id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
fe99c403-155c-44b0-addc-d388da2656a3 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Owain Evans on LLMs, Truthful AI, AI Composition, and More
*(Podcast episode in link)*
I had a discussion with [Owain Evans](https://www.linkedin.com/in/owain-evans-78b210133/) about many topics around language models. We touch on his work with [Truthful AI](https://arxiv.org/abs/2110.06674) and [Ought](https://o... |
0142e0a4-b89b-42c3-9833-ef2b33e91820 | trentmkelly/LessWrong-43k | LessWrong | Counterfactual Contracts
(Cross-posted from my blog)
1.
Imagine coming home from work (back when that was a thing). As you walk in the door, your roommate says “Hey! I bought you a burrito for dinner, mind paying me back?” Some possible responses:
“Thanks! That’s exactly what I wanted for dinner, I will gladly pay ... |
999a7af0-1164-4a22-ba6e-4ec6712bcf2a | StampyAI/alignment-research-dataset/youtube | Youtube Transcripts | Deep Learning 7. Attention and Memory in Deep Learning
okay everyone my name is Alex graves I'm
a research scientist at beep mind which
should no longer say google deepmind but
only deep mine but I haven't updated my
slides obviously and what I'm going to
be talking about in this lecture is
attention and memory and ho... |
48b18af8-4f9a-4d2a-87ef-13e018be44ca | trentmkelly/LessWrong-43k | LessWrong | The Summer Solstice Paradox
In which I call something a paradox that, like most “paradoxes”, isn’t really. But is nonetheless a confusing problem.
After every Winter Solstice I’ve run, I get people asking me “so what’s for Summer Solstice?” It seems like an obvious question – clearly the Winter Solstice is just a sma... |
5d71dc17-917a-4f7c-9d98-ecb05b901c84 | trentmkelly/LessWrong-43k | LessWrong | Some things I've learned in college
In my corner of the internet, there is a healthy contrarian skepticism about education, and about college in particular. Here are some of the points or critiques of the learning that (perhaps) goes on in college, many of which I largely agree with:
Signalling
(i.e. any learning th... |
e645d7a5-2940-421c-94fe-3549b5282479 | trentmkelly/LessWrong-43k | LessWrong | Any LessWrongers in Calgary?
I'm wondering if there'd be any sense in organizing a meet-up. If you're local leave a comment.
|
bcc47b73-f509-4a11-8036-bae7c9239d6a | trentmkelly/LessWrong-43k | LessWrong | Search-in-Territory vs Search-in-Map
Let’s split out two different types of optimization. The first type includes things like Google Maps’ pathfinder, a human making a plan, or Amazon’s logistical algorithms. The second type includes things like a thermostat, a heat-seeking missile, or water flowing downhill.
The dis... |
ef3bd554-bc0f-4997-a596-b77cf96a9c83 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Alignment as Constraints
In order to find the most promising alignment research directions to pour resources into, we can go about it 3 ways
1. Constraints all alignment proposals should have
2. Constraints for current research directions
3. Constraints for new research directions
Constraints all alignment proposals... |
72b2d227-f4ba-4e0b-8d3d-cb01342d5f0e | trentmkelly/LessWrong-43k | LessWrong | What it means to be lucky
This is a linkpost for https://amirbolous.com/posts/luck
* Introduction
* Anecdotes
* Dimension One: Random Good Fortune
* Dimension Two: Luck as a Function
* Dimension Three: Luck as a Serendipity Vehicle
* Closing Thoughts
"When it comes to luck you make your own" - Bruce Springteen... |
eb38e331-bd9b-4e0d-865b-0f1be33eaea9 | trentmkelly/LessWrong-43k | LessWrong | AI Safety Camp, Virtual Edition 2023
AI Safety Camp connects you with an experienced research lead to collaborate on a research project – helping you try your fit for a potential career in AI Safety research.
The applications for AI Safety Camp’s Virtual Edition in 2023 are now open!
AI Safety Camp Virtual 8 will... |
c412c3d4-ac9a-4b38-9c71-1a37f82e260e | trentmkelly/LessWrong-43k | LessWrong | Thoughts about Hugging Face?
I'm curious what the LW community as a whole thinks about the work that Hugging Face is doing. Their main MO seems to be taking whatever new breakthrough in AI there is, and making it open-source and accessible to the public. (as a first order approximation)
I see a few aspects here:
... |
25d5d7f1-88df-4419-8af8-0c8a66895f71 | trentmkelly/LessWrong-43k | LessWrong | Manifund 2025 Regrants
Each year, Manifund partners with regrantors: experts in the field of AI safety, each given an independent budget of $100k+. Regrantors can initiate fast, small grants, seeding early-stage projects with $5k-$50k.
