id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
1b98afe5-d23c-4139-9c55-e9a999147884 | trentmkelly/LessWrong-43k | LessWrong | What is learning?
I want to know what learning actually is. Ideally you can provide a model that is compact and useful as possible, such that when I encounter the word 'learn' in daily life I could replace it with this model while seeing a complete set of unique gears within it such that I can look at a real-world lea... |
e31b4037-bb21-4c39-9499-5789fdee0867 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Forecasting Thread: AI Timelines
This is a thread for displaying your timeline until human-level AGI.
Every answer to this post should be a forecast. In this case, a forecast showing your AI timeline.
For example, [here](https://elicit.ought.org/builder/BsNkKzJoc) are Alex Irpan’s AGI timelines.
How do we decide how to make decisions about the future?
How we, as a society, decide to answer this question is one of the most important moral quandaries we must face. All too often, questions of ... |
7ba1a01c-b989-453b-9e4b-c087174164aa | StampyAI/alignment-research-dataset/arbital | Arbital | 'Beneficial'
[https://arbital.com/p/3d9](https://arbital.com/p/3d9) is a [reserved term](https://arbital.com/p/9p) in [AI alignment theory](https://arbital.com/p/2v), a speaker-dependent variable to denote whatever the speaker means by, e.g., "normative" or "really actually good" or "the outcomes I want to see resulti... |
0dc32b94-60f2-4bf4-819d-dbb50d9a0cc4 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | New open letter on AI — "Include Consciousness Research"
Quick context:
* The potential development of artificial sentience [seems very important](https://forum.effectivealtruism.org/posts/cEqBEeNrhKzDp25fH/the-importance-of-artificial-sentience); it presents large, neglected, and potentially tractable risks.
* 80,00... |
da7f0507-432a-447e-8e28-d6a01413de41 | trentmkelly/LessWrong-43k | LessWrong | [AN #165]: When large models are more likely to lie
Alignment Newsletter is a weekly publication with recent content relevant to AI alignment around the world. Find all Alignment Newsletter resources here. In particular, you can look through this spreadsheet of all summaries that have ever been in the newsletter.
Aud... |
d35a978a-567d-4478-aedf-952b2a4810df | trentmkelly/LessWrong-43k | LessWrong | Understanding the diffusion of large language models: summary
5-minute summary
Time from GPT-3’s publication (May 2020) to…TimeDateModelOrganizations involvedA better model is produced[1]7 months[2]Dec 2020GopherDeepmind
A better or equally good model is open-sourced[3]
AND
A successful, explicit attempt at replica... |
f6fcccdc-f75a-467c-b585-efe7c492ebe8 | trentmkelly/LessWrong-43k | LessWrong | Mako's Notes from Skeptoid's 13 Hour 13th Birthday Stream
Recently, the pretty great skeptic podcast, Skeptoid, celebrated its 13th anniversary with a 13 hour-long livestream, in which host and founder Brian Dunning wore a party hat and received video calls from 24 of his dearest friends and peers.
It was a bit of a ... |
64c5ddcd-eb6f-4e9a-83f8-cb03645783f6 | trentmkelly/LessWrong-43k | LessWrong | cold aluminum for medicine
cold aluminum
Very pure aluminum at the boiling point of hydrogen is a very cool material. At 20K, 99.999% pure aluminum has 1000x the electrical conductivity it has at 0 C. At 4K, it has maybe 4000x.
At such low temperatures, electron free paths in aluminum become macroscopic, which is why... |
83cb19b0-2c3a-413a-bf69-593afba53ca5 | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "In Inaccessible Information, Paul Christiano lays out a fundamental challenge in training machine learning systems to give us insight into parts of the world that we cannot directly verify. The core problem he lays out is as follows.
Suppose we lived in a world that had invented machine learning but not Newtonian mech... |
a3c170b8-25aa-4fec-a19b-55e8999a7a27 | trentmkelly/LessWrong-43k | LessWrong | Transformer Modular Addition Through A Signal Processing Lens
Hello, my name is Benjamin "Frye" Kelley. This post is regarding some independent research I've been doing expanding on the work, Progress Measures for Grokking via Mechanistic Interpretability, by Neel Nanda et al. I've been trying to understand how sinu... |
46a8ea29-ff3b-4b98-abaf-a3b0f98d5d3e | trentmkelly/LessWrong-43k | LessWrong | Weekly LW Meetups: Austin, Berkeley, Budapest, Houston, Madison, Ohio/Washington DC
There are upcoming irregularly scheduled Less Wrong meetups in:
* [Ohio/Washington DC] Interest in Reason Rally meetup?: 24 March 2012 04:14PM
* Houston Meetup: 25 March 2012 01:00PM
* First meetup in Budapest: 25 March 2012 06:00P... |
e23bccfa-8685-404d-8fa9-be91ed15cbf9 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Dath Ilan's Views on Stopgap Corrigibility
*The second half of this linkpost contains significant planecrash spoilers, up through* [*Book 7, null action.*](https://www.glowfic.com/posts/6132)
> [Somewhere in the true dath ilan, carefully blurred out of satellite images by better image-editing software than is suppos... |
fef1ab68-1520-4e9d-a29a-e19827c95689 | trentmkelly/LessWrong-43k | LessWrong | Robust Artificial Intelligence and Robust Human Organizations
This academic note was previously linked but undiscussed in AN#35, and I finally got around to reading it and really liked it so wanted to highlight it with a post.
