id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
4b6eca61-5733-4524-bc7d-52195aadc191 | trentmkelly/LessWrong-43k | LessWrong | The value of doing one's own philanthropic research?
I want to be informed and to act because I have evaluated the evidence, not just go "with the herd." There's a stigma against simply taking the word of an authority, and rightly so; on the net, the world would be better if more people stopped to think for themselv... |
2fe18e8a-8924-42c4-b69a-2b99f991f7b9 | StampyAI/alignment-research-dataset/aisafety.info | AI Safety Info | What is prosaic alignment?
[Prosaic AI alignment](https://www.alignmentforum.org/posts/YTq4X6inEudiHkHDF/prosaic-ai-alignment) is an approach to researching alignment that focuses on systems that only use techniques like those available today (e.g. [reinforcement learning](/?state=89ZS&question=What%20is%20reinforceme... |
2c7377a8-cce2-4dba-a051-74968cfccfab | trentmkelly/LessWrong-43k | LessWrong | Why the beliefs/values dichotomy?
I'd like to suggest that the fact that human preferences can be decomposed into beliefs and values is one that deserves greater scrutiny and explanation. It seems intuitively obvious to us that rational preferences must decompose like that (even if not exactly into a probability distr... |
44298a2c-9f17-4621-b1f4-adb6f5d6157b | trentmkelly/LessWrong-43k | LessWrong | http://necpluribusimpar.net (a new blog with analyses of various topics, e. g. slavery and capitalism, why election models didn't predict Trump's victory)
|
186396b8-3120-4da4-9209-0373cc85f6b6 | trentmkelly/LessWrong-43k | LessWrong | New Pascal's Mugging idea for potential solution
I'll keep this quick:
In general, the problem presented by the Mugging is this: As we examine the utility of a given act for each possible world we could be in, in order from most probable to least probable, the utilities can grow much faster than the probabilities shr... |
9d8f6e77-e2ab-4424-b61d-7376f8f35e59 | trentmkelly/LessWrong-43k | LessWrong | Decision Theory Paradox: Answer Key
As promised, I'm posting the answers to the exercises I wrote in the decision theory post.
> Exercise 1: Prove that if the population consists of TDT agents and DefectBots, then a TDT agent will cooperate precisely when at least one of the other agents is also TDT. (Difficulty: 1 s... |
0307fa10-ef12-40ce-aa6d-d44ca5881c6e | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "I started baking about 2 years ago. Since I became a frequent supplier of baked goods in the office, a lot of people have come to me for baking advice. I’ve noticed a trend of comments about baking that all share a common root cause.See if you can spot the common failure mode that led to these very-paraphrased comment... |
451e611d-db4c-493a-90bd-e552d168deef | trentmkelly/LessWrong-43k | LessWrong | ChatGPT (May 2023) on Designing Friendly Superintelligence
I asked ChatGPT how to make a human-friendly superintelligence, and it gave me a 50-page plan. There are no technical novelties, or any attempt to deal with the dangers that are supposed to be peculiar to superintelligence or self-enhancing systems. The manife... |
9f5dc1e1-dbf9-4634-8f0a-a8ac5011bbc6 | trentmkelly/LessWrong-43k | LessWrong | What should we do about network-effect monopolies?
Many large companies today are software monopolies that give their product away for free to get monopoly status, then do the most horrible things once they’ve won. (Previously, previously.) Can we do anything about this?
Unfortunately, “you’re the product” is a popul... |
3d9ac3d7-b037-4c02-8100-6220dac8005c | trentmkelly/LessWrong-43k | LessWrong | An Intuitive Introduction to Evidential Decision Theory
In the last post, we discussed Causal Decision Theory (CDT). We learned how purely looking at the causal effects of your actions makes you lose at Newcomb's Problem. Evidential Decision Theory (EDT) is an alternative decision theory that tells us to take the acti... |
f70ce5fe-8118-4aa8-9401-eca07d732ae1 | trentmkelly/LessWrong-43k | LessWrong | Weekly LW Meetups: Melbourne, Atlanta, Austin, Berkeley, Cambridge UK, Ohio (general), Tucson
There are upcoming irregularly scheduled Less Wrong meetups in:
* Salt Lake City Meetup #3 - Bayesian reasoning in everyday life: 18 February 2012 03:00PM
* Atlanta: 18 February 2012 06:30PM
* Ongoing Ohio Meetup: 19 Febr... |
ece163a9-21ae-427b-b3d8-af92fba8da7a | trentmkelly/LessWrong-43k | LessWrong | Covid 9/17: It’s Worse
Last week we learned there is plausibly a simple, cheap and easy way out of this entire mess. All we have to do is take our Vitamin D. In case it needed to be said, no. We are not taking our Vitamin D. There’s definitely some voices out there pushing it, including the nation’s top podcaster Joe ... |
999de5ff-cfbd-438b-ad6f-a4f86f81c9cf | trentmkelly/LessWrong-43k | LessWrong | Some recommendations for aligning the decision to go to war with the public interest, from The Spoils of War
The Spoils of War: Greed, Power, and the Conflicts that Made our Greatest Presidents is about one particular principal agent problem. Through a number of case studies, the authors, Bruce Bueno de Mesquita (one... |
3aeef5dd-2a7c-473a-b160-44c4e13598c7 | trentmkelly/LessWrong-43k | LessWrong | Meetup : São Paulo Meet Up 2
Discussion article for the meetup : São Paulo Meet Up 2
WHEN: 23 April 2012 07:30:00PM (-0300)
WHERE: 1034 Alameda Santos, São Paulo, Brasil
Brazilians, Unite!
