id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
c66b1a77-7780-41f2-9d16-2445eedad9d4 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | AI Safety Needs Great Product Builders
In his [AI Safety Needs Great Engineers](https://forum.effectivealtruism.org/posts/DDDyTvuZxoKStm92M/ai-safety-needs-great-engineers) post, Andy Jones explains how software engineers can reduce the risks of unfriendly artificial intelligence. Even without deep ML knowledge, these... |
333013c5-9779-4c9a-bb08-5fbf8c49f85d | trentmkelly/LessWrong-43k | LessWrong | Idealism, Realistic & Pragmatic
Ideal is something considered perfect.
Perfection is not a final step but achieved through a process.
Realism is necessary, meaning in its implementation, it is adjusted according to capability.
Don't forget, even if something is done within one's capabilities, if it lacks utility or... |
9c221056-4356-4934-98b8-b6d6916916d8 | trentmkelly/LessWrong-43k | LessWrong | EIS XIV: Is mechanistic interpretability about to be practically useful?
Part 14 of 12 in the Engineer’s Interpretability Sequence.
Is this market really only at 63%? I think you should take the over.
Only 63%? I think you should take the over.
Five tiers of rigor for safety-oriented interpretability work
Lately,... |
e3179640-b720-41ee-80bd-16d9e83470ba | trentmkelly/LessWrong-43k | LessWrong | Stable self-improvement as a research problem
“Stable self-improvement" seems to be a primary focus of MIRI’s work. As I understand it, the problem is "How do we build an agent which rationally pursues some goal, is willing to modify itself, and with very high probability continues to pursue the same goal after modifi... |
766eba97-9df3-47d6-956f-613e5e7cc551 | trentmkelly/LessWrong-43k | LessWrong | Open & Welcome Thread - January 2023
If it’s worth saying, but not worth its own post, here's a place to put it.
If you are new to LessWrong, here's the place to introduce yourself. Personal stories, anecdotes, or just general comments on how you found us and what you hope to get from the site and community are invit... |
726e4371-072c-47db-a332-2e48d62f10a2 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Washington, D.C.: Availability Heuristic
Discussion article for the meetup : Washington, D.C.: Availability Heuristic
WHEN: 04 October 2015 03:00:00PM (-0400)
WHERE: National Portrait Gallery
Meetup is moving to the Luce Center due to an event in the courtyard.
Crossposted from mailing list. Meeting in co... |
f95237dd-22cb-4cb4-9554-9ad19a66c0dd | trentmkelly/LessWrong-43k | LessWrong | Corrigibility, Self-Deletion, and Identical Strawberries
This post is a follow-up to my post from yesterday (LW version) about LLM-based AI. Here I want to share some exchanges I’ve had with LLMs that I think support that post’s conjectures B and C. To paraphrase, those conjectures were “safety depends on the prompt” ... |
ed302f56-e022-4abf-9e4f-22f1a07cf339 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Distinguishing test from training
Various people propose some variant of “align AGI by making it sufficiently uncertain about whether it’s in the real world versus still in training”. (The idea being that, insofar as it's in training, betraying the apparent-operators predictably gets high loss; so if the AGI is uncert... |
3bb845e9-84d9-4974-9b34-b1f28e2af365 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Teleosemantics!
I wanted to write a long, detailed, analytic post about this, somewhat like my [Radical Probabilism](https://www.lesswrong.com/posts/xJyY5QkQvNJpZLJRo/radical-probabilism-1) post (to me, this is a similarly large update). However, I haven't gotten around to it for a long while. And perhaps it is better... |
6246b12c-6b6d-4c35-a993-9b8fdb8dfa5a | trentmkelly/LessWrong-43k | LessWrong | Autonomous Alignment Oversight Framework (AAOF)
Abstract:
To align advanced AIs, an ensemble of diverse, transparent Overseer AIs will independently monitor the target AI and provide granular assessments on its alignment with constitution, human values, ethics, and safety. Overseer interventions will be incremental an... |
4e43656f-fe2a-47ac-b389-23557861a139 | trentmkelly/LessWrong-43k | LessWrong | Let’s think about slowing down AI
Averting doom by not building the doom machine
If you fear that someone will build a machine that will seize control of the world and annihilate humanity, then one kind of response is to try to build further machines that will seize control of the world even earlier without destroying... |
80aa3385-1082-4e84-b11f-bd7dd92835d0 | trentmkelly/LessWrong-43k | LessWrong | Building an Inexpensive, Aesthetic, Private Forum
Reason for Writing: I want to gather information for a friend who is a professional in the behavioral sciences.
Epistemic Status: Less than 2 hours googling how to build a forum, and a couple youtube videos
Requirements:
1. Aesthetic
2. Inexpensive
3. Private
4. ... |
00ac2873-662d-45ab-99b2-e12c9731ccc0 | trentmkelly/LessWrong-43k | LessWrong | Phonosemantic Duplication
I would like to try to propose a potential novel terminology for a motif found in several languages. The following is a short description of the idea.