For 2025, we’ve raised $2.25m so far, and are excited to announce our first 10 reg... |
364c85a5-590a-4b71-aaa1-fd321834c6e2 | trentmkelly/LessWrong-43k | LessWrong | A Meritocracy of Taste
Epistemic status: the idea it’s not fully fleshed out (there are a bunch of problems that I’m skipping over, the post is meant more as: “has anyone been seriously thinking about this?” and throwing out a starting point, than: “here is a proposal that would actually work”) and I wouldn’t be surpr... |
19eddfc9-0424-4254-bb96-b5819b68031f | trentmkelly/LessWrong-43k | LessWrong | Reasoning upon its own dark fiction
This post is a loving tribute to Unsong, an amazing novel of kabbalistic magic, theodicy, and Biblical whale puns by Scott Alexander. The apocalyptic conclusion of Unsong is set for the evening of May 14th in both our universe and its own. This is not a coincidence – our own world m... |
00dc4f63-d23a-486e-9251-67a8fadd56e3 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Hot Air Doesn't Disagree
Today's post, Hot Air Doesn't Disagree was originally published on 16 August 2008. A summary (taken from the LW wiki):
> "Disagreement" between rabbits and foxes is sheer anthropomorphism. Rocks and hot air don't disagree, even though one decreases in elevation and one increase... |
10e89931-6b70-4cc6-85c5-04bd3584c488 | StampyAI/alignment-research-dataset/blogs | Blogs | Summary of history (empowerment and well-being lens)
*Click lower right to download or find on Apple Podcasts, Spotify, Stitcher, etc.*
> Now that I've finished the "most important century" series, I'll still be putting out one longer piece per week, but they'll be on toned-down/less ambitious topics, as you're ab... |
c177f4df-8a6f-43f1-932b-7290980840e0 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Berkeley: To-do lists and other systems
Discussion article for the meetup : Berkeley: To-do lists and other systems
WHEN: 22 May 2013 07:30:00PM (-0700)
WHERE: Berkeley, CA
Dear all, the topic of tomorrow's meetup is to-do lists, Getting Things Done, and other systems for managing time and tasks. I used to... |
fc3114bc-5c0b-4ddd-8ea8-62a7a68f9c2d | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Will AI See Sudden Progress?
Will advanced AI let some small group of people or AI systems take over the world?
AI X-risk folks and others have accrued lots of arguments about this over the years, but I think this debate has been disappointing in terms of anyone changing anyone else’s mind, or much being resolved. I ... |
b1b18860-626e-46ad-960c-4116cd8a28c0 | trentmkelly/LessWrong-43k | LessWrong | Lessons On How To Get Things Right On The First Try
This post is based on several true stories, from a workshop which John has run a few times over the past year.
John: Welcome to the Ball -> Cup workshop! Your task for today is simple: I’m going to roll this metal ball:
… down this hotwheels ramp:
… and off th... |
0c70b583-478a-453a-bf71-52d488fe5929 | StampyAI/alignment-research-dataset/youtube | Youtube Transcripts | DeepMind x UCL | Deep Learning Lectures | 12/12 | Responsible Innovation
hi my name is Tony Chen and welcome to
today's lecture on responsible
innovation and artificial intelligence
so today's talk will be split into two
parts for the first part I will be going
over some of the research which has been
done taking sur... |
ad86f1df-32d0-454e-9d6b-bf9be11abbd3 | trentmkelly/LessWrong-43k | LessWrong | How to Save the World
Most of us want to make the world a better place. But what should we do if we want to generate the most positive impact possible? It’s definitely not an easy problem. Lots of smart, talented people with the best of intentions have tried to end war, eliminate poverty, cure disease, stop hunger, pr... |
a7486d70-a620-4d45-9aca-77540f823e73 | trentmkelly/LessWrong-43k | LessWrong | "What Mistakes Are You Making Right Now?"
(That's^ all.)
(Crazy, though, that you successfully noticed your mistake from just a prompt...) |
85c4664f-91ea-4235-9033-430a3bf2d47b | trentmkelly/LessWrong-43k | LessWrong | Learning to Learn and Navigating Moods
Moods are important to learning. Clearly, if you’re upset, it will be harder to learn JavaScript, how to cook a fine pesto, or surf. If you’re curious and filled with wonder, it will be much easier. Because your mood can be either promote or hamper your learning success, learning... |
d99ad18c-a342-43ba-bf45-ad4dc2c2698e | trentmkelly/LessWrong-43k | LessWrong | Evolutionary psychology as "the truth-killer"
So, a little background- I've just come out as an atheist to my dad, a Christian pastor, who's convinced he can "fix" my thinking and is bombarding me with a number of flimsy arguments that I'm having trouble articulating a response to, and need help shutting down. The par... |
1e9939a1-2ba9-40ad-bc75-cd4d8b0b4cd4 | trentmkelly/LessWrong-43k | LessWrong | About addition and truth
This is intended to explore a a thought I had, rather than making any particular argument about truth.