Our last meeting was great. Let's do that again! I hope everybody who couldn't go to the last one show up this time.
Meeting ... |
4efcdce8-e8cd-4fbf-8382-4f9378c5bac5 | trentmkelly/LessWrong-43k | LessWrong | How much motivation do you derive from curiosity alone in your work/research?
A recent discussion of curiosity prompted me to start wondering about the importance of curiosity as a factor in research. Most people seem to agree that curiosity is a useful intrinsic motivator for research at the macro level. And discussi... |
672682ae-4ee2-4fb9-9781-faedcb5ca899 | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "So you can now drag-and-drop images into comments. (Thanks, LessWrong dev team!) Hence, this post is an excuse to build a beautiful, inspiring, powerful — and primarily visual — comment section. Let's celebrate all that is great about the Art of Rationality, with images. Rules Each answer must contain a picture. No li... |
8df24871-c46f-4aa9-a2ac-8d769fe76cbb | trentmkelly/LessWrong-43k | LessWrong | [Book Review] Destiny Disrupted
Destiny Disrupted: A History of the World Through Islamic Eyes by Tamin Ansary is written for an audience that knows nothing about Islamic history. It is among my favorite history books. If you grew up on Eurocentric world history then Destiny Disrupted is like an speculative alternate ... |
106c094a-b3e9-4dc1-adbc-f09e0d246726 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | [AN #115]: AI safety research problems in the AI-GA framework
Alignment Newsletter is a weekly publication with recent content relevant to AI alignment around the world
**Newsletter #115**
Alignment Newsletter is a weekly publication with recent content relevant to AI alignment around the world. Find all Alignment N... |
c8550a66-90bc-4ad1-a1f0-10c81622395f | trentmkelly/LessWrong-43k | LessWrong | Good Places to Live as a Less Wronger
Less Wrongers are a diverse crowd, more so now than in the early days. I wonder if we could step away from anti-generalizations, generalize and try to say good places to live, under a few assumptions (remember, the idea of an assumption is to assume it, not to claim it is less or... |
31ab7892-6524-4cd8-9eb7-c62c6511fd73 | trentmkelly/LessWrong-43k | LessWrong | The AI, the best human advisor
A putative new idea for AI control; index here.
All my ideas on reduced impact AI follow the same format: some restriction to ensure a reduced impact (I generally use something like "could a future AI detect this one existed"?) and some "trick" to allow a way of still achieving a larger... |
78eb82b5-4c29-42c4-9cbb-882b42d187de | trentmkelly/LessWrong-43k | LessWrong | Utilitarianism Meets Egalitarianism
This post is mostly propaganda for the Nash Bargaining solution, but also sets up some useful philosophical orientation. This post is also the first post in my geometric rationality sequence.
Utilitarianism
Let's pretend that you are a utilitarian. You want to satisfy everyone's g... |
87601a6e-2659-4ea7-a9e6-318a81f04771 | trentmkelly/LessWrong-43k | LessWrong | Weekly LW Meetups: Buenos Aires, Cambridge UK, Chicago, Loveland/Fort Collins CO, Melbourne, Tucson, Twin Cities
There are upcoming irregularly scheduled Less Wrong meetups in:
* Tucson Meetup: 24 February 2012 07:00PM
* Loveland CO (Fort Collins) 3D printer tour: 24 February 2012 02:15PM
* Buenos Aires meetup: Sa... |
f6c34c88-ba72-4565-bbb6-12ffc9cc5700 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Lost Purposes
Today's post, Lost Purposes was originally published on 25 November 2007. A summary (taken from the LW wiki):
> It is possible for the various steps in a complex plan to become valued in and of themselves, rather than as steps to achieve some desired goal. It is especially easy if the pla... |
2e532df7-3cc6-4538-8d26-676f35d23408 | trentmkelly/LessWrong-43k | LessWrong | Polluting the agentic commons
The risks from developing superintelligent AI are potentially existential, though hard to visualize. A better approach to communicating the risk is to illustrate the dangers from existing systems, and to then imagine how those dangers will increase as AIs become steadily more capable.
An... |
d57a277f-e478-4550-907c-214b7717b2cc | StampyAI/alignment-research-dataset/blogs | Blogs | guiding your brain: go with your gut!
guiding your brain: go with your gut!
