;DR: We develop a robust method to detect when an LLM is lying based on the internal model activations, making the following contributions: (i) We demonstrate the existence of a two-dimensional subspace, along whic... |
3b1df9b3-88a2-4c79-8d99-0e5dd6f0c36e | trentmkelly/LessWrong-43k | LessWrong | Inner Optimization Mechanisms in Neural Nets
I believe that current architecture of neural networks supports mesa-optimization: generally speaking, searching across some vectors in order to select one of them, which will be most useful for producing an answer.
Three inner optimization ways are already possible, and m... |
16ddbbbc-c3e3-4ae3-bc64-539a676cf4a7 | trentmkelly/LessWrong-43k | LessWrong | AGI's Impact on Employment
Here is an interesting post about AGI's impact on human employment. The author basically arguing that AGI will increase human employment in the economy. Basically he seems to think that humans will always manage to find work especially in labor-intensive sectors like health care and educati... |
d480f204-b472-4971-8c0a-aec2f69bd2a1 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | RecurrentGPT: a loom-type tool with a twist
This is a link post for [RecurrentGPT: Interactive Generation of (Arbitrarily) Long Text](https://arxiv.org/abs/2305.13304) paper by a team from ETH Zürich ("Swiss Federal Institute of Technology in Zürich").
[The Compleat Cybornaut](https://www.lesswrong.com/posts/iFBdEqE... |
88d8d670-7c82-49e9-bc3b-313e88ef122f | trentmkelly/LessWrong-43k | LessWrong | Notes on countermeasures for exploration hacking (aka sandbagging)
If we naively apply RL to a scheming AI, the AI may be able to systematically get low reward/performance while simultaneously not having this behavior trained out because it intentionally never explores into better behavior. As in, it intentionally put... |
e391b0eb-1832-48d5-b6a7-4438274ed58e | trentmkelly/LessWrong-43k | LessWrong | Takes and loose predictions on AI progress and some key problems
Thanks to Gavin, Kaivu, Kushal, Nikola for giving feedback.
AI is moving fast. I wanted to put down some thoughts on the future and what I believe will be important, moreso on a technical level than a governance/policy perspective. Give me feedback, pus... |
0b72ab7c-dd30-4ad1-9614-3650f0760761 | trentmkelly/LessWrong-43k | LessWrong | Connection Theory Has Less Than No Evidence
I’m a member of the Bay Area Effective Altruist movement. I wanted to make my first post here to share some concerns I have about Leverage Research.
At parties, I often hear Leverage folks claiming they've pretty much solved psychology. They assign credit to their central r... |
1633dcf3-844d-436f-a8b8-ad6935a2d39b | trentmkelly/LessWrong-43k | LessWrong | Weekly LW Meetups: Atlanta, London, Moscow, Vienna
This summary was posted to LW main on April 12th. The following week's summary is here.
There are upcoming irregularly scheduled Less Wrong meetups in:
* Vienna Meetup #2: 13 April 2013 04:00PM
* London Meetup, 14th April: Hedonic hacks: 14 April 2013 02:00PM
* M... |
46f15454-de9b-4d96-9ec4-e25e2ffa752e | trentmkelly/LessWrong-43k | LessWrong | Impartial ethics and personal decisions
Some moral questions I’ve seen discussed here:
* A trolley is about to run over five people, and the only way to prevent that is to push a fat bystander in front of the trolley to stop it. Should I?
* Is it better to allow 3^^^3 people to get a dust speck in their eye, or o... |
16ea1f5b-c916-4943-9d93-360684d30e17 | trentmkelly/LessWrong-43k | LessWrong | It’s Bayes All The Way Up
[Epistemic status: Very speculative. I am not a neuroscientist and apologize for any misinterpretation of the papers involved. Thanks to the people who posted these papers in r/slatestarcodex. See also Mysticism and Pattern-Matching and Bayes For Schizophrenics]
Bayes’ Theorem is an equation... |
7da46989-05d7-4098-8250-266dbcdc649f | trentmkelly/LessWrong-43k | LessWrong | A singularity scenario
Wired Magazine has a story about a giant data center that the USA's National Security Agency is building in Utah, that will be the Google of clandestine information - it will store and analyse all the secret data that the NSA can acquire. The article focuses on the unconstitutionality of the dom... |
610055b7-75a7-472a-912b-5d6d3e10db1f | trentmkelly/LessWrong-43k | LessWrong | Loss of Alignment is not the High-Order Bit for AI Risk
This post aims to convince you that AI alignment risk is over-weighted and over-invested- in.