The canonical example of a thing which is true without any obvious physical referent is the statement 2+2=4. It is true about fingers, sheep, particles, and galaxies; but intuitively it does... |
f7a7cd53-b3c0-46e8-9f24-f54b62238864 | StampyAI/alignment-research-dataset/special_docs | Other | The Role and Limits of Principles in AI Ethics: Towards a Focus on Tensions
The Role and Limits of Principles in AI Ethics:
Towards a Focus on Tensions
Jess Whittlestone∗
Leverhulme Centre for the Future of Intelligence
University of Cambridge
jlw84@cam.ac.ukRune Nyrup
Leverhulme Centre for the Future of Intelligence
... |
c30e0fce-17f3-4e2d-b889-ca62ec9bb654 | trentmkelly/LessWrong-43k | LessWrong | How to write Pseudocode and why you should
TLDR: Writing pseudocode is extremely useful when designing algorithms. Most people do it wrong. Mainly because they don't intentionally try to keep the code as abstract as possible. Possibly this happens because in normal programming a common workflow is to focus on getting ... |
2f324e1f-bbd7-474a-8474-236eff477586 | trentmkelly/LessWrong-43k | LessWrong | Test
Test. |
41a9d174-82e5-4216-9eaa-3a990a26ef18 | trentmkelly/LessWrong-43k | LessWrong | Löb's Lemma: an easier approach to Löb's Theorem
Related to: Löb's Theorem for implicit reasoning in natural language: Löbian party invitations
tl;dr: Löb's Theorem is much easier to grok if you separate the parts of the proof that use the assumption □p→p from the parts that don't. The parts that don't use □p→p can... |
66d87e47-0617-4746-b45c-5993d423f900 | trentmkelly/LessWrong-43k | LessWrong | A short introduction to machine learning
Despite the current popularity of machine learning, I haven’t found any short introductions to it which quite match the way I prefer to introduce people to the field. So here’s my own. Compared with other introductions, I’ve focused less on explaining each concept in detail, an... |
46d4c77a-a40d-4bb9-9232-c02c0f3aeed5 | trentmkelly/LessWrong-43k | LessWrong | Group Rationality Diary, October 1-15, plus frequency poll
This is the public group instrumental rationality diary for October 1-15.
> It's a place to record and chat about it if you have done, or are actively doing, things like:
>
> * Established a useful new habit
> * Obtained new evidence that made you change y... |
ec11a84f-5df1-4582-96a5-7e525f70ecf8 | trentmkelly/LessWrong-43k | LessWrong | How To Actually Change Your Mind eBook (In Order)
Firstly, if this has already been created, then I apologize, but I couldn't find it and I looked.
I wanted to read the how to change your mind sequence on my eBook reader, but when I looked around the site I couldn't find any eBooks other than large ones with the blog... |
6c8e0b1a-50f1-4584-b41e-63a07c0dbd96 | trentmkelly/LessWrong-43k | LessWrong | Penguicon & Blook
One million cumulative daily visits! Woot n' stuff. Also we're in the top 5,000 of all blogs on Technorati, and one of the top 10 econblogs by Technorati rank.
Seems like a good time to mention that I'll be appearing at Penguicon, a combination open-source/science-fiction convention in Troy, MI, A... |
d3ea1278-b230-4659-ba87-b2be0f4cec37 | trentmkelly/LessWrong-43k | LessWrong | [Interview w/ Rob Miles] The case for taking AI Safety seriously
I recently concluded a fascinating interview with science communicator Rob Miles on the Futurati Podcast.
We cover:
* His background in computer science, and how he became interested in the broader AI Safety debate.
* The resources he recommends for ... |
584742c8-662e-4a1c-8888-3f3d76c0ddbb | trentmkelly/LessWrong-43k | LessWrong | Fertility Rate Roundup #1
Previously: On Car Seats as Contraception
[Editor’s Note: This post assumes the perspective that more people having more children is good, actually. I will not be engaging with any of the arguments against this, of any quality, whether they be ‘AI or climate change is going to kill everyone’... |
b556200c-0344-4eaf-9bfc-934d4cba0345 | trentmkelly/LessWrong-43k | LessWrong | Prompts for Big-Picture Planning
During my metastrategy workshop, Day Two was focused on taking a step back and asking "okay, wait, what am I actually doing and why?". Choosing what area to focus, and what your mid-level strategy is for achieving it, determine at least as much (and I think often much more) of the valu... |
7361c37a-c23f-41be-bb7f-59207f5afd39 | trentmkelly/LessWrong-43k | LessWrong | Decision-Relevance of worlds and ADT implementations
Crossposted on the EA Forum.
Lots of effort has been put into estimating the density of Space-Faring Civilisations (SFCs) in the universe. Further work built on these results to produce decision-relevant information assuming various anthropic updates (SSA, SIA, AD... |
7d8dfeed-3a55-45af-9e3e-044ccd1f8e71 | StampyAI/alignment-research-dataset/special_docs | Other | Estimation with Incomplete Data: The Linear Case.
Estimation with Incomplete Data: The Linear Case
Karthika Mohan1, Felix Thoemmes2andJudea Pearl3
1University of California, Berkeley
2Cornell University
3University of California, Los Angeles
karthika@eecs.berkeley.edu, felix.thoemmes@cornell.edu, judea@cs.ucla.edu
Abs... |
977493a8-6dd8-4e4c-ab3e-0e0ce6e8f833 | trentmkelly/LessWrong-43k | LessWrong | How can I get over my fear of becoming an emulated consciousness?