Phonosemantic Duplication is a linguistic phenomenon where words that represent duplicity, duality, or redundancy exhibit clear internal phon... |
f793b502-a2f8-4d16-ba2a-2bde5d7697e7 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Learning under Misspecified Objective Spaces
1 Introduction
---------------
Progress in robot motion planning and reinforcement learning has enabled faster and more efficient optimization of pre-specified objective functions. This progress on *how* to solve optimization problems has opened up new questions about *w... |
aec4192e-b5b4-4fa0-aa17-cdc2dd68c72b | StampyAI/alignment-research-dataset/special_docs | Other | Rational use of cognitive resources in human planning. Nature Human Behaviour,.
Articles
https:/ / doi.org/10.1038/s41562-022-01332-81Department of Psychology, Princeton University, Princeton, NJ, USA. 2Department of Psychology, University of California, Berkeley, CA, USA. 3Department
of Cognitive Sciences, Universit... |
3a262eaa-7be3-4f06-9f49-c20c81d12ce5 | trentmkelly/LessWrong-43k | LessWrong | Jokes Thread
This is a thread for rationality-related or LW-related jokes and humor. Please post jokes (new or old) in the comments.
------------------------------------
Q: Why are Chromebooks good Bayesians?
A: Because they frequently update!
------------------------------------
A super-intelligent AI walks out ... |
3bc9fc7d-aa0f-4877-95e8-4d31a8e051a3 | StampyAI/alignment-research-dataset/arbital | Arbital | Why is the decimal expansion of log2(3) infinite?
$\log_2(3)$ starts with
1.58496250072115618145373894394781650875981440769248106045575265454109822779435856252228047491808824209098066247505916734371755244106092482214208395062169829949365759223858523444158253630274768530697805168759955447372668346246123642488500475818... |
62bc3d12-5bd6-44de-a0b2-27f175f0ac34 | trentmkelly/LessWrong-43k | LessWrong | Meetup : San Francisco Meetup: Board Games
Discussion article for the meetup : San Francisco Meetup: Board Games
WHEN: 20 April 2015 06:15:00PM (-0700)
WHERE: 1390 Market St., San Francisco, CA
We'll be meeting to hang out and play board games!
This is at our apartment because we have a table, call me at 301-458-0... |
aa6c8d74-a2c4-45de-ac6b-26a2f946cdc1 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Taking a leave of absence from Open Philanthropy to work on AI safety
I’m planning a leave of absence (aiming for around 3 months and potentially more) from Open Philanthropy, starting on March 8, to explore working directly on AI safety.
I have a few different interventions I might explore. The first I explore wi... |
d47a10e6-b801-46f1-b011-ea540c69ef05 | trentmkelly/LessWrong-43k | LessWrong | What are you working on?
Whpearson recently mentioned that people in some other online communities frequently ask "what are you working on?". I personally love asking and answering this question. I made sure to ask it at the Seattle meetup. However, I don't often see it asked here in the comments, so I will ask it:
W... |
385452cd-c271-4372-8eb1-6c1c2380df2a | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] The Martial Art of Rationality
Today's post, The Martial Art of Rationality was originally published on November 22, 2006. A summary (taken from the LW wiki):
> Rationality is the martial art of the mind, building on universally human machinery. But developing rationality is more difficult than developin... |
5bc8bb88-197a-48b1-8811-676b23cfb7f6 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Transformer trained on it's own content?
Please forgive my lack of knowledge and inexperience on these topics. As I'm interested in the intersection of AI, Jungian studies, and the human condition, I've stumbled on this community. I'm continually impressed with the scope and depth of thought I find here. I dont have m... |
9152e3b3-c40a-4402-a4d4-565b2edc628c | trentmkelly/LessWrong-43k | LessWrong | Meetup : Boston / Cambridge - Systems, Leverage, and Winning at Life
Discussion article for the meetup : Boston / Cambridge - Systems, Leverage, and Winning at Life
WHEN: 24 November 2013 02:00:00PM (-0500)
WHERE: Citadel, 98 Elm St, Apt 1, Somerville, MA, 02144
Satvik Beri will be presenting on Systems, Leverage, ... |
87c4b05d-5c62-42c3-98bb-ed8543d42a4d | trentmkelly/LessWrong-43k | LessWrong | Senelda - Reputed IT training Academy in Chennai
|
6b68b739-82a3-4461-81d4-ea448bd82b3f | trentmkelly/LessWrong-43k | LessWrong | Theories That Can Explain Everything
It's generally accepted here that theories are valuable to the extent that they provide testable predictions. Being falsifiable means that incorrect theories can be discarded and replaced with theories that better model reality (see Making Beliefs Pay Rent). Unfortunately, reality ... |
407a0235-0717-46d1-bcad-10580e59f99f | trentmkelly/LessWrong-43k | LessWrong | "Newton's laws" of finance
[NOT financial advice! In fact I'll really be showing some generic math here – and it's up to you if you apply it to investments in stocks, projects, relationships, ideas, skills, etc.]
As I’ve been thinking about some economic theories and working on a personal investment algorithm, I kep... |
b1acc3e9-bc50-49c0-98fc-02af09554e21 | trentmkelly/LessWrong-43k | LessWrong | Corrigibility doesn't always have a good action to take
In a previous critique of corrigibility, I brought up the example of a corrigible AI-butler that was in a situation where it was forced to determine the human's values through its actions - it had no other option.
