id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
327863ca-6aac-4ccb-a255-3806c7bbec66 | trentmkelly/LessWrong-43k | LessWrong | Looking for study
I believe there was a post by Scott Alexander at some point where he referenced a study that demonstrated being depressed made people inaccurate about how happy they would be in 5 years or some time in the future, that they were generally happier than they expected. My googling powers have failed me ... |
b3143d89-ab4d-4a97-a704-c374b00bcca7 | trentmkelly/LessWrong-43k | LessWrong | Discursive Competence in ChatGPT, Part 2: Memory for Texts
I've finished a new working paper. Title above, links, abstract, table of contents, and introduction below.
Academia.edu: https://www.academia.edu/107318793/Discursive_Competence_in_ChatGPT_Part_2_Memory_for_Texts
SSRN: https://ssrn.com/abstract=4585825
Resea... |
5f5b71d7-e67c-4fb7-81e9-7d4c15b93339 | trentmkelly/LessWrong-43k | LessWrong | AI Safety at the Frontier: Paper Highlights, February '25
This is the selection of AI safety papers from my blog "AI Safety at the Frontier". The selection primarily covers ML-oriented research and frontier models. It's primarily concerned with papers (arXiv, conferences etc.).
tl;dr
Paper of the month:
Emergent mi... |
51f4d300-460c-4542-8faa-ea56258bd4a5 | trentmkelly/LessWrong-43k | LessWrong | Selfism and Partiality
Several moral systems spend a great deal of effort in trying to resolve issues of conflict between the self and others. Indeed one of criticisms against consequentialism is that it lacks accommodation for partiality (people’s inherit tendency to give special moral weight to themselves, family an... |
7855ebd2-0719-4816-b263-ee3d1d7e0e28 | trentmkelly/LessWrong-43k | LessWrong | Alignment Targets and The Natural Abstraction Hypothesis
In this post, we explore the Natural Abstraction Hypothesis (NAH), a theory concerning the internal representations used by cognitive systems. We'll delve into how this hypothesis could significantly reduce the amount of data required to align AGI, provided we h... |
a9666127-49c5-43c6-8bbc-f9d77876061c | trentmkelly/LessWrong-43k | LessWrong | Rationality when it's somewhat hard
Tunnel Creek avalanche kills skiers:
The page I've linked to describes a party of sixteen excellent skiers who went on a trip where they easily could have known better. Three of them died. It's common knowledge that large parties increase the risk of avalanche, but the page describ... |
6db27046-463e-4ff6-acc0-f665fe73d8bb | trentmkelly/LessWrong-43k | LessWrong | {Math} A times tables memory.
I have a distinct memory of being 8 years old, or so, and being handed one of those worksheets where they ask you to multiply numbers up through 12x12, and being viscerally disgusted by the implied pedagogy of it. That was over a hundred things you were asking me to memorize. On my own ti... |
f2425df9-7670-47c7-96ef-81d4d6a49191 | trentmkelly/LessWrong-43k | LessWrong | My experience using financial commitments to overcome akrasia
About a year ago I decided to try using one of those apps where you tie your goals to some kind of financial penalty. The specific one I tried is Forfeit, which I liked the look of because it’s relatively simple, you set single tasks which you have to verif... |
0afbbe4d-ff04-46e7-b01b-6bf3729a643d | trentmkelly/LessWrong-43k | LessWrong | The Apprentice Experiment
About two months ago, someone asked me what I would do with more funding. Other than the obvious (i.e. generally improve my own quality-of-life in minor ways), my main answer was: take on an apprentice. I have some models about how best to train people for this sort of work, and an apprentice... |
c0601271-9db3-4c9b-9d54-df9c3ee1cae6 | trentmkelly/LessWrong-43k | LessWrong | Other utopias or searching the reference class for the rationality hub project
Let's find the reference class for the new rational hub project, come up with some parameters to assess them, and quarter-bake a model to constrain our migration expectations.
I arbitrarily selected Free Town Project and Rajneeshpuram. If... |
8a71f4f0-30bc-4c16-b648-8e19883c5764 | trentmkelly/LessWrong-43k | LessWrong | Seeking Input to AI Safety Book for non-technical audience
TLDR: You are invited to make suggestions for ideas/information/framing for an upcoming AI safety book for a non-technical audience.
Context:
I'm still writing an accessible book about AI Safety/Risk for a non-technical audience to serve both the AI safety ... |
6455b961-e39c-40ab-901e-ad0e2e66d0c6 | trentmkelly/LessWrong-43k | LessWrong | Covid 6/24: The Spanish Prisoner
The last scare is underway. Delta is an increasing share of Covid cases around the world, causing cases in many places to rise. Are enough people vaccinated? How bad are things going to get before we turn the corner one final time?