The abstract:
> Every AI system is deployed by a human organization. In high risk applicat... |
a81d4116-269e-4be6-9226-37505e0ef9d1 | trentmkelly/LessWrong-43k | LessWrong | Goal completion: the rocket equations
A putative new idea for AI control; index here.
I'm calling "goal completion" the idea of giving an AI a partial goal, and having the AI infer the missing parts of the goal, based on observing human behaviour. Here is an initial model to test some of these ideas on.
The linear ... |
ffa3ee6b-d3c6-432c-b2eb-52f0b5de5d35 | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "Traditional Rationality is phrased as social rules, with violations interpretable as cheating: if you break the rules and no one else is doing so, you're the first to defect - making you a bad, bad person. To Bayesians, the brain is an engine of accuracy: if you violate the laws of rationality, the engine doesn't run,... |
22a67d9f-3941-4947-99d9-28508796d853 | trentmkelly/LessWrong-43k | LessWrong | Experimentally evaluating whether honesty generalizes
If we train our ML systems to answer questions honestly in cases where humans can check the answer, will they generalize to behave honestly on questions where we can’t check?
I think that we could learn a lot about this question by running experiments today. I th... |
950c32db-69e8-423a-bc74-222c50f98d73 | trentmkelly/LessWrong-43k | LessWrong | Subjective implication decision theory in critical agentialism
This is a follow-up to a previous post on critical agentialism, to explore the straightforward decision-theoretic consequences. I call this subjective implication decision theory, since the agent is looking at the logical implications of their decision acc... |
9d427fe5-3a3d-4697-ae23-13d476012449 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Why GPT wants to mesa-optimize & how we might change this
This post was inspired by orthonormal's post [Developmental Stages of GPTs](https://www.lesswrong.com/posts/3nDR23ksSQJ98WNDm/developmental-stages-of-gpts) and the discussion that followed, so only part of it is original.
First I'll aim to provide a crisper v... |
fa6160d2-d1e4-44e9-b7ec-706d78b2b8d8 | trentmkelly/LessWrong-43k | LessWrong | Information bottleneck for counterfactual corrigibility
Previously, I wrote a post introducing a strategy for corrigibility based on counterfactuals.
There were various objections to this of various strengths. I don't think any defeating the notion of using this style of approach for the stop button problem specifica... |
49def551-6e6a-440e-a41d-08cfc1b26d53 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | The Science Algorithm AISC Project
I put up a project for [AISC](https://aisafety.camp/). The easiest way to evaluate your fit is to look at the [applicant questions](https://docs.google.com/document/d/1M-DMOAujWwWETkxjjAiWHRQeXXCtJm-DrAZKD6sQ5-Q/edit) I created. If the kind of reasoning required to answer these quest... |
5e1d5d1f-81ac-4fed-a562-889a0304d5c3 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Washington, D.C.: Current Events Schelling Point [To-Do List Hacking Postponed]
Discussion article for the meetup : Washington, D.C.: Current Events Schelling Point [To-Do List Hacking Postponed]
WHEN: 12 October 2014 03:00:00PM (-0400)
WHERE: National Portrait Gallery
Edit: The to-do list hacking is postp... |
3f5ab7ae-5169-45bf-85c7-6c8622b9b5d0 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Games in Kocherga club: Fallacymania, Tower of Chaos, Scientific Discovery
Discussion article for the meetup : Games in Kocherga club: Fallacymania, Tower of Chaos, Scientific Discovery
WHEN: 29 March 2017 07:40:00PM (+0300)
WHERE: Moscow, B.Dorogomilovskaya, 5-2
Welcome to Moscow LW community makeshift ga... |
bd5bf46d-e769-4326-ba38-1d79462a0840 | trentmkelly/LessWrong-43k | LessWrong | Subtext is not invariant under linear transformations
You can download the audio and PDFs from the 2007 Cognitive Aging Summit in Washington DC here; they're good listening. But I want to draw your attention to the graphs on page 6 of Archana Singh-Manoux's presentation. It shows the "social gradient" of intelligenc... |
56635229-27cf-4ecc-8fc2-bc1f23aad8e6 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | [Link] Center for the Governance of AI (GovAI) Annual Report 2018
The Center for the Governance of AI (GovAI) just released our annual report, summarising our work in 2018 and our plans for 2019.