Summary
In the last few months I've been suffering from bouts of abject terror related to the thought of experiencing reality as a thinking being only - one with no senses of any kind, as well as no agency - presumably as a result of unintended conseque... |
d741c39b-8fc4-487d-a824-7e6655a366e7 | trentmkelly/LessWrong-43k | LessWrong | Covid 4/21/22: Variants Working Overtime
Earlier this week, a judge overturned the transportation mask mandate, which I covered here. I will also be covering the situation in China in its own post.
No one asked this time around, but a friend recommends this P100 mask if you’re looking to go big rather than go home.
... |
d40129cb-fcb0-4c75-86be-27440af83d3e | trentmkelly/LessWrong-43k | LessWrong | Fetch The Coffee!
This is a reaction to a specific point in the "Debate on Instrumental Convergence between LeCun, Russell, Bengio, Zador, and More"
The Disagreement
"Stuart Russell: It is trivial to construct a toy MDP in which the agent's only reward comes from fetching the coffee. If, in that MDP, there is anothe... |
9e5f7e0b-26b7-4580-ad3b-fbf26b42974d | trentmkelly/LessWrong-43k | LessWrong | Thoughts on the SPIES Forecasting Method?
Here, I discuss the SPIES forecasting method, and ask for the community's thoughts on it.
Not too long ago, I came across the SPIES (Subjective Probability Interval Estimates) method for judgmental forecasting. The method was developed by Uriel Haran, and seems first to have... |
092f876c-8d0b-4224-a15a-bf79965450cc | trentmkelly/LessWrong-43k | LessWrong | What is autonomy, and how does it lead to greater risk from AI?
As with many concepts in discussions of AI risk, terminology around what autonomy is, what agency is, and how they might create risks is deeply confused and confusing, and this is leading to people talking past one another. In this case, the seeming binar... |
efb75d09-9886-44e4-9b4e-b2f26f6180ec | StampyAI/alignment-research-dataset/blogs | Blogs | Symbology for Topia
Symbology for Topia
-------------------
i have come up with what i'd like to think is a cool symbol for [my utopia](two-principles-for-topia.html):
the background is black, for the inherent anarchism of my ideology (singularity doesn't count as state, it's more like an ad... |
8e07ecc5-b15b-4126-be58-240d8033a3b2 | trentmkelly/LessWrong-43k | LessWrong | Real-World Coordination Problems are Usually Information Problems
Let’s start with a few examples of very common real-world coordination problems.
* The marketing department at a car dealership posts ads for specific cars, but the salespeople don’t know which cars were advertised, causing confusion when a customer c... |
2700eb6c-aa8a-413d-8311-f4db46d1068c | trentmkelly/LessWrong-43k | LessWrong | LLMs and hallucination, like white on rice?
Cross posted from New Savanna.
It’s a matter of history, I suppose.
We know that LLMs confabulate. They make stuff up. That’s what’s known as hallucination. Here’s the introductory section of the Wikipedia entry for Hallucination (artificial intelligence):
> In artificia... |
0e0d55a3-d518-4991-8196-e2e70e1e4102 | trentmkelly/LessWrong-43k | LessWrong | Ophiology (or, how the Mamba architecture works)
The following post was made as part of Danielle's MATS work on doing circuit-based mech interp on Mamba, mentored by Adrià Garriga-Alonso. It's the first in a sequence of posts about finding an IOI circuit in Mamba/applying ACDC to Mamba.
This introductory post was als... |
075ee4d2-5cd9-4f8b-8927-ac47d655e7ca | trentmkelly/LessWrong-43k | LessWrong | Netflix's "Start-Up" and sincere work dramatization
I'm in the middle of watching the show Start-Up on Netflix. I just finished episode 5 (some spoilers). Some background: Seo Dal-mi and Do-san are our main characters. They and Do-san's machine-learning-developer buddies are all on the same team at a hackathon. At the... |
7a445410-7af6-485b-a4a7-bb55b14a645d | StampyAI/alignment-research-dataset/arxiv | Arxiv | Reward learning from human preferences and demonstrations in Atari
1 Introduction
---------------
Reinforcement learning (RL) has recently been very successful in solving hard problems in domains with well-specified reward functions (Mnih et al., [2015](#bib.bib31), [2016](#bib.bib32); Silver et al., [2016](#bib.bi... |
949af747-ee82-475f-a1b5-f578423a136e | trentmkelly/LessWrong-43k | LessWrong | Links for Nov 2020
I started collecting various links on FB and figured I'd cross-post. This is the second post, there's a link at the top to the first post a month ago. |
53627143-5999-4c5c-8700-4654a023f427 | trentmkelly/LessWrong-43k | LessWrong | Tool for combating undue hesitation
I sometimes feel negative emotions at the thought that the course of action that I am taking isn't even close to the optimal course of action -- that a more effective mind could sort through whatever situation I'm currently having difficulty with and craft a plan that was much more ... |
51ac39a4-ebbe-4399-b4a2-48f28e9657a9 | StampyAI/alignment-research-dataset/youtube | Youtube Transcripts | Demis Hassabis: The interview - DeepMind: The Podcast (S1, Ep8)
[Music]
so here we are the last episode in this
series of the deep mind podcast my name
is Hanna Frey I am a mathematician and
someone who is deeply intrigued by
artificial intelligence much like you I
imagine since you made it this far now
we've been toy... |
87d1707f-ba9d-4ce5-b2fc-d5a5e1d225af | trentmkelly/LessWrong-43k | LessWrong | Promoting rationality to a broad audience - feedback on methods
We at Intentional Insights , the nonprofit devoted to promoting rationality and effective altruism to a broad audience, are finalizing our Theory of Change (a ToC is meant to convey our goals, assumptions, methods, and metrics). Since there's recently be... |
32dec52c-00e2-462b-aee9-f495e98eae2b | trentmkelly/LessWrong-43k | LessWrong | Does Evidence Have To Be Certain?