The incremental news was not good. Calculations that... |
b4677ddf-a117-436e-8121-8b7d570c9475 | trentmkelly/LessWrong-43k | LessWrong | A case for peer-reviewed conspiracy theories
Conspiracy theories can be thought of as revisionist history that is still political. Speculation is a normal part of analyzing politics. So, while these theories are commonplace historically speaking, the use of the term "conspiracy theory" for stigmatization and idea repr... |
411825c5-926b-4638-9e53-72e79a4cf650 | trentmkelly/LessWrong-43k | LessWrong | Using LLM Search to Augment (Mathematics) Research
In this post, I would like to offer evidence to support the following beliefs:
1. LLM search tools like Perplexity.ai greatly speed up the search portion of thinking about research questions.
2. These tools are more effective the more domain expertise the user has.
3.... |
4403fd83-5120-4639-8cce-bd6cd5af3836 | trentmkelly/LessWrong-43k | LessWrong | Parametrically retargetable decision-makers tend to seek power
This paper—accepted as a poster to NeurIPS 2022— is the sequel to Optimal Policies Tend to Seek Power. The new theoretical results are extremely broad, discarding the requirements of full observability, optimal policies, or even requiring a finite number o... |
9668eb74-0da1-47b4-b1bb-f1f85050e6aa | trentmkelly/LessWrong-43k | LessWrong | How seriously should I take the supposed problems with Cox's theorem?
I had been under the impression that Cox's theorem said something pretty strong about the consistent ways to represent uncertainty, relying on very plausible assumptions. However, I recently found this 1999 paper, which claims that Cox's result actu... |
0a976761-1929-46f8-a939-da722abecddc | trentmkelly/LessWrong-43k | LessWrong | The Goddess of Everything Else
[Related to: Specific vs. General Foragers vs. Farmers and War In Heaven, but especially The Gift We Give To Tomorrow]
They say only Good can create, whereas Evil is sterile. Think Tolkien, where Morgoth can’t make things himself, so perverts Elves to Orcs for his armies. But I think th... |
bd20dba5-161c-4eea-9f05-c9bd89a1abc9 | trentmkelly/LessWrong-43k | LessWrong | Personal experimentation: summary
Cross posted from Overcoming Bias. Comments there.
***
I asked how it could be that experimenting in my own life could be worthwhile, given that if such things were worthwhile other people should have already figured them all out. My suggested explanations:
1. I am strange: nerdy,... |
08182c2b-35be-4d19-be4b-cc37e24ebd5d | trentmkelly/LessWrong-43k | LessWrong | Toy models of AI control for concentrated catastrophe prevention
In this post, we analyze some simple games that model the problem of ensuring AIs never succeed at taking a single catastrophic action (which we'll refer to as "escaping") during deployment. This exercise highlights some of the difficulties associated wi... |
53f2ace5-4d8a-4674-a98f-fdf14ac7ca75 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Ilya Sutskever's thoughts on AI safety (July 2023): a transcript with my comments
There has been a 25 min interview with Ilya conducted by Sven Strohband and released on July 17: <https://www.youtube.com/watch?v=xym5f0XYlSc>
This interview has a section dedicated to AI safety (7 min starting from 14:56). Ilya is now... |
fcb6db25-f42c-4cac-9272-5bc5ac3dd649 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Why should we *not* put effort into AI safety research?
Many people associated with the effective altruism world believe that AI safety research is very valuable. But, there must be good work written on the other side of that debate! So, my question is: what are the best arguments that AI risks are overblown or that A... |
b4d34e25-124f-48d6-8697-92d6878994b9 | trentmkelly/LessWrong-43k | LessWrong | ReSolsticed vol I: "We're Not Going Quietly"
For the past few months I've been working on an AI generated alternate solstice remix album. It's now released on Youtube and Spotify, and should be on Apple Music soon.
ReSolsticed vol I: "We're Not Going Quietly"
My favorite genre of song is "cover that reimagines the or... |
ca7891ae-0d80-4245-9e98-71dadf7c1543 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | The alignment stability problem
The community thinks a lot about how to align AGI. It thinks less about how to align AGI so that it stays aligned for the long term. In many hypothetical cases, these are one and the same thing. But for the type of AGI we're actually likely to get, I don't think they are.
Despite some ... |
164ea5cd-3baa-46ba-9743-c0ab0963f735 | trentmkelly/LessWrong-43k | LessWrong | CoZE 2
This is part 15 of 30 of Hammertime. Click here for the intro.
Another of CFAR’s running themes is: Try Things!
> When you’re considering adopting new habits or ideas, there’s no better way to gather data than actually trying […] This is particularly important because when something does work out, you get to ... |
2bd88c2f-ceff-4c17-9117-eb1885bb0c15 | trentmkelly/LessWrong-43k | LessWrong | How Does The Finance Industry Generate Real Economic Value?
Note: This piece will use “capital” in the popular sense, i.e. as a synonym for “money”.