The Center, housed at the Future of Humanity Institute, strives to help humanity capture the benefits and mitigate the ri... |
18f9ac91-b689-493a-9280-4492dc98324c | trentmkelly/LessWrong-43k | LessWrong | We should start looking for scheming "in the wild"
TLDR: AI models are now capable enough that we might get relevant information from monitoring for scheming in regular deployments, both in the internal and external deployment settings. We propose concrete ideas for what this could look like while preserving the priva... |
f0ff7986-77be-4b91-b3e5-b78f5b8cf95f | trentmkelly/LessWrong-43k | LessWrong | "Arctic Instincts? The universal principles of Arctic psychological adaptation and the origins of East Asian psychology" - Call for Reviewers (Seeds of Science)
Abstract
Modern East Asian populations plausibly retained extensive psychological adaptations in general personality traits and social dynamics to arctic envi... |
2165a077-54dd-4c59-85e7-04bcc98470bd | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Concept extrapolation for hypothesis generation
*[Posted initially](https://www.aligned-ai.com/post/concept-extrapolation-for-hypothesis-generation) on the [Aligned AI website](https://www.aligned-ai.com/). Authored by Patrick Leask, Stuart Armstrong, and Rebecca Gorman.*
There’s an [apocryphal story](https://www.gw... |
45a2555e-774f-4427-ac31-b069d92d1225 | trentmkelly/LessWrong-43k | LessWrong | I Want To Live In A Baugruppe
Rationalists like to live in group houses. We are also as a subculture moving more and more into a child-having phase of our lives. These things don't cooperate super well - I live in a four bedroom house because we like having roommates and guests, but if we have three kids and don't m... |
dbb28623-88dc-4d04-9a3e-6ed0995a888d | trentmkelly/LessWrong-43k | LessWrong | Could we use current AI methods to understand dolphins?
I am told that unsupervised machine translation is a thing. This is amazing. I ask: Could we use it to understand dolphin language? (Or whales, perhaps?)
I don't currently see a convincing reason why not. Maybe dolphins aren't actually that smart or communicativ... |
80a8ec3d-8bde-49b8-8ec0-cd7170e56469 | trentmkelly/LessWrong-43k | LessWrong | "Little glimpses of empathy" as the foundation for social emotions
(Note: I published a much-improved discussion of this topic two years later: See Symbol Grounding & Human Social Instincts. I suggest you read that one instead. By the way, I have started calling it “transient empathetic simulations” instead of “little... |
2c760035-f0a5-40bf-bf82-82f10468aeed | StampyAI/alignment-research-dataset/special_docs | Other | Human-Compatible Artificial Intelligence.
Human-Compatible Artificial Intelligence
Stuart Russell
Computer Science Division, University of California, Berk eley
1
1
Human-Compatible Artificial
Intelligence
1.1 Introduction
Artificial intelligence (AI) has as its aim the creation of intelligent mach ines. An entity
is co... |
15e629b0-44db-42e3-aa24-93acb00efd72 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Role playing game based on HPMOR in Moscow
Discussion article for the meetup : Role playing game based on HPMOR in Moscow
WHEN: 16 July 2016 03:00:00PM (+0300)
WHERE: Большая Дорогомиловская, 5к2
Welcome to our custom made role playing game "Hogwarts and methods of rationality", inspired by HPMOR story! It... |
fa48375c-aab3-4ae3-95c3-adf36c9a745d | trentmkelly/LessWrong-43k | LessWrong | Creating Space to Cultivate Skill
TLDR - Take a day to literally hide all distractions out of sight and attempt every task a second time whether you succeeded or failed
Intro
I want a ritual that does for skill cultivation what Sabbath does for rest. I want a coherent practice, one which systematically promotes s... |
d49c0aaf-1d18-40e5-97a7-95da5503e904 | trentmkelly/LessWrong-43k | LessWrong | Stupidity is also hard
Everyone today talks about Intelligence, but no one talks about the inverse problem: Stupidity. How stupid can you make a system? And what are the (mis)alignment properties of such a system?
If intelligence is the ability to efficiently achieve goals in a wide range of domains, then stupidity i... |
2a9f0464-d621-4ef8-a9a9-d46decea8ca3 | trentmkelly/LessWrong-43k | LessWrong | Stupid Questions (10/27/2014)
I think it's past time for another Stupid Questions thread, so here we go.