It seems like in order to go from P(H) to P(H|E) you have to become certain that E. Am I wrong about that?
Say you have the following joint distribution:
P(H&E) = a
P(~H&E) = b
P(H&~E) = c
P(~H&~E) = d
Where a,b,c, and d, are each larger than 0.
So P(H|E) = a/(a+b). It seems li... |
1b07c9dc-2662-4a58-8f5c-0bd124cd84aa | trentmkelly/LessWrong-43k | LessWrong | Meetup : Washington, D.C.: Sword of Good
Discussion article for the meetup : Washington, D.C.: Sword of Good
WHEN: 19 April 2015 03:00:00PM (-0400)
WHERE: Reynolds Center
We will be meeting in the Kogod Courtyard of the Donald W. Reynolds Center for American Art and Portraiture (8th and F Sts or 8th and G Sts NW, g... |
e9bf2f0b-0b62-46e7-967c-d9ddde0600de | StampyAI/alignment-research-dataset/aisafety.info | AI Safety Info | Who should I talk to about my non-research AI alignment coding project idea?
The [Alignment Development Ecosystem](https://alignment.dev/) is a big central navigation point where you can find collaborators and other resources. After you’ve found some people to get feedback from, our advice is to lean toward just makin... |
c177d2b5-76b5-47a7-8fad-08509e71881f | trentmkelly/LessWrong-43k | LessWrong | Why is our sex drive too strong?
It is a cultural universal that people are discouraged from having sex as often and with as many people as they want to. Every culture I've ever heard of imposes many restrictions on sex. I've never heard of a culture that shames people for being too stingy with sex.
If we assume th... |
feb075e0-bf77-4c9a-8d73-7ad8f593fbb3 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | GPT can write Quines now (GPT-4)
GPT-4 is not only able to write code, more reliably than GPT-3.5, it writes code *that writes code*; see the example below (GPT-3.5 was not able to do this). But first, let me say:
1) @OpenAI: Thank for your openness to the world about your capabilities and shortcomings! Specificall... |
5dd504ba-fe2e-4c35-bea2-3eab64651c19 | trentmkelly/LessWrong-43k | LessWrong | What are the limitations on politically motivated relocation?
I know that this idea has been around for a very long time (probably since representative democracy), and that there are current attempts at it. The Free State Project considered Wyoming, Vermont and Alaska (amongst others), the least populous states in th... |
0d9ba2b8-4c9c-4d95-abf3-bbdda6a16a5e | trentmkelly/LessWrong-43k | LessWrong | Drawing Less Wrong: An Introduction
> You have not found a way to connect your hobby to rationalism yet. It itches. You are not whole. It is forbidden to post an article entitled Rationalist Hobby on Less Wrong. You lie awake at three in the morning, trying to create puns.
> -- Alicorn, Rationality Gothic [2]
(Note ... |
21be4f4e-bce4-4570-897f-1946abbe4211 | trentmkelly/LessWrong-43k | LessWrong | After Overmorrow: Scattered Musings on the Immediate Post-AGI World
Cutting to the case, I've been grappling with various thoughts about the immediate future for over a year now— it was almost exactly a year ago that Yudkowsky's aggressive predictions of absolute and unquestionable doom led me to decide to take a vesp... |
a1c1ef74-1f63-425a-b499-f3986934323c | trentmkelly/LessWrong-43k | LessWrong | Proverbs and Cached Judgments: the Rolling Stone
People have long noted that individuals diagnosed as schizophrenic usually manifest disturbances of language, communication, and abstract thought. One way to examine that disturbance is to ask patients to interpret various common proverbs, as psychiatrists have done si... |
83b7c422-d4e5-41c6-8bd9-fab9378d7cff | StampyAI/alignment-research-dataset/arxiv | Arxiv | AI Forensics: Did the Artificial Intelligence System Do It? Why?