Plenty of people argue that some or all of the modern finance industry is engaged in zero-sum games. In particular, speculators, high-frequency traders, and broker-deale... |
12c4b7a1-41c2-4e68-bb6c-937f4bcc99b6 | StampyAI/alignment-research-dataset/blogs | Blogs | New paper: “Quantilizers”
[](https://intelligence.org/files/QuantilizersSaferAlternative.pdf)MIRI Research Fellow Jessica Taylor has written a new paper on an error-tolerant framework for software agents, “**[Quantilizers: A safer alte... |
603c8e7b-43c4-490a-8b8b-8b7b05e48acf | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Bayesians vs. Barbarians
[Previously](/lw/3h/why_our_kind_cant_cooperate):
>
> Let's say we have two groups of soldiers. In group 1, the privates are ignorant of tactics and strategy; only the sergeants know anything about tactics and only the officers know anything about strategy. In group 2, everyone at all level... |
d0f2d088-93c6-4924-8ac2-b3c46e723b0f | trentmkelly/LessWrong-43k | LessWrong | Clarifying “What failure looks like”
Thanks to Jess Whittlestone, Daniel Eth, Shahar Avin, Rose Hadshar, Eliana Lorch, Alexis Carlier, Flo Dorner, Kwan Yee Ng, Lewis Hammond, Phil Trammell and Jenny Xiao for valuable conversations, feedback and other support. I am especially grateful to Jess Whittlestone for long conv... |
eaf97840-9d7f-4ca9-bcd7-277ab956357c | trentmkelly/LessWrong-43k | LessWrong | How Much Should We Care What the Founding Fathers Thought About Anything?
A while back I saw an interesting discussion between U.S. Supreme Court Justices Stephen Breyer and Antonin Scalia. Scalia is well known for arguing that the way to deal with Constitutional questions is to use the plain meaning of the words in t... |
68b51051-751e-4d54-98e8-621dfdcfcc0d | trentmkelly/LessWrong-43k | LessWrong | Meetup : Test meetup please ignore
Discussion article for the meetup : Test meetup please ignore
WHEN: 27 March 2017 11:18:40AM (-0700)
WHERE: antartica
sorry
Discussion article for the meetup : Test meetup please ignore |
b55600b6-8323-453f-a9b9-eb9933635bf5 | trentmkelly/LessWrong-43k | LessWrong | Latent Variables and Model Mis-Specification
Posted as part of the AI Alignment Forum sequence on Value Learning.
> Rohin's note: So far, we’ve seen that ambitious value learning needs to understand human biases, and that we can't simply learn the biases in tandem with the reward. Perhaps we could hardcode a specific... |
97949dee-14ab-4e48-a7e2-b582340f18d2 | trentmkelly/LessWrong-43k | LessWrong | AI Alignment 2018-19 Review
Preamble
WHAT THIS POST IS
This is a review post of public work in AI alignment over 2019, with some inclusions from 2018. It has this preamble (~700 words), a short version / summary (~1.6k words), and a long version (~8.3k words). It is available as a Google Doc here.
There are many ar... |
e399ae5b-177a-43c4-8f3a-9eecdf930b26 | trentmkelly/LessWrong-43k | LessWrong | The AI Countdown Clock
I made this clock, counting down the time left until we build AGI:
It uses the most famous Metaculus prediction on the topic, inspired by several recent dives in the expected date. Updates are automatic, so it reflects the constant fluctuations in collective opinion.
Currently, it’s sitting ... |
09218619-21dc-4393-b926-6f071f9bd93a | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Brainstorming additional AI risk reduction ideas
It looks as though lukeprog has finished his series on how to purchase AI risk reduction. But the ideas lukeprog shares are not the only available strategies. Can Less Wrong come up with more?
A summary of recommendations from [Exploring the Idea Space Efficiently](/l... |
9a1ce84a-78ee-4faa-af7d-7f5625ef1d8b | trentmkelly/LessWrong-43k | LessWrong | Pay Other Species to Pandemize Vegetarianism for You
(This post is completely jocose. If you can't take it, don't read it. I'm making fun of Rationalists, of Me, of homo economicus, of Vegans and of things I really praise, like Consequentialism and Outsourcing. It is not serious. The Sheldon Sarcasm sign has been lift... |
021d5334-0003-457d-8fd5-6087217dca60 | trentmkelly/LessWrong-43k | LessWrong | Deducing Impact
The solution comes in the next post! Feel free to discuss amongst yourselves.
Reminder: Your sentence should explain impact from all of the perspectives we discussed (from XYZ to humans). |
d4b5d54a-fca3-44c3-a01c-37d0b9a95b3d | trentmkelly/LessWrong-43k | LessWrong | Risk Tokens: Economic Security in AI Safety
As intelligence and safety research continue to progress, I’ve been thinking more and more about how to create potential market dynamics that help with alignment and safer usage of AI. This feels especially important as we likely face cat and mouse games with frontier models... |
e7e8933b-0b33-4f06-842a-92b68d8de498 | trentmkelly/LessWrong-43k | LessWrong | Blues, Greens and abortion
Abortion is one of the most politically-charged debates in the world today - possibly the most politically charged, though that's the subject for another thread. It's an excellent way of advertising whether you are Green or Blue. As a sceptical atheist who thinks guns should be banned and ga... |
598604c6-2e05-4705-bc55-be1a42327b90 | StampyAI/alignment-research-dataset/blogs | Blogs | Privileged Bases in the Transformer Residual Stream
### [Abstract](#abstract)
Our mathematical theories of the Transformer architecture suggest that individual coordinates in the residual stream should have no special significance (that is, the basis directions should be in some sense "arbitrary" and no more likely ... |
310eb723-b868-4a64-9cc3-946488d111cb | trentmkelly/LessWrong-43k | LessWrong | Meta: Test
If I click "Save and continue" -- should it publish the post?