> A further consideration is that sometimes people argue that all of this futurist speculation about AI is really dumb, and that its errors could be readily explained b... |
2f5a0416-3aec-4e46-a042-cae4ec620a77 | trentmkelly/LessWrong-43k | LessWrong | [Linkpost] Treacherous turns in the wild
Mesa-optimizers potentially escape behavioral adversarial training by realizing they're being tested, gaming the test, then defecting when they're deployed. Luke Muelhauser recounts an example of such a treacherous turn in the wild.
Most relevant quote:
> As part of a project... |
66c7ecce-3514-4932-87b0-f1da854424d3 | trentmkelly/LessWrong-43k | LessWrong | What is the difference between AI misalignment and bad programming?
Hi there,
I am new to both AI and computer programming in general, and have been teaching myself about AI alignment. I am wondering if there is a difference between AI misalignment and "bad" or uninformed programming.
My understanding of AI misalign... |
d9fd4e76-b155-43b9-a6a8-b36ffe965dbd | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Time and Effort Discounting
**Related to:** [Akrasia, hyperbolic discounting, and picoeconomics](/lw/6c/akrasia_hyperbolic_discounting_and_picoeconomics/)
If you're tired of studies where you inevitably get deceived, electric shocked, or tricked into developing a sexual attraction to penny jars, you might want to si... |
20ff5a5a-cb81-4665-a39b-01ac991422d8 | trentmkelly/LessWrong-43k | LessWrong | Reasons why coronavirus mortality of young adults may be underestimated.
TL;DR Our estimation of young adult mortality could be biased because of longer survival time before death in ICU and triage.
The data from Asia is that the mortality of young adults between 20-39 is like 0.2 per cent. This sometimes contradicts... |
8b04b759-f8cf-409c-91b7-98a777a556b6 | trentmkelly/LessWrong-43k | LessWrong | Beta - First Impressions
I thought that instead of everyone having to create a separate post for their first impressions, it would be more convenient to create a single post for this discussion. I'll post my own here soon. |
63aa6a62-3b94-4b4d-8bed-538820d45141 | trentmkelly/LessWrong-43k | LessWrong | Best of LessWrong: Jan 10 - Jan 17
I'm an infovore. The failure mode for an infovore is spending too much time reading and not enough time doing. I was bit especially hard by this a few days ago - I hadn't caught up on LW posts in while and had 100+ in my RSS reader. I wanted to read them all but I knew I shouldn't.
... |
819cb158-8a29-48ea-b97c-5907f9da0e71 | trentmkelly/LessWrong-43k | LessWrong | A List Of Questions & Exercises For Reviewing Your Year
As the year comes to an end, I thought it might be useful to share some questions that might help with reviewing your past year.
This is an adapted post, based on what I have written here.
And, I've put together an interactive suite so you can answer the questi... |
1d37ea96-df19-4aab-8486-1ca395a2eabd | trentmkelly/LessWrong-43k | LessWrong | Submission and dominance among friends
|
a1cf7020-cf31-4f78-b669-c0324a06b725 | trentmkelly/LessWrong-43k | LessWrong | AMA: Paul Christiano, alignment researcher
I'll be running an Ask Me Anything on this post from Friday (April 30) to Saturday (May 1).
If you want to ask something just post a top-level comment; I'll spend at least a day answering questions.