This thread is for asking any questions that might seem obvious, tangential, silly or what-have-you. Please respect people trying to fix any ignorance they might have, rather than mocking that ignorance.
|
2e503571-6b7f-4605-af1d-d403355d7462 | trentmkelly/LessWrong-43k | LessWrong | On expected utility, part 1: Skyscrapers and madmen
(Cross-posted from Hands and Cities)
Summary:
Suppose that you’re trying to do something. Maybe: get a job, or pick a restaurant, or raise money to pay medical bills.
Some people think that unless you’re messing up in silly ways, you should be acting “as if” y... |
f837be62-50df-4015-85f9-2d836a4b3da3 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Mathematical Mindset
I agree that [optimization amplifies](https://www.alignmentforum.org/posts/zEvqFtT4AtTztfYC4/optimization-amplifies?postId=zEvqFtT4AtTztfYC4&view=postCommentsNew) things. I also agree that a mathematical mindset is important for AI alignment. I don't, however, think that a "mathematical mindset" i... |
540ed96a-63e5-43b2-a68b-432a07d88bb2 | StampyAI/alignment-research-dataset/arbital | Arbital | Binary notation
When you were first taught how to write [numbers](https://arbital.com/p/54y) and [add](https://arbital.com/p/addition) them together, you were probably told something about the "ones place", the "tens place", the "hundreds place", and so on. Each [https://arbital.com/p/-digit](https://arbital.com/p/-d... |
cd5d2d64-34ff-4430-b157-78ff57f780eb | trentmkelly/LessWrong-43k | LessWrong | Using expected utility for Good(hart)
Goodhart's curse happens when you want to maximise an unknown or uncomputable U, and instead choose a simpler proxy V, and maximise that instead. Then the optimisation pressure on V transforms it into a worse proxy of U, possibly resulting in a very bad result from the U-perspecti... |
14861f46-3843-49b7-88ba-82e50bbcd824 | trentmkelly/LessWrong-43k | LessWrong | Official Less Wrong Redesign: Special pages
Following along from Louie's post and the discussion around it…
User:orthonormal suggested (and many seconded) a better Welcome section and improvements to the About page.
User:bentarm suggested doing something with the comment help link, User:jimrandomh suggested making it... |
7e1b2981-6811-4dd6-9bb7-a55b9785e3fd | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post279
We thank Madeline Brumley, Joe Kwon, David Chanin and Itamar Pres for their helpful feedback. Introduction Controlling LLM behavior through directly intervening on internal activations is an appealing idea. Various methods for controlling LLM behavior through activation steering have been proposed. Most st... |
56d95372-581a-45e6-86a9-3df9b245c135 | trentmkelly/LessWrong-43k | LessWrong | LLMs can strategically deceive while doing gain-of-function research
Recently Apollo Research published an article about LLMs excibit misaligned behavior and strategically deceive people when put under pressure, simulating a stock trading agent. This post is an attempt to replicate their paper, and to demonstrate sim... |
edfda044-e971-41ca-8bd2-e8605d176d28 | trentmkelly/LessWrong-43k | LessWrong | AI Safety Reading Group
If you are interested in AI Safety, come visit the AI Safety Reading Group.
The AI Safety reading group meets on Skype Wednesdays at 18:45 UTC, discussing new and old articles on different aspects of AI Safety. We start with a presentation round, then a summary of the article is presented, fol... |
29160396-617a-43f9-8f6a-4bce0ade401c | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Disentangling four motivations for acting in accordance with UDT
Introduction
------------
In this post, I examine a number of different arguments for acting like an updateless agent. Concretely, I try to provide answers to the following question: **What could move someone who isn’t already born a perfectly coherent*... |
c9b46b63-327f-4fe9-81f9-91a9f4763b39 | StampyAI/alignment-research-dataset/blogs | Blogs | A policy guaranteed to increase AI timelines
*Rick Korzekwa, April 1, 2023*
The number of years until the creation of powerful AI is a major input to our thinking about risk from AI and which approaches are most promising for mitigating that risk. While there are downsides to transformative AI arriving many years fr... |
e9b55efa-62ca-4cb6-b00a-681deb7f49df | trentmkelly/LessWrong-43k | LessWrong | IRL 6/8: Querying the human: Active Reward Learning
Every Monday for 8 weeks, we will be posting lessons about Inverse Reinforcement Learning. This is lesson 6.
Note that access to the lessons requires creating an account here.
This lesson comes with the following supplementary material:
* Gaussian Processes
Have... |
68982bc7-0c4e-427f-93b0-95a7b37d9e2d | trentmkelly/LessWrong-43k | LessWrong | Bad day contingency Dojo
Original post: http://bearlamp.com.au/dojo-bad-day-contingency-plan/
----------------------------------------
The following is an exercise I composed to be run at the Lesswrong Sydney dojos. It took an hour and a half but could probably be done faster with some adaptations that I have includ... |
74ee83ae-06fb-4868-83b5-a84b6b093ca0 | trentmkelly/LessWrong-43k | LessWrong | A Reason to Expect Republics to Perform Better than Absolute Monarchies in the Long-Term
Reading Raemon's recent Shortform on why the long-term leadership succession of small groups often deteriorates made me consider organizations which manage to successfully have several generations of competent leadership.