ETA
Yes it published it. And only afterwards the drop-down list with an option "Drafts for XiXiDu" appeared.
ETA #2
This is a bug, don't punish people with downvotes for publishing their unfinished drafts.
People don't expect their draft to ... |
5687524e-203d-4542-9d82-726bebf77feb | trentmkelly/LessWrong-43k | LessWrong | The memorization-generalization spectrum and learning coefficients
This is part I.5 of the series on “generalization spectra” and SLT. I’ll try to make it readable independently of Part I.
The purpose of this post is to reconceptualize the ideas from the last post in a “more physically principled” (and easier-to-mea... |
d443ca47-c0f2-40c7-8be3-7ca46aba5259 | trentmkelly/LessWrong-43k | LessWrong | The principle of ‘altruistic arbitrage’
Cross-posted from http://www.robertwiblin.com
There is a principle in finance that obvious and guaranteed ways to make a lot of money, so called ‘arbitrages’, should not exist. It has a simple rationale. If market prices made it possible to trade assets around and in the proces... |
0710dc8c-83fa-4433-9415-8ca47b57ad01 | trentmkelly/LessWrong-43k | LessWrong | Analysing AI Actors
Provides a different and interesting perspective for looking at the issue of AI Safety. |
3f9c4c2a-5ea6-4303-8ae1-9637e0f42607 | trentmkelly/LessWrong-43k | LessWrong | How to correct for multiplicity with AI-generated models?
I'm starting to automate my workflow with AI, but don't know where to even start with correcting for multiplicity.
Right now, my workflow is like this:
Think really hard about interesting ideas and do a surface-level examination of data -> generate hypothesis... |
c8a001a5-4362-4d9b-86a6-e78da7d6169a | StampyAI/alignment-research-dataset/arxiv | Arxiv | Theoretically Principled Trade-off between Robustness and Accuracy
1 Introduction
---------------
In response to the vulnerability of deep neural networks to small perturbations around input data [[SZS+13](#bib.bibx43)], adversarial defenses have been an imperative object of study in machine learning [[HPG+17](#bib... |
51e37412-94e5-49ad-a511-4cb44192f839 | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "I previously wrote about some practical game-theoretical (game-practical?) realizations I had while buying a house. Today I want to talk about how bureaucracy is a ritualistic, magical place.In our home-buying process, every step of the way, there were papers to be signed. Paperwork is how the magic of bureaucracy com... |
f31827a3-c14b-4ebf-8386-a13d1e7cb1d9 | trentmkelly/LessWrong-43k | LessWrong | Friendly AI ideas needed: how would you ban porn?
To construct a friendly AI, you need to be able to make vague concepts crystal clear, cutting reality at the joints when those joints are obscure and fractal - and them implement a system that implements that cut.
There are lots of suggestions on how to do this, and a... |
1c5f485a-ffc5-4b2f-adf2-a267923600fa | StampyAI/alignment-research-dataset/arbital | Arbital | Probability distribution: Motivated definition
When discussing probabilities, people will often (informally) say things like "well, the probability $\mathbb P(sick)$ of the patient being sick is about 20%." What does this $\mathbb P(sick)$ notation mean?
Intuitively, $\mathbb P(sick)$ is supposed to denote the probab... |
b3d3754c-68e3-499b-8915-9174b965e47e | trentmkelly/LessWrong-43k | LessWrong | Being hella lost as rationality practice
I've had the privilege of knowing where I was only in the sense of "within this quadrangle of an insufficiently detailed map, with impassable terrain not well indicated" several times. Usually by choice. More often while alone than prudence would dictate. I think they've been p... |
181e1e3e-9379-4b46-b73d-8a5ff48e071f | trentmkelly/LessWrong-43k | LessWrong | What are habits that a lot of people have and don't tend to have ever questioned?
Beside social norms, like shaking hands, and survival needs, like eating. |
3fd1c03f-cb46-42f9-99a7-6d70d479e467 | trentmkelly/LessWrong-43k | LessWrong | God Help Us, Let’s Try To Understand Friston On Free Energy
I’ve been trying to delve deeper into predictive processing theories of the brain, and I keep coming across Karl Friston’s work on “free energy”.
