id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
6bcaf8d9-b91e-45f5-b3ea-470f090879bd | trentmkelly/LessWrong-43k | LessWrong | April 2017 Media Thread
This is the monthly thread for posting media of various types that you've found that you enjoy. Post what you're reading, listening to, watching, and your opinion of it. Post recommendations to blogs. Post whatever media you feel like discussing! To see previous recommendations, check out the o... |
9f2d9c74-c982-4219-b900-40f61bc95377 | trentmkelly/LessWrong-43k | LessWrong | [Link] Marek Rosa: Announcing GoodAI
Eliezer commented on FB about a post Announcing GoodAI (by Marek Rosa GoodAIs CEO). I think this deserves some discussion as it has a quite effective approach to harness the crowd to improve the AI:
> As part of GoodAI’s development, our team created a visual tool called Brain Sim... |
121533a5-1d8a-4180-8fd4-17fb8babe90d | trentmkelly/LessWrong-43k | LessWrong | Engineering and Hacking your Mind
|
e9b0b1c9-19b6-45b1-9083-542d14a46d22 | StampyAI/alignment-research-dataset/arbital | Arbital | Stabiliser is a subgroup
Let $G$ be a [https://arbital.com/p/-3gd](https://arbital.com/p/-3gd) which [acts](https://arbital.com/p/3t9) on the [set](https://arbital.com/p/-3jz) $X$.
Then for every $x \in X$, the [stabiliser](https://arbital.com/p/4mz) $\mathrm{Stab}_G(x)$ is a [subgroup](https://arbital.com/p/-subgroup... |
00c31a08-d823-4afd-96ae-bb792838797e | trentmkelly/LessWrong-43k | LessWrong | Could billions spacially disconnected "Boltzmann neurons" give rise to consciousness?
Imagine two cerebral hemispheres, materialized in a cosmic vacuum by unlikely fluctuations, simulated in advanced virtual reality, or physically disconnected and kept alive by advanced technology.
Both hemispheres are separated, but... |
e9fb0114-eccf-46b2-9bb6-31bb0c631087 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | How truthful can LLMs be: a theoretical perspective with a request for help from experts on Theoretical CS
**Can we apply formal theory analysis to the truth of LLM outputs?**
Hi! I am Sergia. I was doing AI safety research before, interning at CHAI Berkeley, and Google, and being in and about AIS discussions. I'm re... |
fb5d81f5-1c4f-444b-b96c-3bddeb38e7e9 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Safety Aware Reinforcement Learning (SARL)
1 Introduction
---------------
Reinforcement learning (RL) algorithms have seen great research advances in recent years, both in theory and in their applications to concrete engineering problems. The application of RL algorithms extends to computer games (Mnih et al., [201... |
9ad668aa-65fb-4eaf-90b5-a9fd6fd14fa0 | trentmkelly/LessWrong-43k | LessWrong | Mobile users learning faster, in smaller bytes
|
acdafa68-afbd-49c9-a63f-9a53c151c29e | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Should we nationalize AI development?
I know very little about this topic, so please forgive me if this is an ignorant question.
It seems to me that it would be a good idea for the U.S. federal government to ban all private companies from working with sufficiently large models/building AGI (and encourage other countr... |
9dd5e9e2-9320-4ab8-9643-33b83f164d25 | trentmkelly/LessWrong-43k | LessWrong | Success and Fail Rates of Monthly Policies
Inspired by Updates from Boston.
One of the more valuable interventions I tested in the last couple years is the idea of setting a "Monthly Policy" and a "Monthly Theme." I'll share the basic process briefly, and then share a couple counterintuitive lessons with implications... |
908e8ccd-2204-4213-be0d-6b4a82833109 | trentmkelly/LessWrong-43k | LessWrong | P(doom|superintelligence) or coin tosses and dice throws of human values (and other related Ps).
TL;DR P(doom|superintelligence) is around 99% because AI researchers now don't have a precise theory of human values and precise understanding of inner structure of current AI, so they can’t encode one into the other. And ... |
10a4b717-8872-45c2-b94e-25c6b945c4ab | StampyAI/alignment-research-dataset/arxiv | Arxiv | Alignment of Language Agents
1 Introduction
---------------
Society, organizations and firms are notorious for making the mistake of *rewarding A, while hoping for B* (Kerr, [1975](#bib.bib46)), and AI systems are no exception (Krakovna et al., [2020b](#bib.bib49); Lehman et al., [2020](#bib.bib53)).