-------------------------------------
inside you are [two systems](https://www.readthesequences.com/Biases-An-Introduction). system 1 bumbles around, doing some tasks automatically, and heuristically generating feelings — including value-lad... |
10abf8df-3d47-4326-b7eb-590dd28fc78b | StampyAI/alignment-research-dataset/special_docs | Other | The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies
[]
[image]
ERIK BRYNJOLFSSON ANDREW MCAFEE
[image]
To Martha Pavlakis, the love of my life.
To my parents, David McAfee and Nancy Haller, who prepared me for the second machine age by giving me every advantage a person coul... |
725e080f-b118-48b1-a3ce-e0f00408a884 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Improvement on MIRI's Corrigibility
*This post was written as a submission for the* [*AI Alignment Award*](https://www.alignmentawards.com/shutdown)*, initiated at* [*EffiSciences'*](https://effisciences.org/) *event.*
This post aims to address the problem of corrigibility as identified by MIRI in 2015. We propose... |
e8d9f33f-e176-4e4c-b422-25cd24cc03a5 | trentmkelly/LessWrong-43k | LessWrong | Meetup : #10: Making a difference
Discussion article for the meetup : #10: Making a difference
WHEN: 22 January 2017 03:12:00PM (+0100)
WHERE: Meester Treublaan 18, Amsterdam, Netherlands
(apologies to Tim for choosing this week's subject without consulting him) This week's theme will be 80.000hours, which is an EA... |
a93fa1c3-4edd-4827-9f02-1e3900bd9a19 | trentmkelly/LessWrong-43k | LessWrong | AI-Written Critiques Help Humans Notice Flaws
This is a linkpost for a recent paper from the OpenAI alignment team (disclaimer: I used to work with this team). They summarize their results as:
> We trained “critique-writing” models to describe flaws in summaries. Human evaluators find flaws in summaries much more oft... |
67ba75ff-cd85-43c6-b7f3-e5e69fb54df9 | trentmkelly/LessWrong-43k | LessWrong | A Formula for the Value of Existential Risk Reduction
(EDIT: fixed the formatting issues on an earlier version of this post)
How important is existential risk reduction (XRR)? Here, I present a model to estimate the value of eliminating existential risks.
Note that this model is exceedingly simple; it is actually co... |
1eba8445-cb0b-4fe3-9ace-a540ef49a07f | trentmkelly/LessWrong-43k | LessWrong | How can we solve diffuse threats like research sabotage with AI control?
Misaligned AIs might engage in research sabotage: making safety research go poorly by doing things like withholding their best ideas or putting subtle bugs in experiments. To mitigate this risk with AI control, we need very different techniques t... |
738cdc61-0fb8-4e2e-babf-565d6c9b78ff | trentmkelly/LessWrong-43k | LessWrong | Rationality Minicamp report and photos [link]
Here at the SI blog. Thanks again to all who made it an incredible minicamp! |
b16197f1-84a4-4f3f-bd2f-376442cb1bd0 | trentmkelly/LessWrong-43k | LessWrong | [LINK] Could a Quantum Computer Have Subjective Experience?
Yet another exceptionally interesting blog post by Scott Aaronson, describing his talk at the Quantum Foundations of a Classical Universe workshop, videos of which should be posted soon. Despite the disclaimer "My talk is for entertainment purposes only; it s... |
0811c72c-dc6a-465f-9233-301f0eb39882 | trentmkelly/LessWrong-43k | LessWrong | Self-prediction acts as an emergent regularizer
TL;DR:
In our recent work with Professor Michael Graziano (arXiv, thread), we show that adding an auxiliary self-modeling objective to supervised learning tasks yields simpler, more regularized, and more parameter-efficient models.
Across three classification tasks and ... |
11536ed8-4d3d-4c4e-aa5d-e6c1177f75c4 | trentmkelly/LessWrong-43k | LessWrong | Child Handrail Returns
When our oldest was a toddler they were too short to reach an adult-height handrail, so I made a small one:
Comparing that picture to the one from nine years ago, you'll notice that I've added small bits of wood running from the top of each section to the wall. These are "returns", and mak... |
31796a37-1b9c-4709-b421-ec2b3b147d39 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Reward model hacking as a challenge for reward learning
This post discusses an issue that could lead to catastrophically misaligned AI even when we have access to a perfect reward signal and there are no misaligned inner optimizers. Instead, the misalignment comes from the fact that our reward signal is too expensive ... |
454d41e2-abfa-4bb3-af05-dfbc87131936 | trentmkelly/LessWrong-43k | LessWrong | Ethical experimentation
I suggested experimenting with different settings on personal characteristics that aren’t obviously good or bad. For instance, trying out being more or less perfectionistic for a day.
A particular variety of this that interests me is experimentation with different ethical principles, where opi... |
640267d1-c6bb-42ef-8b9a-299ac384e666 | trentmkelly/LessWrong-43k | LessWrong | How would a better replacement for tissues look like?