You can find some background about me here. |
14b10e68-aad3-4809-9863-5ffc119b5b48 | StampyAI/alignment-research-dataset/special_docs | Other | It Takes a Village: The Shared Responsibility of 'Raising' an Autonomous Weapon
1 It Takes a Village: The Shared Responsibility of 'Raising' an
Autonomous Weapon
Amritha Jayanti and Shahar Avin
Submitted to and Presented at RSIS Military Transformations Programme Workshop
Co-organised with the Norwegian Institut... |
caae7f3c-4073-4927-b014-4c9cc67a7cd0 | trentmkelly/LessWrong-43k | LessWrong | Pressure to publish increases scientists' vulnerability to positive bias
More evidence for this hypothesis:
> The growing competition and “publish or perish” culture in academia might conflict with the objectivity and integrity of research, because it forces scientists to produce “publishable” results at all costs. P... |
37bcde06-5c9d-4d3c-b10f-2093de5beedb | trentmkelly/LessWrong-43k | LessWrong | Implications of the Grabby Aliens Model
Hanson’s Grabby Aliens Model is one way to explain the apparent lack of other intelligent life in the universe. Essentially, if you believe it’s possible for extraterrestrial civilizations to expand across the galaxy and transform large amounts of matter, then most observers (e.... |
fb905a04-c824-4412-a710-12a7ed7e2545 | trentmkelly/LessWrong-43k | LessWrong | Random Thoughts on Predict-O-Matic
I'm going to be a bit more explicit about some ideas that appeared in The Parable of Predict-O-Matic. (If you don't want spoilers, read it first. Probably you should read it first anyway.)
[Note: while the ideas here are somewhat better than the ideas in the predict-o-matic story, t... |
7a28dfcd-3216-45cf-bc74-19d5e66a0e7b | trentmkelly/LessWrong-43k | LessWrong | Femininity can’t be that bad
Lisa at Sociological Images:
> Most of us are clear on the idea that patriarchies are defined by sexism… In our American patriarchy, however, this is …perhaps even superseded by something called androcentrism: the valuing of all-things-masculine over all-things-feminine. We know we live ... |
6a72c480-f80d-4e36-9381-546e3bec1786 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | (Extremely) Naive Gradient Hacking Doesn't Work
Tldr
====
A super simple form of gradient hacking might be to perform badly on any change to the value of the model's mesa-objective, in order to preserve the mesa-objective. I use some simple properties of SGD to explain why this method of gradient hacking cannot work.... |
89cb42c5-1c04-43c2-98f5-befe8754d476 | trentmkelly/LessWrong-43k | LessWrong | The Unexpected Hanging Paradox
The Unexpected Hanging Paradox has been described as follows:
> A judge tells a condemned prisoner that he will be hanged at noon on one weekday in the following week but that the execution will be a surprise to the prisoner. He will not know the day of the hanging until the executioner... |
cbf1f8e9-7443-4319-946d-a82fc7757882 | trentmkelly/LessWrong-43k | LessWrong | How An Algorithm Feels From Inside
"If a tree falls in the forest, and no one hears it, does it make a sound?" I remember seeing an actual argument get started on this subject—a fully naive argument that went nowhere near Berkeleyan subjectivism. Just:
> "It makes a sound, just like any other falling tree!"
> "But ... |
f51eddfd-27fd-41ad-8998-1dd6a4c7d6af | trentmkelly/LessWrong-43k | LessWrong | Ethan Caballero on Private Scaling Progress
Some quotes from the latest episode of my podcast, The Inside View. You can access the audio, video and transcript here. The key insight is that we are only seeing the tip of the iceberg w.r.t. Large Language Models Scaling, and Alignment can be seen as an inverse scaling pr... |
fc1f8fbf-771c-4dbd-859e-9dee3d4ea34b | StampyAI/alignment-research-dataset/special_docs | Other | EleutherAI Alignment 101
### \*\*EleutherAI Alignment 101\*\* \*\*Curriculum\*\*
Overview
--------
Meeting Times: {Fill-in}
The first iteration of the reading group has [[concluded]{.ul}](https://docs.google.com/document/d/1W-V7hCr9JEnoi\_SG6occL8SztkcmXhW9-FbP7tOki6c/edit?usp=sharing). Future iterations will proba... |
44a6be4f-f898-4ada-993f-792276023c14 | trentmkelly/LessWrong-43k | LessWrong | Do what we mean vs. do what we say
Written quickly after a CHAI meeting on the topic, haven't thought through it in depth.
If we write down an explicit utility function and have an AI optimize that, we expect that a superintelligent AI would end up doing something catastrophic, not because it misunderstands what huma... |
1a51de86-40fd-4415-8666-28e0675f1c4b | trentmkelly/LessWrong-43k | LessWrong | The case for creating unaligned superintelligence
Roko's basilisk states that
> an otherwise benevolent artificial superintelligence (AI) in the future that would punish anyone who knew of its potential existence but did not directly contribute to its advancement or development, in order to incentivize said advanceme... |
6c6297ff-86a7-45e1-9f24-37bd46e82462 | trentmkelly/LessWrong-43k | LessWrong | Why do we need RLHF? Imitation, Inverse RL, and the role of reward
This is a cross-post from my Substack to get some feedback from the alignment community. I have condensed some sections to get the message across.