Within AI ... |
5130e375-e96e-4883-affe-ea9174c696c6 | trentmkelly/LessWrong-43k | LessWrong | Simulation Typology and Termination Risks
The main idea of the article is that we likely live in a “Fermi simulation”, that is, a simulation which is created by aliens to solve Fermi paradox via simulating possible global risks, or in a "Singularity simulation", – a simulation where future AI models its own origin (ga... |
7e9c4092-fdf1-42c1-8ac9-1b733e92d96f | trentmkelly/LessWrong-43k | LessWrong | Against unstoppable crypto prediction markets
As others have outlined, there are a million ways that prediction machines and betting markets can go wrong and do crazy things like spontaneously inducing unlikely senseless verifiable destruction. I am aware of interest in making better, faster, more expressive predictio... |
e296e1e7-0775-47f1-87f4-d17ba3adbfb3 | trentmkelly/LessWrong-43k | LessWrong | Reflexive Oracles and superrationality: prisoner's dilemma
This grew out of an exchange with Jessica Taylor during MIRI's recent visit to the FHI. Still getting my feel for the fixed point approach; let me know of any errors.
The question is, how can we make use of a reflective oracle to reach outcomes that are not N... |
2163cac5-7cd1-4c69-bcf4-dce3a6fb8a2c | trentmkelly/LessWrong-43k | LessWrong | Why not train reasoning models with RLHF?
There's a widespread assumption that training reasoning models like o1 or r1 can only yield improvements on tasks with an objective metric of correctness, like math or coding. See this essay, for example.
This assumption confuses me, because we already know how to train model... |
ae871124-b689-4850-b803-8bb0df48cb39 | trentmkelly/LessWrong-43k | LessWrong | Naked mole-rats: A case study in biological weirdness
Epistemic status: Speculative, just having fun. This piece isn't well-cited, but I can pull up sources as needed - nothing about mole-rats is my original research. A lot of this piece is based on Wikipedia.
When I wrote about “weirdness” in the past, I called mari... |
692367f7-f472-4a40-bf00-2fa86467244c | StampyAI/alignment-research-dataset/arxiv | Arxiv | Impossibility and Uncertainty Theorems in AI Value Alignment (or why your AGI should not have a utility function)
1 Introduction
---------------
Conversations about the safety of self-driving cars often turn into discussion
of “trolley problems”, where the vehicle must make a decision between several
differently di... |
f037326d-ab58-410b-9758-37d323931067 | StampyAI/alignment-research-dataset/blogs | Blogs | A Fervent Defense of Frequentist Statistics
*[Highlights for the busy: de-bunking standard “Bayes is optimal” arguments; frequentist Solomonoff induction; and a description of the online learning framework.]*
**Short summary.** This essay makes many points, each of which I think is worth reading, but if you are only... |
cfeb78b9-374f-43f8-9865-37aa5882594c | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Announcing the SPT Model Web App for AI Governance
Modeling Cooperation is excited to announce our web app [SPT Model](https://spt.modelingcooperation.com/)—an interactive tool to explore a model of AI com... |
a5877adf-99b1-40c9-a444-391148e497f5 | trentmkelly/LessWrong-43k | LessWrong | Intentionally Making Close Friends
Introduction
One of the greatest sources of joy in my life are my close friends. People who bring excitement and novelty into my life. Who expose me to new experiences, and ways of seeing the world. Who help me learn, point out my blind spots, and correct me when I am wrong. Who I ca... |
fad0c45f-92fd-4fd6-b8a9-41cee53cc91e | trentmkelly/LessWrong-43k | LessWrong | Why won't some people listen to reason?
|
3a5ead48-ab10-4de6-b3a3-565c701b7175 | StampyAI/alignment-research-dataset/special_docs | Other | Transforming Worlds: Automated Involutive MCMC for Open-Universe Probabilistic Models
Introduction Open-universe probabilistic models (OUPMs) enable Bayesian inference about the existence and attributes of latent objects underlying observed data, as well as their interconnections. Prominent applications include seismi... |
fab69072-2b63-475e-a5d1-92ee516f4bed | trentmkelly/LessWrong-43k | LessWrong | Wrong Questions
Where the mind cuts against reality's grain, it generates wrong questions—questions that cannot possibly be answered on their own terms, but only dissolved by understanding the cognitive algorithm that generates the perception of a question.
One good cue that you're dealing with a "wrong question" is ... |
7e17cce9-d50a-4ed9-b5fe-f80d0013bd30 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Longtermism Fund: August 2023 Grants Report
Introduction
------------
In this grants report, the Longtermism Fund team is pleased to announce that the following grants have been recommended by Longview... |
a970f40c-0e20-4adc-b889-621940f5a2b6 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Consequentialism is in the Stars not Ourselves
### Polished from [my shortform](https://www.lesswrong.com/posts/zhhYwM7gk8LZsDzxj/dragongod-s-shortform?commentId=rm5LqmGNmsyFQK7nX).