At first I felt bad for not understanding this. Then I realized I wasn’t alone. There’s an entire not-understand... |
b1ef6402-9af5-42e2-9f33-368740c75ed7 | trentmkelly/LessWrong-43k | LessWrong | Reminder: $250 LessWrong source introduction prize submissions due soon
This is just a reminder that the deadline for $250 prize for the best introduction to the LessWrong source is (Edit: not Nov. 23rd!) next Tuesday Oct. 25th. Good luck to all contestants! |
bf1b5769-cc53-4ae7-b1c7-ef1c23f54327 | StampyAI/alignment-research-dataset/blogs | Blogs | Concrete AI tasks for forecasting
This page contains a list of relatively well specified AI tasks designed for forecasting. Currently all entries were used in the 2016 Expert Survey on Progress in AI.
List
----
1. Translate a text written in a newly discovered language into English as well as a team of human exper... |
4e5502e1-6f08-4013-8486-c0162b869108 | trentmkelly/LessWrong-43k | LessWrong | Status-Regulating Emotions
Elizer Yudkowsky wrote an interesting comment on What Universal Human Experiences Are You Missing Without Realizing It?
> It was the gelling of the HPMOR hatedom which caused me to finally realize that I was blind, possibly I-don’t-have-that-sense blind, to the ordinary status-regulation em... |
2cae4205-eac5-4427-825f-beaeee66d275 | StampyAI/alignment-research-dataset/arxiv | Arxiv | CUAD: An Expert-Annotated NLP Dataset for Legal Contract Review.
1 Introduction
---------------
While large pretrained Transformers (Devlin et al., [2019](#bib.bib6); Brown et al., [2020](#bib.bib2)) have recently surpassed humans on tasks such as SQuAD 2.0 (Rajpurkar et al., [2018](#bib.bib20)) and SuperGLUE (Wang... |
6a9af7f6-3b82-4dd5-a43a-dba73d333823 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | We Shouldn't Expect AI to Ever be Fully Rational
Summary of Key Points[[1]](#fn4w8b2h9of43)
------------------------------------------
LLMs are capable of being rational, but they are also capable of being extremely irrational, in the sense that, to quote EY's [definition of rationality](https://www.lesswrong.com/pos... |
788bf28a-b819-4ce2-8b33-3900d01dcb99 | trentmkelly/LessWrong-43k | LessWrong | How Microsoft's ruthless employee evaluation system annihilated team collaboration.
A shining case study of how “instilling fear in employees to boost productivity" can backfire catastrophically.
It's a great example of how the inability to inspire, motivate, or sufficiently empathize with people (admittedly, not an... |
8e5e1706-920e-4870-a0cf-857804531aeb | trentmkelly/LessWrong-43k | LessWrong | The Outside View Of Human Complexity
One common question: how complex is some aspect of the human body? In addition to directly evaluating the available evidence for that aspect, one fruitful tactic in making this kind of prediction is to analyze past predictions about similar phenomena and assume that the outcome wil... |
90c1a696-7f97-488a-b049-a2fdfe7a14aa | trentmkelly/LessWrong-43k | LessWrong | Risk of Mass Human Suffering / Extinction due to Climate Emergency
The good folks at Extinction Rebellion are working crazy hard to promote the scientific case for impending societal and economic collapse due to climate change.
This is their talk presenting that case: Heading for Extinction and What to Do About It
... |
fe05bc74-8c02-4c6e-b119-ba9f47dfc5e1 | trentmkelly/LessWrong-43k | LessWrong | How to specify an alignment target
[Crossposted from my substack Working Through AI.]
It’s pretty normal to chunk the alignment problem into two parts. One is working out how to align an AI to anything at all. You want to figure out how to control its goals and values, how to specify something and have it faithfully ... |
7f1b8647-ffa5-4f54-9e7b-473f1d306e0f | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Alignment as Function Fitting
*In this post, I spell out some fairly basic ideas, that seem important to me. Maybe they are so basic that everyone finds them obvious, but in the spirit of* [*this piece*](https://www.lesswrong.com/posts/mLubC65xXekk5tkug/no-you-need-to-write-clearer#What_You__Personally__Need_to_Do_Dif... |
e5980290-b9ae-4705-a909-5f2068e85abd | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Bayesian Utility: Representing Preference by Probability Measures
This is a simple transformation of standard expected utility formula that I found conceptually interesting.
For simplicity, let's consider a finite discrete [probability space](http://en.wikipedia.org/wiki/Probability_space) with non-zero probability ... |
9bc94469-37e5-4cbe-ba1c-c33d26f26203 | trentmkelly/LessWrong-43k | LessWrong | [Link]: 80,000 hours blog
Some of you probably aren't aware yet of the rather excellent High Impact Careers / 80,000 hours blog.