1 Introduction
---------------
The number of “smart” devices is growing rapidly which makes them valuable for investigations; they are confiscated with the hope to find evidence on them [[18](#bib.bib8 "IoT Ignorance is Digital Forensics Research Blis... |
fb46b069-3dbb-4c83-aa5b-03fc47e95d43 | trentmkelly/LessWrong-43k | LessWrong | A bridge to Dath Ilan? Improved governance on the critical path to AI alignment.
Summary: This post showcases my 2nd-place-winning entry in the Future of Life Institute's AI worldbuilding contest. It imagines:
1. How we might make big improvements to decisionmaking via mechanisms like futarchy and liquid democracy... |
e171adcc-058a-4262-92e3-2728297fe1ad | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Making decisions under moral uncertainty
*Cross-posted [to the EA Forum](https://forum.effectivealtruism.org/posts/ex834aaANLhamLkvf/making-decisions-under-moral-uncertainty). Updated substantially since initial publication.*
Overview/purpose of this sequence
---------------------------------
While working on an (... |
7ad35ec7-ee72-419f-99ef-9e4c4923320b | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | US public perception of CAIS statement and the risk of extinction
Summary
=======
On June 2nd-June 3rd 2023, Rethink Priorities conducted an online poll of US adults, to assess their views regarding a recent open statement from the Center for AI Safety (CAIS). The statement, which has been signed by a number of prom... |
ee48e4b5-ebe4-493e-9646-27372810b2ed | trentmkelly/LessWrong-43k | LessWrong | Interpretable by Design - Constraint Sets with Disjoint Limit Points
cart;horse: How can we constrain our models to be interpretable? Convex, linear sets make for more interpretable parameter spaces, and the simplex and the Birkhoff Polytope are great examples of this that have other desirable properties.
An interpre... |
f2e9166a-a701-4e30-b00f-f6c8522259e9 | StampyAI/alignment-research-dataset/arbital | Arbital | Ideals are the same thing as kernels of ring homomorphisms
In [ring theory](https://arbital.com/p/3gq), the notion of "[ideal](https://arbital.com/p/ideal_ring_theory)" corresponds precisely with the notion of "[kernel](https://arbital.com/p/5r6) of [https://arbital.com/p/-ring_homomorphism](https://arbital.com/p/-rin... |
01a47e44-bd68-4def-9cf0-dd5ed2b324fe | trentmkelly/LessWrong-43k | LessWrong | Practical use of the Beta distribution for data analysis
Many data scientists and data analysts working today use the Gaussian (Normal) distribution when computing probabilities from yes/no count data. The appropriate distribution for binary count data is the Beta distribution, and the Gaussian is only an approximatio... |
e81a4f1f-a312-4ec1-8363-e1d3070369d4 | trentmkelly/LessWrong-43k | LessWrong | Meetup : LessWrong Tel Aviv: Rational Career Development
Discussion article for the meetup : LessWrong Tel Aviv: Rational Career Development
WHEN: 18 September 2014 08:33:43PM (+0300)
WHERE: Yigal Alon Street 98, Tel Aviv-Yafo, Israel
The organization of this meetup was presided by Aur Saraf. The text below is by h... |
c5298b96-6117-4f37-8391-d7b20cf6449b | trentmkelly/LessWrong-43k | LessWrong | Just another day in utopia
(Reposted from discussion at commentator suggestion)
Thinking of Eliezer's fun theory and the challenge of creating actual utopias where people would like to live, I tried to write a light utopia for my friends around Christmas, and thought it might be worth sharing. It's a techno-utopia, b... |
ced68832-8f90-4691-9b8f-c5d93eb09c14 | trentmkelly/LessWrong-43k | LessWrong | Seeking Samsara
Thousands of years ago, the Greeks determined that παν μέτρον άριστον, or alternatively, that virtue is a golden mean. Despite the fact that Western culture has at least superficially internalized the idea — the average Joe on the street likely reflexively concurs with squiggly formula — there are stil... |
3919d891-ba56-4cf7-b2a8-671e10be0d1d | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post2775
Suppose you are a friendly AI A and have a mysterious black box B . B outputs a sequence of bits. You want to predict the next bits that B will output. Fortunately, you have a magic Turing machine oracle O . You can give O any computable function f ( Turing machine, does it Halt? , What does it output? ,... |
32381210-7d8c-42a4-ba39-0e0d926050b1 | trentmkelly/LessWrong-43k | LessWrong | ARENA 4.0 Impact Report
If you're interested in helping to run the ARENA program, note that we're currently hiring for an Operations Lead! For more details, and to apply, see here.
Summary
The purpose of this report is to evaluate ARENA 4.0’s impact according to our four success criteria:
1. Source high-quality pa... |
f585a47d-6cf3-4d43-a98d-ca0baa4390cb | trentmkelly/LessWrong-43k | LessWrong | Yoda Timers 2
This is part 12 of 30 of Hammertime. Click here for the intro.