### Epistemic Status
Thinking out loud.
---
Introduction
============
[I've argued that system wide/total optimisation for an o... |
f3ca5c5d-fd82-45ec-ace8-523b10a052e8 | trentmkelly/LessWrong-43k | LessWrong | TSR #8 Operational Consistency
This is part of a series of posts where I call out some ideas from the latest edition of The Strategic Review (written by Sebastian Marshall), and give some prompts and questions that I think people might find useful to answer. I include a summary of the most recent edition, but it's not... |
9bf418b5-8a98-420a-ba70-e30fefb30d2d | StampyAI/alignment-research-dataset/special_docs | Other | Existential risks from a Thomist Christian perspective
Existential risks from a
Thomist Christian
perspective
Stefan Riedener
Global Priorities Institute | January 2021
GPI Working Paper No . 1-2021
Existential risks
from a Thomist Christian perspective
Stefan Riedener, University of Zu... |
b6939369-19ac-4cc9-ae2c-96dc36585135 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | The Alignment Newsletter #9: 06/04/18
**Highlights**
--------------
**[Playing hard exploration games by watching YouTube](http://arxiv.org/abs/1805.11592)** *(Yusuf Aytar, Tobias Pfaff et al)*: There are many YouTube videos demonstrating how to play levels of eg. Montezuma's Revenge. Can we use these demonstrations ... |
00aed0d4-9357-4d27-8926-b2cecd7a75bb | trentmkelly/LessWrong-43k | LessWrong | An overall schema for the friendly AI problems: self-referential convergence criteria
A putative new idea for AI control; index here.
After working for some time on the Friendly AI problem, it's occurred to me that a lot of the issues seem related. Specifically, all the following seem to have commonalities:
* Nasty... |
cdb83165-736a-4c7b-b9b0-27d2a9ecbc55 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Washington DC books meetup
Discussion article for the meetup : Washington DC books meetup
WHEN: 11 August 2013 03:00:00PM (-0400)
WHERE: National Portrait Gallery courtyard, Washington, DC 20001, USA
We'll be meeting to talk about and exchange books.
Discussion article for the meetup : Washington DC book... |
a7914803-89ba-4ef7-b90c-a32942238035 | trentmkelly/LessWrong-43k | LessWrong | Predictive model agents are sort of corrigible
TLDR: Agents made out of conditioned predictive models are not utility maximisers, and, for instance, won't try to resist certain kinds of shutdown, despite being able to generally perform well.
This is just a short cute example that I've explained in conversation enough... |
ff1df79a-03a8-4d4c-9bbf-5213c81135b3 | trentmkelly/LessWrong-43k | LessWrong | Set Up for Success: Insights from 'Naïve Set Theory'
,,,,,,,,,
Foreword
This book has been reviewed pretty thoroughly already. Rather than restate each chapter, I'll be sharing insights: some book-specific, some general.
I am quite proud of my time-to-completion for this book - just over a week, working around a ve... |
7c8eea8c-c761-46c4-99c5-0cc4c2e56e7b | trentmkelly/LessWrong-43k | LessWrong | You can tell a drawing from a painting
Epistemic status: I've had this concept for a while, and applied it in practice many times. It's fairly easy to convey in an argument, and can clear up some confusion. But there are many loose ends, and I'm not sure it's well formulated. Also, I'm stuck, so I'll just throw it out... |
3e66cdd1-a14b-4228-a987-04f6cb211734 | trentmkelly/LessWrong-43k | LessWrong | [LINK] Transcendence (2014) -- A movie about "technological singularity"
Transcendence is an upcoming movie, starring Johnny Depp:
Two leading computer scientists work toward their goal of Technological Singularity, as a radical anti-technology organization fights to prevent them from creating a world where computers... |
6ed9f5a2-cf33-4e0c-be09-30f7d74207d5 | StampyAI/alignment-research-dataset/blogs | Blogs | September 2015 Newsletter
| |
| --- |
|
**Research updates**
* New analyses: [When AI Accelerates AI](https://intelligence.org/2015/08/03/when-ai-accelerates-ai/); [Powerful Planners, Not Sentient Software](https://intelligence.org/2015/08/18/powerful-planners-not-sentient-software/)
* New at AI Impacts: [Research ... |
a571f3eb-9a85-4ac9-a23c-87656695ff1e | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Reducing profit motivations in AI development
*(Written up quite quickly — policy suggestions are intended to be illustrative, not detailed proposals. Feedback greatly appreciated!)*
Introduction
------------
In the past few days (!) we've seen a [number](https://twitter.com/SigGravitas/status/1642181498278408193) o... |
25a378a0-6d56-4fb5-9cab-47b63dadf4f0 | trentmkelly/LessWrong-43k | LessWrong | Whole Brain Emulation & DL: imitation learning for faster AGI?