It covers topics about how to have the biggest impact with your career, including
* how likely you are to become Prime Minister
* Decision Making under Moral Uncertainty
* Temporal Conc... |
66309b87-5033-4b77-a55b-f70162791b9a | trentmkelly/LessWrong-43k | LessWrong | The map of agents which may create x-risks
Recently Phil Torres wrote an article where he raises a new topic in existential risks research: the question about who could be possible agents in the creation of a global catastrophe. Here he identifies five main types of agents, and two main reasons why they will create a... |
a570d073-1eff-4a89-b87c-b82806813a24 | trentmkelly/LessWrong-43k | LessWrong | Trolley Problem Experiment Run IRL
https://www.youtube.com/watch?v=1sl5KJ69qiA
Neat show. Claims to be the first time the Trolley Experiment was actually run in the real world.
The host of this show set up a Trolley Problem. Subjects were convinced they were part of a focus group about commuter rail. They're placed... |
18e5eb4c-730b-4e1e-a38c-971254003389 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | AI x-risk, approximately ordered by embarrassment
Advanced AI systems could lead to existential risks via several different pathways, some of which may not fit neatly into traditional risk forecasts. Many previous forecasts, for example the well known [report](https://arxiv.org/abs/2206.13353) by Joe Carlsmith, decomp... |
813614b4-7679-41df-b4be-c1ac9a677ecc | trentmkelly/LessWrong-43k | LessWrong | The 2017 Effective Altruism Survey - Please Take!
This year, the EA Survey volunteer team is proud to announce the launch of the 2017 Effective Altruism Survey.
-
PLEASE TAKE THIS SURVEY NOW! :)
If you're short on time and you've taken the survey in prior years, you can take an abridged donations-only version of th... |
7980973a-82d2-4f0e-a9c8-9c31d79afdf8 | trentmkelly/LessWrong-43k | LessWrong | Forecasting Newsletter: November 2020
Highlights
* DeepMind claims a major breakthrough in protein folding.
* OPEC forecasts slower growth
* Gnosis announces futarchy experiment
Index
* Highlights
* In The News
* Prediction Markets & Forecasting Platforms
* United States Presidential Elections Post-mortems
*... |
649e6d97-c29b-43d4-bf59-3f6d18f6652d | trentmkelly/LessWrong-43k | LessWrong | Report & retrospective on the Dovetail fellowship
In September last year I posted an ad for a fellowship. This current post is the retrospective on how that went. Or, more accurately, it's the in-progress report of how it's going. The de-facto length of the fellowship was 3 months, but it went well enough that I'm ext... |
332d90fa-102f-4545-bda7-a7feb3a0f827 | trentmkelly/LessWrong-43k | LessWrong | September Bragging Thread
Thought I'd try reviving an old LessWrong experiment: the Bragging Thread. (See this old one for some context)
LessWrong mostly rewards people for writing blogposts. That's cool and all, but is not the only thing worth doing. What's an awesome thing you've done lately?
You are encouraged to... |
4a42fb9f-e52c-4fd9-815c-7f1a52e2ee98 | trentmkelly/LessWrong-43k | LessWrong | Intertheoretic utility comparison: examples
A previous post introduced the theory of intertheoretic utility comparison. This post will give examples of how to do that comparison, by normalising individual utility functions.
The methods
All methods presented here obey the axioms of Relevant data, Continuity, Individu... |
b9c0b0b3-3c95-4ce9-ae68-3533dd9bf524 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Permission for mind uploading via online files
Giulio Prisco [made a blog post](http://giulioprisco.blogspot.com/2010/09/mind-uploading-via-gmail.html) giving permission to use the data in his Gmail account to reconstruct an uploaded copy of him.
>
>
> To whom it may concern:
>
> I am writing this in 2010.... |
c9ff815c-fe8b-402a-8ca8-4a5cb613abb7 | StampyAI/alignment-research-dataset/arbital | Arbital | Kernel of ring homomorphism
Given a [https://arbital.com/p/-ring_homomorphism](https://arbital.com/p/-ring_homomorphism) $f: R \to S$ between [rings](https://arbital.com/p/3gq) $R$ and $S$, we say the **kernel** of $f$ is the collection of elements of $R$ which $f$ sends to the zero element of $S$.
Formally, it is $$... |
a14c5ca4-b204-4e4e-9853-49798b634462 | trentmkelly/LessWrong-43k | LessWrong | Why has the replication crisis affected RCT-studies but not observational studies?
Everyone here is probably familiar with the reproducibility crisis in psychology and various other fields. I've recently been thinking there's something very odd about all this. Namely, the reproducibility crisis seems to be almost enti... |
58575600-4ca2-4985-828a-cbaa117f32fd | trentmkelly/LessWrong-43k | LessWrong | The importance of Not Getting the Joke
> One of my favorite genres in the prestige press is the Self-Refuting Article. These are articles that contain all the facts necessary to undermine the premise of the piece, but reporters, editors, and readers all conspire together in an act of collective stupidity to Not Get th... |
dda07576-18d3-48c7-a2c4-f9c3b74a8a98 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Proposal: Using Monte Carlo tree search instead of RLHF for alignment research
Currently the most powerful techniques for getting a language model to act as an agent is via [RLHF](https://www.lesswrong.com/tag/rlhf) and similar approaches. For example, ChatGPT was trained to be an agent that tries to give humans answe... |
27dd562d-a854-450f-bff1-ddd1a4453e0e | StampyAI/alignment-research-dataset/blogs | Blogs | our deepest wishes
our deepest wishes
------------------
*there is a christian out there, who wants there to be god*
*while another would like the truth, to disbelieve or not*
*utilitarians who want to max out happiness*
*and negative ones more concerned with suffering unnoticed*
*a humble wish of luxury,... |
f0d462b9-d708-4b7c-a3b5-3ad2c1e20f9e | StampyAI/alignment-research-dataset/arxiv | Arxiv | Learning Visual Robotic Control Efficiently with Contrastive Pre-training and Data Augmentation.