> Anyone who can muster their willpower for thirty seconds, can make a desperate effort to lift more weight than they usually could. But what if the weight that needs lifting is a truck? Then desperate efforts won’t suffice; you’ll have to... |
1abe26f2-d40f-460f-830c-245c4663bbcc | trentmkelly/LessWrong-43k | LessWrong | Can HCH epistemically dominate Ramanujan?
Srinivasa Ramanujan is an Indian mathematician who is famously known for solving math problems with sudden and inexplicable flashes of insight. From his Wikipedia page:
> Imagine that you are on a street with houses marked 1 through n. There is a house in between (x) such tha... |
bd247154-beb1-45ee-b1ed-b121c92c5c6e | trentmkelly/LessWrong-43k | LessWrong | On Reflection
Introduction
The world is full of wasted motion. Many things are far below where they should be, many actions fail to achieve their goals, and this includes many of your actions. And this is not because you’re failing, but because achieving your goals is hard.
This is obviously not a problem that can ev... |
e807d11f-a12f-4fbf-838b-262ad99f0e62 | trentmkelly/LessWrong-43k | LessWrong | And the Darkness Answered
Epistemic Status: Endorsed
Part of the Series: Open Portals
Previous Post: My Journey to the Dark Side
Author: Octavia
“This horror will grow mild, this darkness light”
On August 22nd 2020, Shiloh finally reached enough of a breaking point to be willing to let go of the artifacts which were... |
a06edcf1-c8b5-41b0-a702-9af8b18561f9 | trentmkelly/LessWrong-43k | LessWrong | Open Thread
This is an experiment to see whether people would like an open thread for small topics and conversation. |
ff6277e2-2938-4b18-84be-d9e4246e0ab6 | trentmkelly/LessWrong-43k | LessWrong | A bunch of videos in comments
Hi. Welcome to my personal blog on lesswrong. I like youtube videos. I've posted lists of them before. Since then, more youtube videos have been made that I want to share, and tools for summarizing videos have come out! So I've wanted to share them here. To allow this to be evergreen wher... |
ce7076c4-06db-498c-be93-196bb8c1fe64 | trentmkelly/LessWrong-43k | LessWrong | Best empirical evidence on better than SP500 investment returns?
I am not smart enough to be a real rationalist, and I am really lazy. But I am smart enough to understand and implement rationalist advice and I suspect y'all are a better source of advice than normal investors. What is the highest level of returns I can... |
04daf6a8-a414-45d2-b260-3eb87c2ab18e | trentmkelly/LessWrong-43k | LessWrong | [ASoT] Natural abstractions and AlphaZero
I just read Acquisition of Chess Knowledge in AlphaZero and it's both really cool and has interesting implications for the Natural Abstractions Hypothesis. AZ was trained with no human data and yet it settles on relatively interpretable abstractions for chess.[1]
They trained... |
06cb4464-d268-4010-bf5a-08e01f1b1218 | StampyAI/alignment-research-dataset/arxiv | Arxiv | A Bounded Measure for Estimating the Benefit of Visualization
1 Related Work
---------------
Claude Shannon’s landmark article in 1948 [[40](#bib.bib40)] signifies the birth of information theory.
It has been underpinning the fields of data communication, compression, and encryption since.
As a mathematical framewo... |
32270159-f09e-4183-991b-ad6ab94992bf | trentmkelly/LessWrong-43k | LessWrong | Replaceability as a virtue
I propose it is altruistic to be replaceable and therefore, those who strive to be altruistic should strive to be replaceable.
As far as I can Google, this does not seem to have been proposed before. LW should be a good place to discuss it. A community interested in rational and ethical beh... |
32c353f2-e974-4b74-966c-7d311e2e54d6 | trentmkelly/LessWrong-43k | LessWrong | [LINK] Terrorists target AI researchers
Something a number of LWers should probably be cautious of.
http://www.nature.com/news/2011/110822/full/476373a.html |
d58dc0d0-d416-47ff-b767-eeab3a1f7c34 | trentmkelly/LessWrong-43k | LessWrong | What are the mostly likely ways AGI will emerge?
I'd rather not bias answers with my specific guesses, but what I'm trying to get at are bottom-up economic, engineering, and organizational realities-based answers. For example, if this question were asked in the year 2000 about the most likely way the next massive comm... |
d41aa629-0941-486e-8de9-7950f27c465e | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Focusing your impact on short vs long TAI timelines
**Summary/Key Points**
======================
I compare considerations for prioritizing impact in short vs. long transformative AI timeline scenarios. Though lots of relevant work seems timelines-agnostic, this analysis is primarily intended for work whose impact is... |
fa2fdfe1-2a8b-4041-8bb4-94a46705ad44 | trentmkelly/LessWrong-43k | LessWrong | AI Alignment via Slow Substrates: Early Empirical Results With StarCraft II
A few months ago, I wrote a post about using slower computing substrates as a possibly new way to safely train and align ASI.
If you haven't read that post, basically the idea is that if we consider compute speed as a factor in Total Intellig... |
1167f361-2925-472c-9495-38424031c7bd | trentmkelly/LessWrong-43k | LessWrong | Artificial Moral Advisors: A New Perspective from Moral Psychology
Tired: can humans solve artificial intelligence alignment?