None |
0d05e686-be89-4f38-b0e5-1e118a00b10f | trentmkelly/LessWrong-43k | LessWrong | Delayed Gratification vs. a Time-Dependent Utility Function
Ideally, a utility function would be a rational, perfect, constant entity that accounted for all possible variables, but mine certainly isn't. In fact, I'd feel quite comfortable claiming that no humans at the time of writing do.
When confronted with the fac... |
c378871c-04b4-43f3-a268-eb4e39fa0668 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Parametric Bounded Löb's Theorem and Robust Cooperation of Bounded Agents
1 Background and Overview
--------------------------
The arc of this paper begins and ends with a discussion of the Prisoner’s Dilemma, but it passes through a new result in provability logic. Thus, it will hopefully be of interest to game th... |
78f20598-77da-4215-a65d-7857deaa082c | trentmkelly/LessWrong-43k | LessWrong | New website contents
Rob and Federico of GWWC have made a nice summary of metacharities, which might be useful to those of you interested in cause prioritization as a cause. For future reference, I’ve put in my online collection of useful things (just for useful things which need help with their web presence).
I’ve a... |
c4cd9944-4839-485f-abbb-d707ea61bcb9 | trentmkelly/LessWrong-43k | LessWrong | Logit Prisms: Decomposing Transformer Outputs for Mechanistic Interpretability
ABSTRACT: We introduce a straightforward yet effective method to break down transformer outputs into individual components. By treating the model’s non-linear activations as constants, we can decompose the output in a linear fashion, expres... |
635ddd61-f2fb-4f48-a954-75440eabc1a6 | trentmkelly/LessWrong-43k | LessWrong | Welcome to EA Latvia [Edit With Your Details]
(The following are our suggestions for what kind of information is best to include in the welcome post of your group, feel free to replace them with whatever you think is best)
What kind of events does your group usually run? What does it usually do?
How frequently does ... |
7cbf3fc9-a12c-4d3f-8b62-a1755399afc2 | trentmkelly/LessWrong-43k | LessWrong | Embedded World-Models
(A longer text-based version of this post is also available on MIRI's blog here, and the bibliography for the whole sequence can be found here)
(Edit: This post had 15 slides added on Saturday 10th November.) |
1e36cfe0-1e0d-4cb5-b71d-61f77fababe3 | trentmkelly/LessWrong-43k | LessWrong | Belief in Belief
Carl Sagan once told a parable of someone who comes to us and claims: “There is a dragon in my garage.” Fascinating! We reply that we wish to see this dragon—let us set out at once for the garage! “But wait,” the claimant says to us, “it is an invisible dragon.”
Now as Sagan points out, this doesn’t ... |
6629f8b7-1ca6-4d75-b44d-043627e604d2 | trentmkelly/LessWrong-43k | LessWrong | AISN #16: White House Secures Voluntary Commitments from Leading AI Labs and Lessons from Oppenheimer
Welcome to the AI Safety Newsletter by the Center for AI Safety. We discuss developments in AI and AI safety. No technical background required.
Subscribe here to receive future versions.
----------------------------... |
5efb0d60-7c41-45c2-b50d-c4d57bce5aa6 | trentmkelly/LessWrong-43k | LessWrong | Alexander and Yudkowsky on AGI goals
This is a lightly edited transcript of a chatroom conversation between Scott Alexander and Eliezer Yudkowsky last year, following up on the Late 2021 MIRI Conversations. Questions discussed include "How hard is it to get the right goals into AGI systems?" and "In what contexts do A... |
3e30b37e-77b0-44b1-a310-e6ebc9c86061 | trentmkelly/LessWrong-43k | LessWrong | Balancing Audio in Software
A few months ago I ported my whistle controlled synthesizer to the Raspberry Pi. I configured it with mono output, and ran it through a DI box to get a balanced XLR output I could send to a sound board. I realized a few weeks ago, however, that I don't actually need a DI box, and I can bala... |
e77c65a5-3c97-44b6-a48b-79715c498aec | trentmkelly/LessWrong-43k | LessWrong | Final Words
Sunlight enriched air already alive with curiosity, as dawn rose on Brennan and his fellow students in the place to which Jeffreyssai had summoned them.
They sat there and waited, the five, at the top of the great glassy crag that was sometimes called Mount Mirror, and more often simply left unnamed. The... |
3112d19d-b8db-4a54-9423-2c2ace32f677 | trentmkelly/LessWrong-43k | LessWrong | The power of information?