It's rarely questioned in the mainstream why we need RLHF in the first place? If we can do supervised f... |
e89ffd65-7f4e-4be2-8321-476e6f85a57c | trentmkelly/LessWrong-43k | LessWrong | 3 Interview-like Algorithm Questions for Programmers
These are my favorite interview-style algorithm questions for programmers. I like these questions because they test how well you understand information density instead of your ability to regurgitate facts.
Comments to help clarify puzzle details such as how the Eni... |
d0e39bdc-6cb5-407c-a118-ac25808a5b6b | trentmkelly/LessWrong-43k | LessWrong | Hiring engineers and researchers to help align GPT-3
My team at OpenAI, which works on aligning GPT-3, is hiring ML engineers and researchers. Apply here for the ML engineer role and here for the ML researcher role.
GPT-3 is similar enough to "prosaic" AGI that we can work on key alignment problems without relying on... |
373bd569-d8df-4ad3-855b-0987d6eb8a6b | trentmkelly/LessWrong-43k | LessWrong | Plausibility of cyborgism for protecting boundaries?
Most of my boundaries work so far has been focused on protecting boundaries "from the outside". For example, maybe davidad's OAA could produce some kind of boundary-defending global police AI.
But, imagine parenting a child and protecting them by keeping them insid... |
50470f4a-d835-466b-ad6b-1696185cffca | StampyAI/alignment-research-dataset/blogs | Blogs | ML Safety Newsletter #5
Welcome to the 5th issue of the ML Safety Newsletter. In this special edition, we focus on prizes and competitions:
* ML Safety Workshop Prize
* Trojan Detection Prize
* Forecasting Prize
* Uncertainty Estimation Prize
* Inverse Scaling Prize
* AI Worldview Writing Prize
Awards for Best Paper... |
4eba2bbb-3faf-4de4-9bce-d267e0dc813d | StampyAI/alignment-research-dataset/arbital | Arbital | Union
The union of two sets $A$ and $B$, denoted $A \cup B$, is the set of things which are either in $A$ or in $B$ or both.
Formally stated, where $C = A \cup B$
$$x \in C \leftrightarrow (x \in A \lor x \in B)$$
That is, [https:/... |
fbd01a67-05af-4fac-b33e-fbc534e5bd24 | trentmkelly/LessWrong-43k | LessWrong | Portland SSC Meetup 6/9/18
Come join us for the Saturday, June 9th meetup! We're meeting at 12:30 -- email me for location details. (Email in rot19: ajnygba125@tznvy.pbz) |
02548b59-69ae-4e5f-b69b-b30b57bf3c52 | trentmkelly/LessWrong-43k | LessWrong | Universal agents and utility functions
I'm Anja Heinisch, the new visiting fellow at SI. I've been researching replacing AIXI's reward system with a proper utility function. Here I will describe my AIXI+utility function model, address concerns about restricting the model to bounded or finite utility, and analyze some ... |
16e55089-72cb-44d1-a292-3189c039b0f0 | trentmkelly/LessWrong-43k | LessWrong | The Progress Paradox
LessWrong Intro: I'm new to the LessWrong community, but was inspired to write this based on recent reading of The Sequences. It seems to me there's a paradox to social progress that tends to undermine itself - societies making rapid ethical progress will (necessarily) become ashamed of their rec... |
cee3413e-e862-4d12-a6ea-9193ec403267 | trentmkelly/LessWrong-43k | LessWrong | To contribute to AI safety, consider doing AI research
Among those concerned about risks from advanced AI, I've encountered people who would be interested in a career in AI research, but are worried that doing so would speed up AI capability relative to safety. I think it is a mistake for AI safety proponents to avoid... |
5d4d22d3-25bc-4ad3-8f60-3356984ee384 | trentmkelly/LessWrong-43k | LessWrong | EA, Veganism and Negative Animal Utilitarianism
Veganism is on the face of it an extremely simple moral philosophy. If Animals have feelings, then eating animals is bad. Duh.