Reading... |
9d99520b-ef0a-467e-a9a9-7dd9b1b64ff0 | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post2487
Meta This is a story where the alignment problem is somewhat harder than I expect, society handles AI more competently than I expect, and the outcome is worse than I expect. It also involves inner alignment turning out to be a surprisingly small problem. Maybe the story is 10-20th percentile on each of th... |
5f0ab88b-0f5a-4165-a847-bd639b6869a0 | trentmkelly/LessWrong-43k | LessWrong | Good things to have learned....
I was looking at a discussion of what should be in a college curriculum, and as such discussions seem to go, there was a big list of things everyone should study, and some political claims about what's being offered but shouldn't be.
Instead, what do you wish you'd studied in college?... |
d36dc929-9a9a-4cc4-a5eb-a0cae444a1ec | trentmkelly/LessWrong-43k | LessWrong | Gaia Network: An Illustrated Primer
NB: this post has lost footnotes. Please read the version on Effective Altruism Forum with footnotes.
In our first LW post on the Gaia Network, we framed it as a solution to the challenges of building safe, transformative AI. However, the true potential of Gaia as a “world-wide web... |
3f1f2d2f-06ec-41d1-a496-147631c43404 | trentmkelly/LessWrong-43k | LessWrong | The Foundational Toolbox for Life: Introduction
(Co-authored with Sailor Vulcan)
Follows from:
What Do We Mean by Rationality?
No Really, Why Aren't Rationalists Winning?
Related resources:
Inadequate Equilibria
Instrumental Rationality sequence
How the ‘Magic: The Gathering’ Color Wheel Explains Humanity
Wh... |
77d725e0-8503-4284-a780-542d63409b91 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Humans can be assigned any values whatsoever...
.mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowr... |
a5cda109-9c91-49e7-b8d8-359f6ab8efec | trentmkelly/LessWrong-43k | LessWrong | Grey Goo Requires AI
Summary: Risks from self-replicating machines or nanotechnology rely on the presence of a powerful artificial intelligence within the machines in order to overcome human control and the logistics of self-assembling in many domains.
The grey goo scenario posits that developing self-replicating mac... |
111e7f05-ba32-432f-a7c2-429ad0d5c39c | trentmkelly/LessWrong-43k | LessWrong | Simulated Elon Musk Lives in a Simulation
All of the "Lsusr" lines were written by me unless marked "[simulated]", in which case they were written by GPT-3. All of the "Elon Musk" lines, including the final line, were written by GPT-3. The GPT-3 lines were curated. I often rolled several different responses and kept t... |
a9be70dc-0b27-4ace-9753-60c13ffd425d | trentmkelly/LessWrong-43k | LessWrong | Latent variable prediction markets mockup + designer request
I've been trying to get places like Manifold or Metaculus to add latent variables to their prediction systems. As a brief summary, latent variables would allow you to make bets and look up predictions on abstract concepts such as "How well will Ukraine do in... |
967488a1-e48a-4fd7-a5dd-c11d92eb6711 | trentmkelly/LessWrong-43k | LessWrong | Fred the Heretic, a GPT for poetry
Paul Fishwick at U of Texas at Dallas has created a poetry-generating GPT based on the poetry of Frederic Turner. I've played around with it a bit:
Fred the Heretic goes to New Orleans and appropriates some culture [Shine on Titanic]
Fred the Heretic, GPT, does poetry, channeling... |
2b6acbb8-39ba-4446-b683-b3810dc7e118 | StampyAI/alignment-research-dataset/blogs | Blogs | Historic trends in telecommunications performance
*Published February 2020*
*January 2023 note: This page contains errors that have not yet been corrected. Our overall conclusion, that there were likely no discontinuities in our metrics for telecommunications performance are likely unaffected by these errors.*
The... |
cb36851c-af66-406b-a3ac-c8095135c481 | StampyAI/alignment-research-dataset/aisafety.info | AI Safety Info | What is "coherent extrapolated volition (CEV)"?
The idea of [coherent extrapolated volition (CEV)](https://arbital.greaterwrong.com/p/cev/) is to train a superintelligence to discover, extrapolate, and combine human values and desires. This is one of Eliezer Yudkowsky’s [suggestions](https://intelligence.org/files/CEV... |
0ad9d4c1-9041-473c-9030-559d52f5c5b2 | trentmkelly/LessWrong-43k | LessWrong | What posts on finance would your find helpful or interesting?