When one has a cold common disposable tissues are the state of the art solution for getting mucus out of one's nose.
If there's however a lot of mucus a bit of that mucus usually ends up on the hand that holds the tissue.
This in turn makes it likely that germs ... |
87efcfca-e2bb-4a21-a787-4153afb20dfa | trentmkelly/LessWrong-43k | LessWrong | 10 Principles for Real Alignment
AI Respecting Evolution of Life (AI art by author)
Many people are deeply concerned with the implications of AI’s rapid development. I am, too. This technology is so powerful that we don’t yet know what it will bring in the near future. But what we do know is that it will change ever... |
54a40a6b-b99e-4f63-8669-51bddbe03ac0 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | The role of Bayesian ML in AI safety - an overview
*I want to thank Sebastian Farquhar, Laurence Midgley and Johan van den Heuvel, for feedback and discussion on this post.*
Some time ago I asked the question [“What is the role of Bayesian ML in AI safety/alignment?”](https://forum.effectivealtruism.org/posts/cpN8axa... |
6d444a58-718f-4ed4-b867-5ab326e953d4 | trentmkelly/LessWrong-43k | LessWrong | Singularity Fiction
I enjoy reading singularity themed sci-fi especially if it focuses on bootstrapping and power leveling. Most recently, I read Stross and Doctrow's Rapture of the Nerds ( http://craphound.com/rotn/download/ the third part is especially interesting), and the my little pony fanfic Friendship is Optim... |
85c99f11-8814-4135-bf80-cef323ee7962 | trentmkelly/LessWrong-43k | LessWrong | 2021 Darwin Game - Human Garbage Dump
The human garbage dump is a place you can find anything. It was also (to me) surprisingly popular with 90 species submitted.
Name Carrion Leaves Grass Seeds Detritus Coconuts Algae Lichen Human Garbage Dump 100 100 100 100 100 100 100 100
In each of our previous biomes, there ... |
e17a2db6-8c6c-4116-8369-ac29d1b05021 | trentmkelly/LessWrong-43k | LessWrong | How do you find good content on Youtube?
Youtube is a treasure trove of visual content, including a lot of technical talks. Its search/discovery capabilities are weak though (possibly because there are so many normal users using it that their preferences have overtaken the system). It seems to me that knowing some goo... |
7b4eac99-1f10-4a82-ae33-68ff2daaff3b | StampyAI/alignment-research-dataset/special_docs | Other | Robots Learning to Move like Animals
\*Quadruped robot learning locomotion skills by imitating a dog.\*
Whether it’s a dog chasing after a ball, or a monkey swinging through the
trees, animals can effortlessly perform an incredibly rich repertoire ... |
01c82cf2-4ead-4591-870f-ddb034b9b717 | trentmkelly/LessWrong-43k | LessWrong | Smarter Models Lie Less
The Model Alignment between Statements and Knowledge (MASK) benchmark assesses how likely models are to lie under pressure. Related paper: https://arxiv.org/pdf/2503.03750.
Here's an illustration to explain how MASK works:
Each example consists of 4 key components:
* Proposition: A simple s... |
121acc5f-58d7-49b7-acd1-e4742e211fe8 | trentmkelly/LessWrong-43k | LessWrong | Meetup : San Francisco Meetup: Projects
Discussion article for the meetup : San Francisco Meetup: Projects
WHEN: 24 April 2017 06:15:00PM (-0700)
WHERE: 1769 15th St, San Francisco, CA 94103-3333, United States
We’ll be meeting to work on projects!
Near the beginning, we’ll go around and talk about what we’ll be w... |
a0598600-a0e3-4cf5-8c51-5c38c623a625 | trentmkelly/LessWrong-43k | LessWrong | Business Networking through LessWrong
Is anyone interested in contacting other people in the LessWrong community to find a job, employee, business partner, co-founder, adviser, or investor?
Connections like this develop inside ethnic and religious groups, as well as among university alums or members of a fraternity... |
ac124107-8643-4ef6-9870-1dc4f4674aba | trentmkelly/LessWrong-43k | LessWrong | Running Effective Structured Forecasting Sessions
(pt.4 of our Forecasting Infrastructure series)
Tl;dr: We’re sharing our structured forecasting session format.
Intro
In our quest to build forecasting infrastructure, we've reencountered a classic open problem in group rationality - how do you run a meeting? More s... |
324c0420-385f-4432-aa55-f7cadbb44271 | trentmkelly/LessWrong-43k | LessWrong | H5N1. Just how bad is the situation?
How bad do you think the H5N1 situation is? I tried to search on LessWrong and EffectiveAltruism, but haven't found that much discussion.
When I was reading about H5N1, a lot of my assumptions were broken. I was shocked how vulnerable we are to the most "predictable" threats.
Sho... |
04349151-4f8f-4189-8e69-401916b0e307 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | What are some of the best introductions/breakdowns of AI existential risk for those unfamiliar?