Effective Altruists tend to be people who actually take moral arguments at face value rather than ignoring them, and unsurprisingly many effect... |
428ccd38-2911-4358-af09-dee65acd5b8f | trentmkelly/LessWrong-43k | LessWrong | Chapter 19: Delayed Gratification
Blood for the blood god! Skulls for J. K. Rowling!
----------------------------------------
Draco had a stern expression on his face, and his green-trimmed robes somehow looked far more formal, serious, and well-turned-out than the same exact robes as worn by the two boys behind him... |
984c8da8-69de-4e1a-8289-0f5059a6eca2 | trentmkelly/LessWrong-43k | LessWrong | Synthetic Neuroscience
Abstract: Neuroscience as engineering -- understanding how the brain works by building simplified, effective models of it, while requiring the model to behave as a real animal would.
Motivation
Suppose we want to build an AGI. The only working human-level general intelligence we know of is ... |
67d7386b-28c5-49a7-ae2b-f685e9a8b2e1 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Meta-Learning with Implicit Gradients
1 Introduction
---------------
A core aspect of intelligence is the ability to quickly learn new tasks by drawing upon prior experience from related tasks. Recent work has studied how meta-learning algorithms [[51](#bib.bib99 "Evolutionary principles in self-referential learnin... |
2be47c55-0584-4ae0-8d8e-d364d3243b77 | trentmkelly/LessWrong-43k | LessWrong | FHI paper published in Science: interventions against COVID-19
This is a linkpost for https://science.sciencemag.org/lookup/doi/10.1126/science.abd9338
Summary: this is basically a Bayesian regression of the COVID-19 cases and deaths in 41 countries against the measures taken by those countries. It requires a rough m... |
307a2546-88b0-458d-8516-248abb9c2ebd | trentmkelly/LessWrong-43k | LessWrong | Longtermist Implications of the Existence Neutrality Hypothesis
Crossposted on the EA Forum.
We describe the implications the Existence Neutrality Hypothesis[1] could have for impartial longtermists, and then provide quantitative impact estimates for four idealized interventions when accounting or not for this hypoth... |
4ee53fdb-05e2-49e7-aed6-934a2ee12485 | trentmkelly/LessWrong-43k | LessWrong | Review of 'Debate on Instrumental Convergence between LeCun, Russell, Bengio, Zador, and More'
I think that 'robust instrumentality' is a more apt name for 'instrumental convergence.' That said, for backwards compatibility, this post often uses the latter.
In the summer of 2019, I was building up a corpus of basic r... |
dea7985c-7d61-4ec3-871f-97c9006ce274 | trentmkelly/LessWrong-43k | LessWrong | Building Phenomenological Bridges
Naturalized induction is an open problem in Friendly Artificial Intelligence (OPFAI). The problem, in brief: Our current leading models of induction do not allow reasoners to treat their own computations as processes in the world.
The problem's roots lie in algorithmic information t... |
b6c1c3c0-c8f4-4338-a5b3-8ac2cf37aa4e | StampyAI/alignment-research-dataset/blogs | Blogs | don't censor yourself, silly !
don't censor yourself, silly !
------------------------------
there's this meme that, as you grow up, you learn that adults don't really know what they're doing either, they've just become good at pretending they do. i think the meme holds up; you'd be surprised how many people — even ... |
4d572e7c-f6bb-4ccc-81f7-f714512c2826 | trentmkelly/LessWrong-43k | LessWrong | Housing Without Street Parking: Implemented
In 2017, I wrote:
> What if, in places like Somerville where all parking is already by-permit-only, we added a new category of housing unit, one that didn't come with any rights to street parking?
It turns out this was included in the 2019 zoning overhaul (I missed this am... |
d23a8220-0c71-4f08-ba19-292fe1e7edd7 | trentmkelly/LessWrong-43k | LessWrong | Long Term Charities: Apply for SFF Funding
In the last funding round of the Survival and Flourishing Fund (SFF), I was one of the recommenders whose evaluations helped distribute millions of dollars in charitable grants. I wrote up my experiences here.
Now that applications for the next round are open, I want to stro... |
4b2527cf-f611-4122-ba7e-42e28126a5a1 | trentmkelly/LessWrong-43k | LessWrong | Meetup : RTLW Meetup: Contra Dance!
Discussion article for the meetup : RTLW Meetup: Contra Dance!