I Introduction
---------------
Recent advances in deep reinforcement learning (RL) have given rise to unprecedented capabilities in autonomous decision making.
Notable successes include learning to solve a diverse set o... |
7bb6fd9e-bbfb-427a-a2d5-fe1a39bebfc1 | StampyAI/alignment-research-dataset/aisafety.info | AI Safety Info | What is instrumental convergence?
<iframe src="https://www.youtube.com/embed/ZeecOKBus3Q" title="Why Would AI Want to do Bad Things? Instrumental Convergence" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowfullscreen></iframe>
A *termin... |
4c236e44-4d38-43b7-afc7-c2d66c1385aa | trentmkelly/LessWrong-43k | LessWrong | Fundamental Uncertainty: Epilogue
I wrote a whole book! What's next?
I'm currently doing an edit pass on the entire book. I need to rewrite some of the early sections, fix some consistency issues, and generally look with fresh eyes on words I wrote months or years ago. Many of you provided helpful comments, and I'm u... |
73ef792a-87e1-4883-8771-91ff1416bb78 | StampyAI/alignment-research-dataset/youtube | Youtube Transcripts | Reinforcement Learning 8: Advanced Topics in Deep RL
let's get started
so we discussed a lot of different
topics so far item see if I can somehow
get this thing out of the way maybe not
oh well so what I thought I do today is
to go through some of the active
research topics right now I won't go
into that much breadth ... |
110b6c51-02e1-4f6f-9beb-89644178ce61 | trentmkelly/LessWrong-43k | LessWrong | Astronomy, Astrobiology, & The Fermi Paradox I: Introductions, and Space & Time
This is the first in a series of posts I am putting together on a personal blog I just started two days ago as a collection of my musings on astrobiology ("The Great A'Tuin" - sorry, I couldn't help it), and will be reposting here. Much h... |
851d1057-e717-4d76-ae0f-304663c03453 | trentmkelly/LessWrong-43k | LessWrong | Open Thread, Feb. 20 - Feb 26, 2017
If it's worth saying, but not worth its own post, then it goes here.
----------------------------------------
Notes for future OT posters:
1. Please add the 'open_thread' tag.
2. Check if there is an active Open Thread before posting a new one. (Immediately before; refresh the l... |
5f69724c-2712-49b3-8300-15164def670d | trentmkelly/LessWrong-43k | LessWrong | Looking for reading recommendations: Theories of right/justice that safeguard against having one's job automated?
Hey folks,
Can anyone recommend any books or articles on AI automation of jobs?
Specifically, books that develop or discuss a theory of right and then apply it to the question of whether we should let j... |
543eeea6-3fd7-475b-85c0-a069fe955adb | trentmkelly/LessWrong-43k | LessWrong | Why underestimate acceptable partners?
Cross posted from Overcoming Bias. Comments there.
***
The romantic view of romance in Western culture says a very small fraction of people would make a great partner for you, customarily one.
Some clues suggest that in fact quite a large fraction of people would make a suitab... |
14f55cf6-8008-44c3-bc77-2c8289bbf05d | trentmkelly/LessWrong-43k | LessWrong | What if Agent-4 breaks out?
There are several scenarios that, unlike AI 2027, explores consequences of AIs “going rogue” and freely replicating over the cloud. This post examines what these scenarios suggest might happen if OpenBrain—the leading AI company in the AI 2027 scenario—loses control of Agent-4.
I’m writing... |
906c6e7b-876c-4956-b04a-c8d7e5d42158 | trentmkelly/LessWrong-43k | LessWrong | What key nutrients are required for daily energy?
I recently found out that I was constnatly tired every day because I was getting virtually zero Vitamin D in my diet.
In addition to Vitamin D, sleep, and Fiber, what other key nutrients are necessary to have energy and a healthy brain?
Linking me to a previous maste... |
7a9de76a-bfb1-4a79-acdd-521ca736acbf | trentmkelly/LessWrong-43k | LessWrong | New LW Meetup: Bangalore
This summary was posted to LW Main on June 6th. The following week's summary is here.