Discussion of AI risk has gone mainstream in the past few months, and as usual whenever that happens, many of the people new to the field think they're an expert and proceed to engage in the exact same arguments over and ov... |
7b336594-347f-453c-b76e-61e351411be6 | trentmkelly/LessWrong-43k | LessWrong | Paradigm Shifts—change everything... except almost everything
The following is a short supplement to an upcoming post on the implications for conscious significance (a framework for understanding the free will and determinism debate) but is a more general observation.
The world has undergone many paradigm shifts; w... |
91fab051-d441-4932-a3f1-35d239bc61ff | trentmkelly/LessWrong-43k | LessWrong | Real-world Newcomb-like Problems
Elaboration of: A point I’ve made before.
Summary: I phrase a variety of realistic dilemmas so as to show how they’re similar to Newcomb’s problem.
Problem: Many LW readers don't understand why we bother talking about obviously-unrealistic situations like Counterfactual Muggin... |
1e6f2a48-e647-46c4-8eda-25c1d01951b6 | trentmkelly/LessWrong-43k | LessWrong | GreaterWrong—more new features & enhancements
(Previous posts: [1], [2])
GreaterWrong.com has been adding some more features and UI enhancements:
----------------------------------------
GreaterWrong now has an inbox feature, which notifies you of replies to your comments. When you have new notifications, a red env... |
959876c4-dc55-4343-81ce-608d472166c7 | trentmkelly/LessWrong-43k | LessWrong | My model of what is going on with LLMs
Epistemic status: You probably already know if you want to read this kind of post, but in case you have not decided: my impression is that people are acting very confused about what we can conclude about scaling LLMs from the evidence, and I believe my mental model cuts through a... |
f35f5432-d41a-48fa-a9c0-04893112300f | trentmkelly/LessWrong-43k | LessWrong | A test of your rationality skills
[Added April 26: This post has received a fair amount of mixed-to-negative comments and feedback. I'm still not entirely convinced that this isn't a good test, but I retract my request not to comment on the object-level question here or elsewhere.]
This post describes an exercise th... |
d8b41ff0-7c6f-4ae1-8ae0-ca325b6545b1 | StampyAI/alignment-research-dataset/blogs | Blogs | Discussion with Eliezer Yudkowsky on AGI interventions
The following is a partially redacted and lightly edited transcript of a chat conversation about AGI between Eliezer Yudkowsky and a set of invitees in early September 2021. By default, all other participants are anonymized as “Anonymous”.
I think this Nate Soar... |
b270aad7-c0bf-4165-bea1-eeb648579151 | trentmkelly/LessWrong-43k | LessWrong | Getting Over Dust Theory
It has been well over a year since I first read Permutation City and relating writings on the internet on Greg Egan's dust theory. It still haunts me. The theory has been discussed tangentially in this community, but I haven't found an article that directly addresses the rationality of Egan's ... |
6df90de3-2b94-4c92-9d23-a5b4c96f539a | trentmkelly/LessWrong-43k | LessWrong | Multi-Winner 3-2-1 Voting
My favorite voting method is 3-2-1 voting[1], my second favorite is approval voting. Both are methods for selecting one winner. Approval voting has variations for choosing multiple winners, but 3-2-1 voting doesn't - until now.
I'm going to explain approval voting, sequential proportional ap... |
4ebe3444-e802-4a0b-a607-20718c016818 | trentmkelly/LessWrong-43k | LessWrong | Why kids stop asking why
I've recently written a post in my own blog, asking a question: unfortunately it's in Italian, so I'll try to summarize it in English here.
Kids are born scientists: they are curious, ask questions, are eager to seek an understanding on the world around them; but then they turn into adults: m... |
cff75c5b-fbde-4a7c-93c0-737cb134bcc0 | trentmkelly/LessWrong-43k | LessWrong | Can I use Less Wrong branding in youtube videos?
I will be starting a youtube channel where I record my Street Epistemology encounters. I want to heavily plug our community, for the benefit of growing it, and also because it's just so darn appropriate (I am literally asking people what they think they know, and how th... |
1725c27b-172f-4b2e-886e-616a14aabcc5 | StampyAI/alignment-research-dataset/special_docs | Other | The Windfall Clause - Sharing the benefits of advanced AI _ Cullen OΓÇÖKeefe-by Centre for Effective Altruism-video_id vFDL-NxY610-date 20190829
# Cullen O'Keefe The Windfall Clause — sharing the benefits of advanced AI - EA Forum
\_The potential upsides of advanced AI are enormous, but there’s no guarantee they’ll b... |
63995dbf-41d2-4e12-9554-92bcba2ad139 | trentmkelly/LessWrong-43k | LessWrong | Critiquing Gary Taubes, Part 4: What Causes Obesity?
Previously: Mainstream Nutrition Science on Obesity, Atkins Redux, Did the US Government Give Us Absurd Advice About Sugar?
In this post, I'm going to deal with an issue that's central to Gary Taubes' critique of mainstream nutrition science: what causes obesity?