There have been a number of post on finance and trading on LW recently. gilch and I realized that we don't have a great sense of how much people on LW know already and what they would find helpful. |
ac31cfe3-1778-4243-bd00-9201438bb41f | trentmkelly/LessWrong-43k | LessWrong | Rhythm 0 and the Absolution of Responsibility
Performance art is a little weird if we think about it. It involves you, the artist, essentially adopting the status of a zoo or circus animal, performing quaint tricks for an indulgent audience to enjoy. The only difference is, we call it art, sometimes saying the word ‘a... |
b7dcd06e-735e-46d1-b8ef-c99b62987780 | StampyAI/alignment-research-dataset/blogs | Blogs | Metabolic Estimates of Rate of Cortical Firing
Cortical neurons are estimated to spike around 0.16 times per second, based on the amount of energy consumed by the human neocortex.[1](https://aiimpacts.org/metabolic-estimates-of-rate-of-cortical-firing/#easy-footnote-bottom-1-144 "The article alternates between ‘... |
cc8ebfee-75f8-4cdc-8024-234af2cf1642 | trentmkelly/LessWrong-43k | LessWrong | [LINK] Hyperbolic discounting is rational
I just came across a paper that shows that hyperbolic discounting is the right thing to do when future discount rates are uncertain. Here is a link to the paper:
Hyperbolic discounting is rational: Valuing the far future with uncertain discount rates
by J. Doyne Farmer and Jo... |
8a24d53c-9b66-4c67-83a8-73b31708e86a | trentmkelly/LessWrong-43k | LessWrong | "Which Future Mind is Me?" Is a Question of Values
This is written right after reading Rob Bensinger's relevant post and andeslodes' comments. That discussion touched on a topic I have long held a strong belief about. I purposed first-person perspective is not physically reducible and ought to be regarded as primitiv... |
dd0b8047-5bee-4c07-8273-90300cb7adbd | trentmkelly/LessWrong-43k | LessWrong | What are fiction stories related to AI alignment?
I would have expected this question to already have been asked, but I couldn't find it. |
ace8837e-f7ed-456c-a4f5-c5281fb4433b | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: 534fd622320a8979
Prefix cache untrusted monitors: a method to apply after you catch your AI Redwood Research blog Subscribe Sign in Share this post Redwood Research blog Prefix cache untrusted monitors: a method to apply after you catch your AI Copy link Facebook Email Notes More Prefix cache untrusted monitors: a... |
03a007d4-1c79-4e99-83be-583f652acf7d | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Compendium of problems with RLHF
*Epistemic status: This post is a distillation of many comments/posts. I believe that my list of problems is not the best organization of sub-problems. I would like to make it shorter, and simpler, because cool theories are generally simple unified theories, by identifying only 2 or 3 ... |
bae225db-6bba-43b5-a4e0-2fbd60b5f755 | trentmkelly/LessWrong-43k | LessWrong | Musk on AGI Timeframes
Elon Musk submitted a comment to edge.org a day or so ago, on this article. It was later removed.
> The pace of progress in artificial intelligence (I'm not referring to narrow AI) is incredibly fast. Unless you have direct exposure to groups like Deepmind, you have no idea how fast-it is growi... |
5539cd99-3a4e-4314-bbdf-df3e2d1b72e9 | trentmkelly/LessWrong-43k | LessWrong | Bedbugs are a Solved Problem - DIY Bio-Weapon works.
Bed bugs are quite resistant to almost all pesticides; chemical extermination usually fails. Heat treatment works, but you need to heat infected areas to at least 120F. This can be pretty expensive. Landlords are required to pay for bedbug extermination, but they c... |
2bae43ea-3a39-4761-b105-411d3fafffa3 | trentmkelly/LessWrong-43k | LessWrong | On akrasia: starting at the bottom
[ Why theorizing about your addictions fails to help you combat them: an addict-thinker's version of "just do it." ]
----------------------------------------
For a few years, I've struggled to overcome several addictions. I'll list them here briefly: (nicotine, nail-biting, overeat... |
d0678c9d-0c3f-4465-bb2e-79d64437016a | trentmkelly/LessWrong-43k | LessWrong | Paul Christiano's views on "doom" (video explainer)
I have just published a video explainer of Paul Christiano's Lesswrong post: My views on "doom".
This is the second Lesswrong post I make a video explainer about, the first one being Richard Ngo's Clarifying and Predicting AGI.
Feel free to mention in the comments... |
22ba13ce-4274-4ea9-8f56-d3687ceee492 | trentmkelly/LessWrong-43k | LessWrong | “PR” is corrosive; “reputation” is not.