I'm thinking about how to model an ecosystem of recursively self-improving computer programs. The model I have in mind assumes finite CPU cycles/second and finite memory as resources, and that these resources are already allocated at time zero. It models the rate of production of new inform... |
99871e28-1db0-45e7-a811-e76096c49a84 | trentmkelly/LessWrong-43k | LessWrong | Even if your Voice Shakes
I try not to talk too much about my personal life but some of my existing patrons suggested that I make this post. I’ve ended up in a sort of rough situation recently and could use some help. |
ef4d4a5a-fbd9-4b5d-be6f-75cd3d63ea02 | trentmkelly/LessWrong-43k | LessWrong | [Link] Where did you get that idea in the first place? | Meaningness
Where did you get that idea in the first place? | Meaningness |
c9cbbdb6-18fb-40e5-bda5-04fbde9e0bab | StampyAI/alignment-research-dataset/lesswrong | LessWrong | EIS IV: A Spotlight on Feature Attribution/Saliency
Part 4 of 12 in the [Engineer’s Interpretability Sequence](https://www.alignmentforum.org/s/a6ne2ve5uturEEQK7).
Thanks to Tony Wang for a helpful comment.
If you want to become more familiar with feature attribution/saliency, a tutorial on them that may offer use... |
ff672a0e-77a5-4441-a98f-21fad9268c7e | trentmkelly/LessWrong-43k | LessWrong | Notes on Fairness
This post examines the virtue of fairness. It is meant mostly as an exploration of what others have learned about this virtue, rather than as me expressing my own opinions about it, though I’ve been selective about what I found interesting or credible, according to my own inclinations. I wrote this n... |
0221e0e8-24fd-4d00-bd77-604ac5d59726 | trentmkelly/LessWrong-43k | LessWrong | This is the way the world… is
10% of the world’s countries experience violent regime change once a year on average. In middle income countries regimes last 12.5 years on average. 90% of the world’s countries could expect a political revolution, coup d’état or violent change in a human lifetime. Throughout history, tha... |
a6628163-36d4-486e-b4d2-621ec83af8b9 | trentmkelly/LessWrong-43k | LessWrong | Recommendations for donating to an anti-death cause
I've recently had the bad luck of having numerous people close to me die. Though I've wanted to contribute to anti-aging and anti-death research for a while, I'm only now in the position of being stable and materially well-off enough to throw around semi-serious cash... |
0234c283-b6bb-4bd3-b1dc-b33514ec6163 | trentmkelly/LessWrong-43k | LessWrong | Against AI Understanding and Sentience: Large Language Models, Meaning, and the Patterns of Human Language Use
Abstract
Large language models such as ChatGPT are deep learning architectures trained on immense quantities of text. Their capabilities of producing human-like text are often attributed either to mental cap... |
d1ef57ca-614b-4dd0-9c31-64d442c4679c | trentmkelly/LessWrong-43k | LessWrong | Re-formalizing PD
The Prisoner's Dilemma has been discussed to death here on OB/LW, right? Well, here's a couple new twists to somewhat... uh... expand the discussion.
Warning: programming and math ahead.
Scenario 1
Imagine a PD tournament between programs that can read each other's source code. In every match, ... |
3f95eb0b-aaba-41b4-81e3-25b7bd15306f | trentmkelly/LessWrong-43k | LessWrong | Conjecture: Internal Infohazard Policy
This post benefited from feedback and comments from the whole Conjecture team, as well as others including Steve Byrnes, Paul Christiano, Leo Gao, Evan Hubinger, Daniel Kokotajlo, Vanessa Kosoy, John Wentworth, Eliezer Yudkowsky. Many others also kindly shared their feedback and ... |
43a7f53c-f072-4493-a8f3-fb8478837603 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Research ideas to study humans with AI Safety in mind
Premise
=======
Recently I spent some time thinking about ways in which studying the human side of human-machine systems would be beneficial to build aligned AIs. I discussed these ideas informally and people seemed interested and wanted to know more. Thus, I deci... |
6be6cf93-ecc5-4a36-b2d8-091415e8e545 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Pairwise Weights for Temporal Credit Assignment.
1 Introduction
---------------
The following *umbrella problem* (Osband et al. [2019](#bib.bib12)) illustrates a fundamental challenge in most reinforcement learning (RL) problems, namely the *temporal credit assignment* (TCA) problem. An RL agent takes an umbrella a... |
58e9d7be-2eb7-47fb-9ea8-12ddd4879244 | trentmkelly/LessWrong-43k | LessWrong | Let's Design A School, Part 1
The American school system, grades K-12, leaves much to be desired.
While its flaws are legion, this post isn’t about that. It’s easy to complain.
This post is about how we could do better.