WHEN: 23 May 2013 06:50:00PM (-0400)
WHERE: 100 N Greensboro St, Carrboro NC 27510
In lieu of our regular coffee and discussion, this week's meetup will be held at the Century Center in Carrboro, where A GREAT DANCE w... |
b2a1c55a-9a01-41d8-9b5a-346609866678 | trentmkelly/LessWrong-43k | LessWrong | Priors and Surprise
I don’t want to be too dogmatic about this claim, but Godzilla is unrealistic. I don’t want to be too non-dogmatic about this claim either. OK then, just how dogmatic should I be? I have all sorts of reasons for thinking that skyscraper sized lizards or dinosaurs don’t actually exist. Honestly,... |
c200f8cb-b562-49ad-8f2f-1873dea98b39 | trentmkelly/LessWrong-43k | LessWrong | "I Don't Know" (2020 Solstice Speech)
This was my primary speech at this year's Solstice. I have meta-commentary on it I will share in a bit.
The virtue of darkness is to face reality at its harshest, without looking away.
There are many forms of metaphorical darkness. But perhaps the most primal is to Not Know Thi... |
99004678-02f8-4194-ac66-e1f68fad561d | trentmkelly/LessWrong-43k | LessWrong | Forecasting an 80% chance of an effective anti COVID-19 drug (probably Remdesivir)
I'm gathering info on anti-COVID-19 drug efforts. Here is what I've compiled so far. I've combined the number of trials in various phases with big data on expected success rates for drugs in those phases to forecast an 80% probability t... |
fe87ed61-0db3-44f6-8249-c5e8a3eee990 | StampyAI/alignment-research-dataset/arbital | Arbital | Arithmetical hierarchy: If you don't read logic
The arithmetical hierarchy is a way of stratifying statements by how many "for every number" and "there exists a number" clauses they contain.
Suppose we say "2 + 1 = 1 + 2". Since this is only a statement about three specific numbers, this statement would occupy the l... |
239dabb0-cf88-4516-9ccf-baf741092944 | trentmkelly/LessWrong-43k | LessWrong | Which Fields Are Underserved?
I'm currently about 2/3rds through Jane Jacobs' "The Death and Life of Great American Cities". This is one of the defining works of modern urban planning, and Jane Jacobs is considered one of our most important urban planning thinkers.
It's a fine book, but I found myself surprised at ju... |
824f059b-d2f7-4408-8992-6a9394e0a823 | trentmkelly/LessWrong-43k | LessWrong | June gwern.net newsletter
None |
577c7904-8541-429f-80f3-d4fcff151608 | trentmkelly/LessWrong-43k | LessWrong | Why aren't assurance contracts widely used?
A priori, dominant assurance contracts seem like awesome tools for solving a fairly broad range of collective action problems. Why aren't they used much? Or is it just that they are a new idea and we should expect them to grow in prominence in the next few decades?
|
3b5b47ba-d9fb-43fb-8f28-0954206ee260 | StampyAI/alignment-research-dataset/special_docs | Other | Planning for Autonomous Cars that Leverage Effects on Human Actions
Planning for Autonomous Cars
that Leverage Effects on Human Actions
Dorsa Sadigh, Shankar Sastry, Sanjit A. Seshia, and Anca D. Dragan
University of California, Berkeley, {dsadigh, sastry, sseshia, anca}@eecs.berkeley.edu
Abstract —Traditionally, auto... |
0001171c-7095-4bdc-82aa-beab9c129b76 | trentmkelly/LessWrong-43k | LessWrong | A spreadsheet/template for doing an annual review
[Sharing this as a forum post because it may be useful for some forum users - mostly a copy and paste of a social media post]
🤔 Are you looking for a way to review your year and plan for the future?
Many people I know do an "annual review".
When I started in 2018, ... |
e2fc0598-fef3-4473-acf5-e00ed2ab5163 | trentmkelly/LessWrong-43k | LessWrong | Bandwagon effect: Bias in Evaluating AGI X-Risks
Note on unusual formatting: Sentences are split into lines so you can parse parts precisely.
The tendency to do (or believe) things
because many other people do (or believe) the same.