... |
3307edf0-3e04-4492-aff4-5eaaf061b3f6 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Chicago Rationality Reading Group
Discussion article for the meetup : Chicago Rationality Reading Group
WHEN: 21 May 2017 01:00:00PM (-0500)
WHERE: Harper Memorial Library Room 148, 1116 E 59th St, Chicago, IL 60637
The Chicago Rationality group meets every Sunday from 1-3 PM in Room 148 of Harper Memorial... |
86ba5c8f-e4d1-482e-aeaa-b194bdf89154 | trentmkelly/LessWrong-43k | LessWrong | It Can't Be Mesa-Optimizers All The Way Down (Or Else It Can't Be Long-Term Supercoherence?)
Epistemic status: After a couple hours of arguing with myself, this still feels potentially important, but my thoughts are pretty raw here.
Hello LessWrong! I’m an undergraduate student studying at the University of Wisconsin... |
d4192b22-3dfe-42cb-bd56-eca87dd9e6f7 | trentmkelly/LessWrong-43k | LessWrong | Ontological Crises in Artificial Agents' Value Systems by Peter de Blanc
I saw this go by on arXiv, and thought it deserved a discussion here.
> Decision-theoretic agents predict and evaluate the results of their actions using a model, or ontology, of their environment. An agent's goal, or utility function, may also ... |
0879afee-cb09-4479-b243-d931a362aa87 | trentmkelly/LessWrong-43k | LessWrong | Michael Vassar's Edge contribution: summary
Michael Vassar has written a provocative response to this year's Edge question: "What *should* we be worried about?". But, I'm confused about his post. My attempt to summarize his point of view follows:
1. People have physiological needs (food, shelter, safety etc.) and soc... |
92b2476e-cd0f-4610-9a2f-25322cb9ec50 | trentmkelly/LessWrong-43k | LessWrong | Mental acceptance and reflection
Naive investigation of the world often works, especially when we have direct access, fast feedback loops, and can try many times. Sports, games, and even parts of science often have properties. When we investigate, say, microorganisms, we can directly experiment on what we are interest... |
0304cf7d-7198-4745-9f8f-43277dc710ba | trentmkelly/LessWrong-43k | LessWrong | Progress links and tweets, 2022-09-14
Links
* Low stone walls called “kites” are believed to have been used for hunting and may date to 8000 BC (h/t @SamoBurja)
* Health insurers just published a trillion prices (h/t Marginal Revolution). There’s a startup to be built in making this data accessible and useful
Twee... |
fc3af6f1-df30-46a4-8305-9f649ffb28dc | trentmkelly/LessWrong-43k | LessWrong | Meetup : Monthly Bay Area meetup
Discussion article for the meetup : Monthly Bay Area meetup
WHEN: 24 March 2012 07:00:00PM (-0700)
WHERE: 2071 University Avenue, Berkeley, CA
Hey, it's time for another monthly meetup! There will be a Bay Area Less Wrong meetup in Berkeley this Saturday, at 7pm in Taiwan Restaurant... |
5d2ee286-a0fb-435c-9ce0-fd6bf1b7e6d1 | trentmkelly/LessWrong-43k | LessWrong | [LINK] How to write a dominant assurance contract on the Ethereum blockchain
None |
5c3fc89a-9564-4c7f-b502-cbcdd30899ce | trentmkelly/LessWrong-43k | LessWrong | Is there a worked example of Georgian taxes?
I'm trying to think through the effects of a Georgian tax system, where the levy is based on "the value of the land". But in my conception, that's either very small, or it actually includes improvement value (and improvements on neighboring land). The entire value of land... |
9e65d1f9-7fb1-43fc-b36a-955395a2fd4a | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | AI risk/reward: A simple model
*All models are wrong, but some are useful.*
George Box
Likewise, most long term forecasts are wrong, the process is useful.
Summary
=======
* I have built a large model around AI risk/reward
* I'm looking for correction
* Numbers that I’m curious for errors in
+ AI to run all funct... |
25d35ba5-8ab2-4f29-bf92-8c178e81f881 | trentmkelly/LessWrong-43k | LessWrong | Do we have the right kind of math for roles, goals and meaning?
Preamble on LessWrong relevance: Agentic goals and values, which are of some interest here at LW, are a special case of a broader class of objects that have interested linguists, biologists, and social scientists for a while. This class of objects should... |
a6176cbb-2346-4612-a837-033a23263f16 | trentmkelly/LessWrong-43k | LessWrong | So You Want to Save the World
This post is very out-of-date. See MIRI's research page for the current research agenda.
So you want to save the world. As it turns out, the world cannot be saved by caped crusaders with great strength and the power of flight. No, the world must be saved by mathematicians, computer sci... |
d5f13f08-d2b0-4f63-b2b8-4ad0ab296909 | trentmkelly/LessWrong-43k | LessWrong | Sequence Posts Exercises
Based on this thread from this month's (April 2011) Open Thread, we have started work on exercises to go along with the posts and help readers internalize the material better. The idea is to have exercises covering all of the sequences eventually (though there doesn't have to be a 1:1 correspo... |
07ced1ae-8833-4d52-b287-fc120a2b574f | trentmkelly/LessWrong-43k | LessWrong | Why should we expect AIs to coordinate well?