To be clear, I’m talking about redesigning public education, so “just use the X model” where X i... |
a531a2e2-d054-465d-b484-45ac47834372 | trentmkelly/LessWrong-43k | LessWrong | Proposing the Conditional AI Safety Treaty (linkpost TIME)
Technological progress can excite us, politics can infuriate us, and wars can mobilize us. But faced with the risk of human extinction that the rise of artificial intelligence is causing, we have remained surprisingly passive. In part, perhaps this was because... |
14a24360-56d0-4810-81dc-ee10e1bdea37 | trentmkelly/LessWrong-43k | LessWrong | Takeoff speeds have a huge effect on what it means to work on AI x-risk
The slow takeoff hypothesis predicts that AGI emerges in a world where powerful but non-AGI AI is already a really big deal. Whether AI is a big deal right before the emergence of AGI determines many super basic things about what we should think o... |
bc689bea-a796-403b-b1ab-92455c38d019 | trentmkelly/LessWrong-43k | LessWrong | Meetup : West LA—In Apprehending Hard Stuff
Discussion article for the meetup : West LA—In Apprehending Hard Stuff
WHEN: 20 November 2013 07:00:00PM (-0800)
WHERE: 10850 West Pico Blvd, Los Angeles, CA 90064, USA
How to get in: Go to the Westside Tavern in the upstairs Wine Bar (all ages welcome), located inside th... |
34988c1f-0352-4bc1-b3b9-c9be41be1c24 | trentmkelly/LessWrong-43k | LessWrong | Meetup : West LA Meetup-indexical and logical uncertainty
Discussion article for the meetup : West LA Meetup-indexical and logical uncertainty
WHEN: 29 January 2014 02:32:41PM (-0800)
WHERE: 10850 West Pico Blvd, Los Angeles, CA 90064, USA
How to get in: Go to the Westside Tavern in the upstairs Wine Bar (all ages ... |
2dca1b44-97d3-4d49-a0e6-6212ae30bb38 | trentmkelly/LessWrong-43k | LessWrong | Metacompilation
A post that is going to be part of my sequence on rethinking programming languages.
The baisc idea
A compiler is a piece of machine code C, that takes as input a text string describing a program p and returns the compiled machine code C(p)
Let Opt be a function that takes in a machine code program a... |
3c4363b0-e9b0-4c38-9ecc-13344968e974 | trentmkelly/LessWrong-43k | LessWrong | The time I got really into poker
One time I decided it would be good to learn to play poker. I had probably learned to play some form of poker a couple of other times before, and forgotten. One way to play a game a lot is to play it with a computer rather than other people. An iPad turns Agricola from one of the slowe... |
93521951-13d8-4136-9b64-e71ac9fa4c58 | trentmkelly/LessWrong-43k | LessWrong | [LINK] System 2 thinking decreases religious belief
This experiment, to be published in Science, used priming (cues like hard-to-read fonts, showing participants the sculpture The Thinker) and problem-solving tasks to induce "analytical thinking" in the participants, and found that it seemed to reduce their degree of ... |
867ca2e7-24bd-4908-8b78-8c2078c58184 | trentmkelly/LessWrong-43k | LessWrong | Rationality Reading Group: Part V: Value Theory
This is part of a semi-monthly reading group on Eliezer Yudkowsky's ebook, Rationality: From AI to Zombies. For more information about the group, see the announcement post.
----------------------------------------
Welcome to the Rationality reading group. This fortnigh... |
4d3c8a48-dba9-4728-a652-2c3e9552be57 | trentmkelly/LessWrong-43k | LessWrong | On the UBI Paper
Would a universal basic income (UBI) work? What would it do?
Many people agree July’s RCT on giving people a guaranteed income, and its paper from Eva Vivalt, Elizabeth Rhodes, Alexander W. Bartik, David E. Broockman and Sarah Miller was, despite whatever flaws it might have, the best data we have so... |
9b2c79e7-b025-405a-9238-e5f49da0099f | trentmkelly/LessWrong-43k | LessWrong | Meetup : Vancouver!
Discussion article for the meetup : Vancouver!
WHEN: 23 February 2013 03:00:00PM (-0800)
WHERE: 2505 west broadway, vancouver
Our recent format has been working well, so we'll continue. Please come out and join us at 2505 west broadway (Benny's Bagels) at 15:00 on Saturday.
Topic is undetermine... |
546f94e1-1b9e-4376-94bf-ba402e67402e | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | The Gears of Impact
 
) recently opened [applications](https://airtable.com/shr3XO3xZVR9HNPS2) for the second iteration of the [M... |
54918564-b54e-4c32-bd8b-325c2ad00c86 | trentmkelly/LessWrong-43k | LessWrong | Meetup Tactics Open Thread
I think we need to have more discussion of meetup tactics on LW. At my local meetup, we've been feeling a bit lost about what works best, so I hereby propose that we have semi-regular meetup tactics discussions like the open and quotes threads.