New meetups (or meetups with a hiatus of more than a year) are happening in:
* Bangalore meetup: 29 June 2014 04:40PM
Irregularly scheduled Less Wrong meetups are taking place in:
* Atlanta June meetup ... |
4919b3d0-1115-4f3a-94e3-a78ad15d5586 | trentmkelly/LessWrong-43k | LessWrong | Can AIXI be trained to do anything a human can?
There is some discussion as to whether an AIXI-like entity would be able to defend itself (or refrain from destroying itself). The problem is that such an entity would be unable to model itself as being part of the universe: AIXI itself is an uncomputable entity modellin... |
d8deafa9-1419-4c14-9fbd-74287bf263a0 | trentmkelly/LessWrong-43k | LessWrong | Can AI Outpredict Humans? Results From Metaculus's Q3 AI Forecasting Benchmark
Metaculus's Q4 Benchmark series is now live. Click here to learn more and to compete for $30,000. Benchmark the state of the art in AI forecasting against the best humans on real-world questions.
Metaculus's Q3 AI Benchmarking Series aime... |
0a979a16-641c-495c-b8aa-68346a0cccbb | StampyAI/alignment-research-dataset/arxiv | Arxiv | Can Machines Learn Morality? The Delphi Experiment
1 Introduction and Motivation
------------------------------
Futurists like Nick Bostrom (bostrom\_yudkowsky\_2014), Max Tegmark (life-30-tegmark), and Stuart Russell (warn-russell-npr) warn of “super-intelligent” AI with no moral compass that could destroy humanit... |
f2800140-0e93-4773-a7c3-e3f2ff9a3ad8 | trentmkelly/LessWrong-43k | LessWrong | Weekly LW Meetups
This summary was posted to LW Main on March 18th. The following week's summary is here.
Irregularly scheduled Less Wrong meetups are taking place in:
* Ann Arbor Meetup - Meditation: 18 March 2016 07:00PM
* European Community Weekend: 02 September 2016 03:35PM
* San Francisco Meetup: Board Games... |
68e636d5-ccec-45d4-80ae-fdc2ae5c90eb | trentmkelly/LessWrong-43k | LessWrong | Some Experiments I'd Like Someone To Try With An Amnestic
A couple years ago, I had a great conversation at a research retreat about the cool things we could do if only we had safe, reliable amnestic drugs - i.e. drugs which would allow us to act more-or-less normally for some time, but not remember it at all later on... |
bacfecf4-bb3d-43e7-83bf-d219f32bc231 | trentmkelly/LessWrong-43k | LessWrong | One Hundred Billion Children’s Sky
This is the first post of my 'Death' series of essays. I wrote all the current essays in the Death series before I started posting on LW regularly, but I think it is probably one of the more important things I've created and thus I over the next week or so I'm going to be posting th... |
55410710-6635-4f74-8100-d767904eaf50 | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post3863
A putative new idea for AI control; index here . A conversation with Jessica has revealed that people weren't understanding my points about AI manipulating the learning process. So here's a formal model of a CIRL-style AI, with a prior over human preferences that treats them as an unchangeable historical ... |
217522ac-2f8c-477e-819b-764e08b39016 | trentmkelly/LessWrong-43k | LessWrong | LessWrongers at Eastercon, 22-25 April?
Given that LessWrong readers tend to read science fiction, I wonder if any of us will be at this year's Eastercon in Birmingham (UK), 22-25 April? I will be, from Friday afternoon until some time on Monday afternoon. Anyone else? Anyone wants to say hello, I look like this.
ETA... |
90423663-e7c9-495f-87ed-8f6f951a971f | trentmkelly/LessWrong-43k | LessWrong | Current AI harms are also sci-fi
A common argument I've seen for focusing on current AI harms over future ones is that the future ones are sci-fi (i.e. the ideas originated in the science fiction genre). This argument is fallacious though, because many (perhaps all?) current AI harms are also sci-fi. It is an isolated... |
ef352f67-732d-4bc8-9662-b6f111f31b28 | StampyAI/alignment-research-dataset/special_docs | Other | Institutionalizing ethics in AI through broader impact requirements
### Subjects
\* [Conferences and meetings](/subjects/conferences-and-meetings)
\* [Policy](/subjects/policy)
\* [Publishing](/subjects/publishing)
Abstract
--------
Turning principles into practice is one of the most pressing challenges of art... |
350419e4-574b-4a49-8471-03c47d69be87 | trentmkelly/LessWrong-43k | LessWrong | 2017 LessWrong Survey
The 2017 LessWrong Survey is here! This year we're interested in community response to the LessWrong 2.0 initiative. I've also gone through and fixed as many bugs as I could find reported on the last survey, and reintroduced items that were missing from the 2016 edition. Furthermore new items hav... |
43deca9b-f418-42e9-944a-fc43c059d784 | trentmkelly/LessWrong-43k | LessWrong | AI doing philosophy = AI generating hands?
I've been playing around with Stable Diffusion recently, and an analogy occurred to me between today's AI's notoriously bad generation of hands and future AI's potentially bad reasoning about philosophy.
In case you aren't already familiar, currently available image generati... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.