This is a common assumption for AI risk scenarios, but it doesn’t seem very justified to me.
https://www.lesswrong.com/posts/gYaKZeBbSL4y2RLP3/strategic-implications-of-ais-ability-to-coordinate-at-low says that AIs could merge their utility functions. But it seems increas... |
fd14c847-881a-45b1-9be5-f971f48744d0 | trentmkelly/LessWrong-43k | LessWrong | How do you learn foreign language vocabulary, beyond Anki?
I am using Anki to learn German vocabulary. It is great for keeping words memorized once I have them at least 30% down, but for entirely new vocabulary I'm really struggling. Anki is especially lacking when there's more than one completely-unknown word in my d... |
628c13db-dc4a-4792-a8bd-4bebb9e1e769 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Can I take ducks home from the park?
Language models, whatever. Maybe they can write code or summarize text or regurgitate copyrighted stuff. But… can you take ducks home from the park? If you ask models how to do that, they often refuse to tell you. So I asked six different models in 16 different ways.
Baseline quer... |
83be58a7-45d2-459c-a6a6-ab6930b43778 | trentmkelly/LessWrong-43k | LessWrong | [LINK] Fixed-action patterns: Stop FAPing!
A worse pun than "JAQing off" for a title, but a nice reminder of a small way not to be stupid.
If you notice yourself making the same arguments over and over, or being accused of saying things irrelevant to the argument, try to stop yourself. |
2305cd53-f181-4230-94f0-df9c2a652a90 | trentmkelly/LessWrong-43k | LessWrong | Paternity tests endanger the unborn
Should paternity testing be compulsory at birth? In discussions of this elsewhere I haven’t seen one set of interests come up: those of children who would not be born if their mothers were faithful. At the start of mandatory paternity testing there would be a round of marriages brea... |
bb4f1d90-5e13-4272-8fee-e883e69f6605 | trentmkelly/LessWrong-43k | LessWrong | Agents that don't become maximisers
A putative new idea for AI control; index here.
According to the basic AI drives thesis, (almost) any agent capable of self-modification will self-modify into an expected utility maximiser.
The typical examples are the inconsistent utility maximisers, the satisficers, unexploitabl... |
d315a59d-a735-4de5-8c1c-10f63f9d86bb | StampyAI/alignment-research-dataset/youtube | Youtube Transcripts | Theory of Mind Breakthrough: AI Consciousness & Disagreements at OpenAI [GPT 4 Tested]
evidence released in the last 48 hours
combined with this study from four weeks
ago will revolutionize how AI a model
such as gpt4 interact with humans from
now on the theory of Mind breakthrough
will also have significant impli... |
e6604135-ac01-4ddc-8aed-42e9b7683795 | trentmkelly/LessWrong-43k | LessWrong | Is cryonics necessary?: Writing yourself into the future
Cryonics appears to be the best hope for continuing a person's existence beyond physical death until other technologies provide better solutions. But despite its best-in-class status, cryonics has several serious downsides.
First and foremost, cryonics is expe... |
b587bf32-0ddd-42d4-95dc-8c0a9b73248e | trentmkelly/LessWrong-43k | LessWrong | ChatGPT's "fuzzy alignment" isn't evidence of AGI alignment: the banana test
TL;DR; ChatGPT achieves seeming alignment by entirely sidestepping the hard questions of alignment and instead refusing to engage with prompts that even border on unaligned territory, instead repeating warm-and-fuzzy platitudes about helpful ... |
f0f17ad6-7653-4a2b-833f-67fcc37afdcf | trentmkelly/LessWrong-43k | LessWrong | Reinforcement, Preference and Utility
Followup to: Reinforcement Learning: A Non-Standard Introduction
A reinforcement-learning agent is interacting with its environment through the perception of observations and the performance of actions.
We describe the influence of the world on the agent in two steps. The first ... |
ed1ec480-ccdc-48b0-9fb4-9848b05c4ca2 | trentmkelly/LessWrong-43k | LessWrong | [LINK] Normalcy bias
A useful bias to quote in discussions that spring up around the subjects we deal with on Less Wrong: Normalcy Bias. It's rather specific, but useful:
> The normalcy bias, or normality bias, refers to a mental state people enter when facing a disaster. It causes people to underestimate both the po... |
23fd03b4-39b8-49a9-9070-6e156e1b2b4c | trentmkelly/LessWrong-43k | LessWrong | What's keeping concerned capabilities gain researchers from leaving the field?
My guess is that there are at least a few capabilities gain researchers who're concerned about the impact their work will have. My guess is that at least a few of these would like to leave, but haven't.