So here's a few questions to start us off:
*... |
86590ecc-398f-441a-88c9-e2e8a94af954 | trentmkelly/LessWrong-43k | LessWrong | Experiment Idea: RL Agents Evading Learned Shutdownability
Preface
Produced as part of the SERI ML Alignment Theory Scholars Program - Winter 2022 Cohort. Thanks to Erik Jenner who explained to me the basic intuition for why an advanced RL agent may evade the discussed corrigibility measure. I also thank Alex Turner, ... |
9e931e7b-88fa-4afb-abcc-4c5e62ec6668 | trentmkelly/LessWrong-43k | LessWrong | METAPHILOSOPHY - A Philosophizing through logical consequences
CAUSALITY & LOGICAL CONSEQUENCES
Cause and effect are not universal, but relative.
On the other hand, the universality of cause and effect is only relative, not absolute, because wherever it is (universally)
RELATIVELY ADAPTABLE
So something is relativ... |
9378cb03-ff08-4139-b89e-d52b46cd82f2 | trentmkelly/LessWrong-43k | LessWrong | Common Probability Distributions
When we output a forecast, we're either explicitly or implicitly outputting a probability distribution.
For example, if we forecast the AQI in Berkeley tomorrow to be "around" 30, plus or minus 10, we implicitly mean some distribution that has most of its probability mass between 20 a... |
46eb8caa-2825-402d-802e-b46482c3168d | trentmkelly/LessWrong-43k | LessWrong | Is GPT3 a Good Rationalist? - InstructGPT3 [2/2]
Summary
* IGPT3 mimics two important human epistemic biases:
* IGPT3 exhibits motivated reasoning.
* IGPT3 exhibits a form of bad faith.
* IGPT3 has a lot of knowledge of the world. And in a specific setup, I show that it’s quite hard to get it to tell its knowl... |
adf3e58f-6d21-4851-af06-1aabd5db74cc | StampyAI/alignment-research-dataset/arxiv | Arxiv | A Technique to Create Weaker Abstract Board Game Agents via Reinforcement Learning
1 Introduction
---------------
Reinforcement Learning (RL) is a powerful model to create learning agents for complex problems. It is a research sub-area in the broader field of artificial intelligence (AI) where, in RL, an agent atte... |
34980490-63b6-444b-9831-7b7dc71ac951 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Predictable updating about AI risk
(Cross-posted from [my website](https://joecarlsmith.com/2023/05/08/predictable-updating-about-ai-risk). Podcast version [here](https://www.buzzsprout.com/2034731/12809255-predictable-updating-about-ai-risk), or search "Joe Carlsmith Audio" on your podcast app.)
> *"This present mo... |
3ae955cd-65c2-4af3-8ae6-ba4aff529850 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Linear Mode Connectivity and the Lottery Ticket Hypothesis
Linear Mode Connectivity and the Lottery Ticket Hypothesis
Jonathan Frankle1Gintare Karolina Dziugaite2Daniel M. Roy3 4Michael Carbin1
Abstract
We study whether a neural network optimizes to
the same, linearly connected minimum under dif-
ferent samples of SGD... |
ba40d6c1-6c47-4d91-8f87-ec9758f56eeb | trentmkelly/LessWrong-43k | LessWrong | EA & LW Forums Weekly Summary (14th Nov - 27th Nov 22')
Supported by Rethink Priorities
This is part of a weekly series summarizing the top posts on the EA and LW forums - you can see the full collection here. The first post includes some details on purpose and methodology. Feedback, thoughts, and corrections are wel... |
a2b63602-2b9e-4f44-aa7e-68372863c08a | StampyAI/alignment-research-dataset/arbital | Arbital | Group orbit
When we have a [group](https://arbital.com/p/3gd) [acting](https://arbital.com/p/3t9) on a set, we are often interested in how the group acts on a particular [element](https://arbital.com/p/group_element). One natural way to do this is to look at the set of all the elements that different group actions wil... |
ef3ce695-06ee-4fd3-a422-d701606d3559 | trentmkelly/LessWrong-43k | LessWrong | The Vostok Ice Cores Revisited pt. I
An Investigation
Introduction
It's impressive how fast the focus in things climate change has recently shifted from Thunberg back to Milankovic again in just a few weeks.
Even more impressive is the fact that, though astronomic or cosmic cycles and positions can be predicted with... |
087e3fd0-d7f2-4d3a-ad9f-8ddb1f7df074 | trentmkelly/LessWrong-43k | LessWrong | Self-fulfilling correlations
Correlation does not imply causation. Sometimes corr(X,Y) means X=>Y; sometimes it means Y=>X; sometimes it means W=>X, W=>Y. And sometimes it's an artifact of people's beliefs about corr(X, Y). With intelligent agents, perceived causation causes correlation.