This is in some sense a small detail, but one important enough to be worth write-up and critique: AFAICT, “PR” is a corrupt concept, in the sense that if you try to “navigate PR concerns” about yourself / your organization / your cause area / etc., the concept will guide you tow... |
54db7a4d-54ca-4344-afa8-424b905c4cee | trentmkelly/LessWrong-43k | LessWrong | [LINK] Why Talk to Philosophers: Physicist Sean Carroll Discusses "Common Misunderstandings" about Philosophy
Why Talk to Philosophers? Part I. by philosopher of science Wayne Myrvold.
See also Sean Carroll's own blog entry, Physicists Should Stop Saying Silly Things about Philosophy.
Sean classifies the disparagin... |
68a9208b-33cb-49b8-9514-4014f464c13b | trentmkelly/LessWrong-43k | LessWrong | Qualitatively Confused
I suggest that a primary cause of confusion about the distinction between "belief", "truth", and "reality" is qualitative thinking about beliefs.
Consider the archetypal postmodernist attempt to be clever:
> "The Sun goes around the Earth" is true for Hunga Huntergatherer, but "The Earth goes ... |
9feb59a7-3bff-4f49-9d41-468fce425493 | trentmkelly/LessWrong-43k | LessWrong | Discussions for information
Information and ideas percolate through society in many forms: via research papers, media publications, schools, advocacy campaigns, and perhaps most ubiquitously, private conversations.
Private conversations fulfill purposes other than information processing and transfer, so they cannot b... |
f1752daa-8f1a-4119-bd2f-cd4cff972d1c | trentmkelly/LessWrong-43k | LessWrong | #4: Introduction to life insurance for cryonics
This is post 4 of 10 in my cryonics signup sequence, and the first of five posts on life insurance. I know that sounds like a lot. It's gonna be okay.
Introduction
Why life insurance?
There are multiple ways to fund your cryopreservation, but most people use life in... |
b34d845f-9548-4fbb-9b77-af5c576601b0 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Artificial Intelligence, Values and Alignment
Introduction
------------
The development and growth of artificial intelligence raises new and important questions for technologists, for humanity, and for sentient life more widely. Foremost among these is the question of what—or whose—values AI systems ought to align ... |
3ed16d57-b002-476c-9c4d-a2b58bccfe09 | trentmkelly/LessWrong-43k | LessWrong | Against Modal Logics
Continuation of: Grasping Slippery Things
Followup to: Possibility and Could-ness, Three Fallacies of Teleology
When I try to hit a reduction problem, what usually happens is that I "bounce" - that's what I call it. There's an almost tangible feel to the failure, once you abstract and generali... |
56bdf42e-356e-456f-937a-330e2975ea65 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Stuck In Throat
Today's post, Stuck In Throat was originally published on November 30, 2008. A summary:
> A new attempt to summarize Yudkowsky's position, and what parts of that position Hanson thinks are unsupported.
Discuss the post here (rather than in the comments to the original post).
This pos... |
eba5c4d2-2a18-44c8-9c72-c19d75919e68 | trentmkelly/LessWrong-43k | LessWrong | Imposing conditions that would have been evidence about optimal behaviour in the EEA
Warning: armchair evopsych speculation follows.
Related to: Summer vs Winter Strategies
A couple months ago, I had a large amount of tedious work to do. Whenever I sat down to do it, I would be distracted by other, less mentally... |
e7de6823-9c72-4157-b46c-5194f8384966 | trentmkelly/LessWrong-43k | LessWrong | [Link] Max Tegmark and Nick Bostrom Speak About AI Risk at UN International Security Event
http://webtv.un.org/watch/chemical-biological-radiological-and-nuclear-cbrn-national-action-plans-rising-to-the-challenges-of-international-security-and-the-emergence-of-artificial-intelligence/4542739995001
Exactly what it say... |
bd016c78-91ea-4e43-8a84-e7bb98d89663 | trentmkelly/LessWrong-43k | LessWrong | Extending control evaluations to non-scheming threats
Buck Shlegeris and Ryan Greenblatt originally motivated control evaluations as a way to mitigate risks from ‘scheming’ AI models: models that consistently pursue power-seeking goals in a covert way; however, many adversarial model psychologies are not well describe... |
8ffc9906-9e0d-450c-91d7-21bd6d55b58a | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | CHAI, Assistance Games, And Fully-Updated Deference [Scott Alexander]
**I.**
This Machine Alignment Monday post will focus on this imposing-looking article ([source](https://arbital.com/p/updated_deference/)):
[ talking about the Great Filter.
> However, recent astronomy tells us that there ar... |
0481b8ce-7305-441e-82bb-4186a0376159 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Proof of posteriority: a defense against AI-generated misinformation
Summary
=======
* A *proof of priority* shows that a piece of data must have been created *before* a certain time.
* A *proof of posteriority* shows that a piece of data must have been created *after* a certain time.