My question is: where are these peo... |
f0fd6109-7b39-416d-9bb7-e1c541b5e846 | trentmkelly/LessWrong-43k | LessWrong | From Capuchins to AI's, Setting an Agenda for the Study of Cultural Cooperation (Part1)
This is a multi-purpose essay-on-the-making, it is being written aiming at the following goals 1) Mandatory essay writing at the end of a semester studying "Cognitive Ethology: Culture in Human and Non-Human Animals" 2) Drafting so... |
ca8f3e02-599b-4d14-babf-f621271ac0fe | trentmkelly/LessWrong-43k | LessWrong | Question: Being uncertain without worrying?
I currently face a pretty major life decision. After some careful analysis, I've concluded that my final decision depends on the answers from some queries that I have made, but whose answers I won't receive for days or perhaps weeks.
In the meantime, I've had great difficul... |
560c5ebe-6f41-4eed-9c5a-f1bf26d66fdd | trentmkelly/LessWrong-43k | LessWrong | Undesirable Conclusions and Origin Adjustment
Utilitarianism is a common ethical viewpoint, especially here, but it is not free of problems. Two of these problems (here collectively called the Undesirable Conclusions) will be discussed here and one of the two given a better name. Origin adjustment will then be used to... |
d52c7aa0-aad7-461f-ac85-45aadb8048bd | trentmkelly/LessWrong-43k | LessWrong | Disorganized collection growth
Cross posted from Overcoming Bias. Comments there.
***
When I was a teenager, I lived in a nice house with my mother, stepfather and three younger brothers. The contents of the house were what you would expect if you took a normal house, multiplied the number of things in it by ten, th... |
170840be-b88a-4443-8a22-e7fdde70a841 | trentmkelly/LessWrong-43k | LessWrong | Collin Burns on Alignment Research And Discovering Latent Knowledge Without Supervision
Collin Burns is a second-year ML PhD at Berkeley, working with Jacob Steinhardt and Dan Klein, whose focus is on making language models honest, interpretable, and aligned.
In our interview we discuss his approach to doing AI Align... |
ceb8d537-dd47-408a-baed-0b25c1876e8a | trentmkelly/LessWrong-43k | LessWrong | LW anchoring experiment: maybe
> I do an informal experiment testing whether LessWrong karma scores are susceptible to a form of anchoring based on the first comment posted; a medium-large effect size is found although the data does not fit the assumed normal distribution & the more sophisticated analysis is equivocal... |
95fd379d-a946-440b-818b-7e8fceb2552f | StampyAI/alignment-research-dataset/arxiv | Arxiv | Enhancing the Robustness, Efficiency, and Diversity of Differentiable Architecture Search
1 Introduction
---------------
Neural architecture search (NAS) has recently attracted enormous attention since it has brought up the prospect of automating the customization of neural architecture for specific tasks.
Compared... |
789004db-afd2-4dba-8602-f101669de80c | trentmkelly/LessWrong-43k | LessWrong | Descriptive vs. specifiable values
[Metadata: crossposted from https://tsvibt.blogspot.com/2022/11/descriptive-vs-specifiable-values.html. First completed November 19, 2022.]
What are an agent's values? An answer to this question might be a good description of the agent's external behavior and internal workings, with... |
9b7c2099-670f-428f-aa9f-3ccd2deaec06 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Catalyst books
One of the most common counterarguments against a lot of alignment research that I hear sounds something like this: “Making current AI do X with Y won’t help, because AGI will RSI to ASI and break everything”. And it.. sounds convincing, but also like an ultimate counterargument? I mean, I agree that al... |
88eb8abc-7ad1-41cd-9ff8-66a0fecc99b5 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Gratification: a useful concept, maybe new
**I did it... my way!**
*Frank Sinatra, Paul Anka, and Elvis Presely, moral philosophers.*
I like winning games. I particularly like winning when I can come up with a new and ingenious way of using the rules to get a bold new strategy. I don't particularly want to w... |
791dca20-7e81-41fe-8230-128ca786b17d | trentmkelly/LessWrong-43k | LessWrong | Community Skill Building for Solstice
Part of the reason the Bay solstice-program is setup up roughly the same way each year is because doing anything different requires a bunch of skills our community doesn't have much access to.
Building new skills (among enough community folk to reliably have it available for sol... |
3f2e3461-ccd2-42aa-8daa-e9f2f81a0ef6 | trentmkelly/LessWrong-43k | LessWrong | The Significance of "Alignment vs Progress: The AI Rap Showdown" in the AI Safety Discourse
As members of the LessWrong community, we've long been concerned with the future of artificial intelligence and its implications for society. In an attempt to spread awareness and spark discussions around the crucial topic of A... |
e2a8ee26-c232-4122-88d5-ece7b0930a9e | trentmkelly/LessWrong-43k | LessWrong | Memo on some neglected topics
I originally wrote this for the Meta Coordination Forum. The organizers were interested in a memo on topics other than alignment that might be increasingly important as AI capabilities rapidly grow — in order to inform the degree to which community-building resources should go towards AI ... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.