While this is similar in effect
to the 'Availability Cascade',
this b... |
aafa3987-278c-47d1-98ab-99c5535a2a6e | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Little attention seems to be on discouraging hardware progress
There is a lot of attention being given to AI software progress, potential regulation etc., however articles such as [this](https://www.allaboutcircuits.com/news/interface-type-memristive-device-pushes-neuromorphic-computing-onward/) celebrating potential ... |
ffe289ff-029e-4325-9478-f13dbf5d4b5e | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Birds, Brains, Planes, and AI: Against Appeals to the Complexity/Mysteriousness/Efficiency of the Brain
*[Epistemic status: Strong opinions lightly held, this time with a cool graph.]*
I argue that an entire class of common arguments against short timelines is bogus, and provide weak evidence that anchoring to the hu... |
46a4969c-3bad-4a83-b11c-c0fdb557ed7b | trentmkelly/LessWrong-43k | LessWrong | Meetup : Saskatoon - Gauging the strength of evidence and Bayesian reasoning.
Discussion article for the meetup : Saskatoon - Gauging the strength of evidence and Bayesian reasoning.
WHEN: 23 November 2013 01:00:00PM (-0600)
WHERE: 2318 8th St E, Saskatoon, SK
Another Saskatoon meetup at the same place and time as ... |
0402fead-9a2c-4fd1-bdf5-bfd4623c2b93 | StampyAI/alignment-research-dataset/special_docs | Other | CHAI Newsletter #4 2019
5/4/22, 3:36 PM CHAI Q4 Newsletter + Intern Applications Due Soon
https://mailchi.mp/13c99f6887ba/chai-q4-newsletter -intern-applications-due-soon 1/7Human Compatible: Artificial Intelligence
and the Problem of Control
Stuart Russell's new book, Human Compatible: Artificial Intelligence and t... |
30a50479-577b-4a0c-b76f-02a16bdbf384 | trentmkelly/LessWrong-43k | LessWrong | Should I Divest from AI?
In recent years, AI has been all the rage in the stock market, and there is no reason to see that slowing down. With the picture of AGI on the horizon becoming clearer and clearer, faster and smarter models being released, and more and more investment being poured into AI stocks, it seems inev... |
37e2896a-ff0e-48f1-8622-85c7bb852264 | trentmkelly/LessWrong-43k | LessWrong | How do bounded utility functions work if you are uncertain how close to the bound your utility is?
If you are trying to calculate the value of a choice using a bounded utility function, how can you be sure whether you are close or far from the bound, whatever the bound is? How do you account for uncertainty about ho... |
134b0373-a470-4d51-a06d-8258a0ff1110 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Solving the Mechanistic Interpretability challenges: EIS VII Challenge 2
We solved\* the second Mechanistic Interpretability challenge (Transformer) that [Stephen Casper](https://www.alignmentforum.org/users/scasper) posed in [EIS VII](https://www.alignmentforum.org/posts/KSHqLzQscwJnv44T8/eis-vii-a-challenge-for-mech... |
5688c27d-c54d-4844-b610-54d0a93bc9fc | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Four usages of "loss" in AI
*Summary:* What does it *mean* for a loss function to be "aligned with" human goals? I perceive four different concepts which involve "loss function" in importantly different ways:
1. *Physical-loss:* The physical implementation of a loss function and the loss computations,
2. *Mathematica... |
642c224f-39fd-473d-bc69-055176f00e83 | trentmkelly/LessWrong-43k | LessWrong | First and Last Questions for GPT-5*
Suppose that in the very near future, a research group finds that their conversational AI has begun to produce extremely high-quality answers to questions. There's no obvious limit to its ability, but there's also no guarantee of correctness or good intention, given the opacity of h... |
8c730009-f2d0-4294-9d23-c652e8eefce6 | trentmkelly/LessWrong-43k | LessWrong | Should we stop using the term 'Rationalist'?
This might be an unpopular opinion, but I really dislike the term 'rationalist' for four reasons:
1) It makes you sound self-aggrandizing. The term gives the impression that you think you are already rational and therefore smarter-than-thou.
2) It's already a term used to... |
ee76476b-7994-41b5-a6c5-6d68c3712b49 | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post3865
A putative new idea for AI control; index here . It seems that stratification is more flexible than I initially thought. That's because the default action or policy ∅ , which I was envisioning as a null action (or maybe the AI turning itself off) can actually be more than that. For instance, ∅ could be an... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.