Volvos are believed by man... |
251e61e6-7bfb-42d3-85d7-34807c72942b | trentmkelly/LessWrong-43k | LessWrong | Thoughts about Dr Stone and Mythology
I highly recommend the anime/manga Dr Stone. It's a science-backed adventure in a world where modern technology has been lost and needs to be resurrected. If you care about technology, you'll probably enjoy it quite a lot.
There's something you should know before setting off. The... |
c2d964cd-0b1a-40f9-80ac-3cec2454111a | trentmkelly/LessWrong-43k | LessWrong | Psychological skills
|
01baf76c-51f9-4935-be89-7613a394f17d | trentmkelly/LessWrong-43k | LessWrong | Meetup : Sydney Meetup - September
Discussion article for the meetup : Sydney Meetup - September
WHEN: 24 September 2014 06:30:06PM (+1000)
WHERE: Sydney City RSL, 565 George St, Sydney, Australia 2000
6:30 PM for early discussion 7PM general dinner-discussion after dinner we'll have our rationality exercise and a... |
a771c063-f26a-4dd1-ac09-0f709996e0d8 | trentmkelly/LessWrong-43k | LessWrong | AI arms race
Racing to the Precipice: a Model of Artificial Intelligence Development
by Stuart Armstrong, Nick Bostrom, and Carl Shulman
This paper presents a simple model of an AI arms race, where several development teams race to build the first AI. Under the assumption that the first AI will be very powerful and ... |
0f95b03a-bd94-4711-a74d-3a6e32375b8d | trentmkelly/LessWrong-43k | LessWrong | Gains from trade: Slug versus Galaxy - how much would I give up to control you?
Edit: Moved to main at ThrustVectoring's suggestion.
A suggestion as to how to split the gains from trade in some situations.
The problem of Power
A year or so ago, people in the FHI embarked on a grand project: to try and find out if t... |
481a57bf-4039-4558-be24-0a649ce6ea05 | trentmkelly/LessWrong-43k | LessWrong | 0. The Value Change Problem: introduction, overview and motivations
Given that AI systems are becoming increasingly capable and widely deployed, we have to think carefully about the multifarious effects their adoption will have on our individual and collective lives. In this sequence, I will focus specifically on how... |
81bbd8ab-dbb6-4411-8b1c-6fdab2f65967 | trentmkelly/LessWrong-43k | LessWrong | Does decidability of a theory imply completeness of the theory?
I know right now that completeness implies decidability of a theory, but the question is essentially the converse of this:
Given an arbitrary formal theory that is somehow decidable by a Turing Machines, in the sense that for any sentence in the theory, ... |
7ac4c908-7471-46b5-b638-879af4758e7e | trentmkelly/LessWrong-43k | LessWrong | On-line google hangout on approaches to the control problem (2017/6/13 7PM UTC)
I'd like to get more discussion on-line about approaches to the control problem, so I'm hosting a hangout.
I'll run it lean coffee style meeting with a broad theme of what to do about the control problem. So people propose topics, we vote... |
01fef7dc-da19-44c7-855d-f9dc02ac18f9 | trentmkelly/LessWrong-43k | LessWrong | Ritual Report: NYC Less Wrong Solstice Celebration
Note: Secular Solstice has evolved a bit since this original post (most noteably, it no longer has a major Lovecraft theme.
Last Friday, the NYC Less Wrong community held their first Winter Solstice Celebration. Approximately twenty of us gathered for dinner and a n... |
0c91fd5f-80e7-4a36-933b-e51652dbdd2b | trentmkelly/LessWrong-43k | LessWrong | Demystifying "Alignment" through a Comic
Disclaimer: This explanatory comic is not specifically aimed at the Less Wrong contributor.
|
58f2da76-1701-496f-9acf-6c4a91e835da | StampyAI/alignment-research-dataset/special_docs | Other | Can Humans Do Less-Than-One-Shot Learning?.
Can Humans Do Less-Than-One-Shot Learning?
Maya Malaviya1,*, Ilia Sucholutsky2,*, Kerem Oktar1, Thomas L. Griffiths1,2
1Department of Psychology, Princeton University
2Department of Computer Science, Princeton University
*Equal contribution.
{mayamala, is2961, oktar, tomg }@... |
e93a0a24-8388-498e-93bc-fba0283cefab | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] The Meditation on Curiosity
Today's post, The Meditation on Curiosity was originally published on 06 October 2007. A summary (taken from the LW wiki):
> If you can find within yourself the slightest shred of true uncertainty, then guard it like a forester nursing a campfire. If you can make it blaze up... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.