* By combining the two, you can ... |
2c4b58a0-c52a-4c10-a3e3-377eece59e2a | trentmkelly/LessWrong-43k | LessWrong | Disentangling Corrigibility: 2015-2021
Since the term corrigibility was introduced in 2015, there has been a lot of discussion about corrigibility, on this forum and elsewhere.
In this post, I have tied to disentangle the many forms of corrigibility which have been identified and discussed so far. My aim is to offer ... |
c6fc202f-d831-4d7d-8729-03b32b407169 | trentmkelly/LessWrong-43k | LessWrong | An easier(?) end to the electoral college
There's currently a campaign to do an end-run around the United States' electoral college system, the National Popular Vote Interstate Compact. (There's a lot of reasons why you might or might not want to abandon the electoral college system, which I will not discuss here.) If... |
083bd72a-1e10-432e-87ae-b89f380b6950 | trentmkelly/LessWrong-43k | LessWrong | State of Generally Available Self-Driving
After a lot of discussions online where people try to argue about where self-driving tech is going but seem pretty confused about where the tech currently is, I wanted to give a bit of an overview of the current state.
There are two main approaches: taxis and personal vehicle... |
e1df12db-9ab2-4d3c-93b8-34b9605ade83 | trentmkelly/LessWrong-43k | LessWrong | Covid 7/21/22: Featuring ASPR
The slow transition from ‘almost all about Covid’ to ‘only somewhat about Covid’ continues, with the non-Covid various-things-to-note portion this week reaching something like 25%. My presumption is that this portion will continue rising over time.
What’s happening with Covid? More of th... |
97f0e1e9-b13a-4a0e-be73-9659e4f7339a | trentmkelly/LessWrong-43k | LessWrong | [Link] An argument on colds
Source.
> It's illegal to work around food when showing symptoms of contagious diseases. Why not the same for everyone else? Each person who gets a cold infects one other person on average. We could probably cut infection rates and the frequency of colds in half if sick people didn't come... |
458917bc-29b7-444b-be97-e0f8219b6262 | StampyAI/alignment-research-dataset/special_docs | Other | Fixation patterns in simple choice reflect
optimal information sampling.
ZK[KF ZIN FZ\OIS K
Ltxktt{z |ktto~z€ tz€ty|wo ms{tmo ~oqwomt
{|ttykw tzq{~yktt{z €ky|wtzr
L~ono~tmv Ikwwk�k� OJ
4*.Fz�{zt{ Zkzrow5.\s{yk� S1M~tqqt�s�4.6
4Jo|k~�yoz� {qW��ms{w{r�. W~tzmo �{z]zt�o~�t��. W~tzmo �{z.Uo� Qo~�o�. ]zt�on [�k�o� {qFyo~tm... |
9d8b4f60-c49a-4ecb-b4e4-ccfa41f7e924 | trentmkelly/LessWrong-43k | LessWrong | Avoiding metaphysics means giving bad philosophy a free pass
I have noticed a culture among AI Alignment field builders, a certain reluctance to let conversations lead to metaphysical underpinnings. The framing and unsaid assumptions beneath models that predict the existential risk from AI is not explained, rather sel... |
1b849573-cf83-4a74-b36f-9c8ea6293d93 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Talking publicly about AI risk
In the past year, I have started talking about AI risk publicly - in mainstream newspapers, public radio, national TV, some of the most popular podcasts. The twist, and reason why you probably haven't noticed is I'm doing this in Czech. This has a large disadvantage - the positive impac... |
050522b4-a248-4d9e-87e2-13b02355cc3b | trentmkelly/LessWrong-43k | LessWrong | AIs teams will probably be more superintelligent than individual AIs
Summary
Teams of humans (countries, corporations, governments, etc) are more powerful and intelligent than individual humans. Our prior should be the same for AIs. AI organizations may then face coordination problems like management overhead, the pr... |
3183695f-cc3c-4d72-af27-8b4685e5f4ea | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Alignment Newsletter One Year Retrospective
*[Crossposted](https://www.alignmentforum.org/posts/3onCb5ph3ywLQZMX2/alignment-newsletter-one-year-retrospective) from the Alignment Forum.*
On April 9, 2018, the first Alignment Newsletter was sent out to me and one test recipient. A year later, it has 889 subscribers an... |
98604f1d-55fe-4396-a545-3bfabb6678ad | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | AXRP Episode 10 - AI’s Future and Impacts with Katja Grace
[Google Podcasts link](https://podcasts.google.com/feed/aHR0cHM6Ly9heHJwb2RjYXN0LmxpYnN5bi5jb20vcnNz/episode/MDQ5OGFmYzgtMWUzNi00NTVkLWI3NjAtYWQ4ZmEyMDRkNTNh)
This podcast is called AXRP, pronounced axe-urp and short for the AI X-risk Research Podcast. Here,... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.