Wired: can artificial intelligence solve human alignment?
“Philosophers have recently put forward the possibility of achieving moral enhancement through artificial intelligence… proposing var... |
9dd9f12a-aa1d-4d0e-89ff-baf08c5be561 | trentmkelly/LessWrong-43k | LessWrong | How accurate are standard Dark Triad personality scales?
Dark Triad, the group of three traits—Machiavellianism, narcissism, and psychopathy—has been studied by well-known researchers Delroy L. Paulhus and Kevin M. Williams, and they published their study in a research paper in 2002. The research suggests that every i... |
0f58bf81-c932-4c55-8180-ea0fe28a3884 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | AXRP Episode 13 - First Principles of AGI Safety with Richard Ngo
[Google Podcasts link](https://podcasts.google.com/feed/aHR0cHM6Ly9heHJwb2RjYXN0LmxpYnN5bi5jb20vcnNz/episode/OTlmYzM1ZjEtMDFkMi00ZTExLWExYjEtNTYwOTg2ZWNhOWNi)
This podcast is called AXRP, pronounced axe-urp and short for the AI X-risk Research Podcast... |
5ebc6e34-51fc-49e6-8fdd-1a76afc14233 | trentmkelly/LessWrong-43k | LessWrong | A thought experiment
Say you are about to flip a quantum coin, and the coin has an equal probability of coming up heads or tails.
If it comes up heads, a machine creates 1,000 simulations of you before flipping the coin (with the same subjective experience so that you cannot tell if you yourself is a simulation) and ... |
4088b68b-4cb4-475e-80d8-8c09c75002e1 | trentmkelly/LessWrong-43k | LessWrong | The role of neodeconstructive rationalism in the works of Less Wrong
Summary: Yudkowsky's fiction emphasizes neodeconstructive rationalism, which serves as a bridge between class and sexual identity. Materialist libertarianism (in the metaphysical sense) implies quantum nonrealism, but examining the works of Vinge, Gi... |
6e45e65c-a717-436c-937a-885b92973452 | trentmkelly/LessWrong-43k | LessWrong | If alignment problem was unsolvable, would that avoid doom?
Suppose there is a useful formulation of the alignment problem that is mathematically unsolvable. Suppose that as a corollary, modifying your own mind while ensuring any non-trivial property of the resulting mind was also impossible.
Would that prevent a new... |
0f330038-9a28-47c6-8d43-8efec7933a64 | trentmkelly/LessWrong-43k | LessWrong | Take 9: No, RLHF/IDA/debate doesn't solve outer alignment.
As a writing exercise, I'm writing an AI Alignment Hot Take Advent Calendar - one new hot take, written every day (ish) for 25 days. Or until I run out of hot takes. And now, time for the week of RLHF takes.
I see people say one of these surprisingly often.
... |
ec9c114c-8d76-47aa-b3fc-f2de83cbb24d | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | The inordinately slow spread of good AGI conversations in ML
Spencer Greenberg wrote [on Twitter](https://twitter.com/SpencrGreenberg/status/1536760598486196225):
> Recently @KerryLVaughan has been critiquing groups trying to build AGI, saying that by being aware of risks but still trying to make it, they’re reckles... |
2b1852ce-c7fb-437b-bccb-e9ff9b0fecef | trentmkelly/LessWrong-43k | LessWrong | Meetup : Next Meetup: Fermi Estimates and Singing
Discussion article for the meetup : Next Meetup: Fermi Estimates and Singing
WHEN: 06 April 2014 03:00:00PM (-0700)
WHERE: National Portrait Gallery, Washington DC
We'll be meeting to do fermi estimates and maybe sing.
(I won't be making it)
Discussion article fo... |
08e9b09c-b58e-4edc-9dfc-9def9748af53 | trentmkelly/LessWrong-43k | LessWrong | Ritual Report: Boston Solstice Celebration
A week after the large-scale Solstice celebration in NYC, we held a smaller one in Boston at Citadel house, with around 20 people attending. The content was essentially a subset (given below) of the 2012 NYC solstice set list - a mix of silly and serious songs and readings, a... |
2fba2a92-c39a-405b-9678-3dc327a8af98 | trentmkelly/LessWrong-43k | LessWrong |
LW Update 2019-1-09 – Question Updates, UserProfile Sorting
First update of 2019!
Streamlined Mobile Above-the-Fold
On mobile, logged out users now see cleaner, slightly sleeker collections of Rationality: A-Z, Codex and HPMOR. Logged in users see only one recommended sequence instead of 3.
Improved Q&A Features
... |
53717d39-3390-4ce2-9139-746e00b8f72a | trentmkelly/LessWrong-43k | LessWrong | Predictions for 2021
Just as I did last year, I have some probabilistic predictions for 2021. In January 2022 I will return to grade them, just as in a week or two I’ll grade my 2020 predictions. This year the predictions fall into four categories: U.S. politics (#1-17 below), COVID (#18-39), Miscellaneous (#40-53), a... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.