Eliezer pointed out that, in his view of corrigi... |
e7677511-fdc3-438a-a2d1-cc4a34a9aec5 | trentmkelly/LessWrong-43k | LessWrong | Fuzzy Boundaries, Real Concepts
Summary: Certain basic concepts are still very useful, even if they have fuzzy or contested boundaries, or break down in edge cases. This is basically just working out a few examples of The Cluster Structure of Thingspace. Two important examples are honesty and consent.
Adam: I can't b... |
71dfa6cc-0297-4fb0-8744-8177fd9bf2fd | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | [Job Ad] SERI MATS is hiring for our summer program
**TL;DR:** [SERI MATS](https://www.serimats.org/) islooking for a Community Manager, 1-3 Scholar Support Specialists, and 4 operations staff to support our next cohort of AI safety researchers. [Apply](https://airtable.com/shrn7rmHWssw0PC7J) by 6/12 to help grow the ... |
06bb9383-f5fd-4423-abac-2e61482ffe6c | trentmkelly/LessWrong-43k | LessWrong | Some thoughts on decentralised prediction markets
**Thought experiment 1 – arbitrage opportunities in prediction market**
You’re Mitt Romney, biding your time before riding in on your white horse to win the US republican presidential preselection (bear with me, I’m Australian and don’t know US politics). Anyway, you’... |
f5d49cac-1092-407f-be86-cf85b2c71263 | trentmkelly/LessWrong-43k | LessWrong | The best way so far to explain AI risk: The Precipice (p. 137-149)
When it comes to explaining AI risk to someone for the first time, it seems to me like it's the kind of thing where there's just a ton of ways to mess up and do it wrong. The human brain might just be really bad at explaining concepts that are complica... |
f59f7d93-4edf-4c56-84b2-927293c1ba6b | StampyAI/alignment-research-dataset/lesswrong | LessWrong | A terrifying variant of Boltzmann's brains problem
I am new to this forum even though I have read a lot on the subject of boltzmann's brains and have come to a terrifying conclusion, the classic argument which for example Sean Carroll used, "The cognitive instability" (<https://arxiv.org/abs/1702.00850>), is not enoug... |
c135f598-cfe7-4839-bb62-ec817614aa7d | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | How much do markets value Open AI?
Summary: A BOTEC indicates that Open AI might have been valued at 220-430x their annual recurring revenue, which is high but not unheard of. Various factors make this multiple hard to interpret, but it generally does not seem consistent with investors believing that Open AI will capt... |
70de7260-7574-4fe8-9e1d-bf438363d56a | trentmkelly/LessWrong-43k | LessWrong | dalle2 comments
(i.)
On April 6, OpenAI announced “DALLE-2” to great fanfare.
There are two different ways that OpenAI talks to the public about its models: as research, and as marketable products.
These two ways of speaking are utterly distinct, and seemingly immiscible. The OpenAI blog contains various posts writt... |
36dc91a6-9094-4816-b40f-571952f90557 | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post306
Summary: If we validate automated alignment research through empirical testing, the safety assurance work will still need to be done by humans, and will be similar to that needed for human-written alignment algorithms. Three levels of automated AI safety Automating AI safety means developing some algorithm... |
80736054-03cd-4ce9-b87d-dfe0d42e513c | trentmkelly/LessWrong-43k | LessWrong | AI Model Registries: A Regulatory Review
This article is the third in a series of ~10 posts comprising a 2024 State of the AI Regulatory Landscape Review, conducted by the Governance Recommendations Research Program at Convergence Analysis. Each post will cover a specific domain of AI governance (e.g. incident reporti... |
8c6f4726-4ce4-47b8-a762-82e369225911 | trentmkelly/LessWrong-43k | LessWrong | How much is karma worth, after all?
It's been a couple days since the funding plea, so I thought I'd like to take this chance to compare self-reported donations to short-term karma gains. Naturally, I voted on none of these comments. Note that after posting this, the karma on these posts will almost definitely change;... |
c723714a-20c6-4245-a4cc-fac967f49d5c | trentmkelly/LessWrong-43k | LessWrong | Why I'm Worried About AI
Preamble
In this sequence of posts, I want to lay out why I am worried about risks from powerful AI and where I think the specific dangers come from. In general, I think it's good for people to be able to form their own inside views of what’s going on, and not just defer to people. There are s... |
a633c784-4fb3-4b13-8d07-fccdeaf107e1 | trentmkelly/LessWrong-43k | LessWrong | An Intuitive Introduction to Causal Decision Theory
Like any decision theory, Causal Decision Theory (CDT) aims to maximize expected utility; it does this by looking at the causal effects each available action in a problem has. For example, in Problem 1, taking box A has the causal effect of earning $100, whereas taki... |
f02b32ad-aefb-4be4-9428-fd280bc42aea | trentmkelly/LessWrong-43k | LessWrong | "The Singularity Is Nearer" by Ray Kurzweil - Review
Over the past week I read Ray Kurzweil's new book "The Singularity Is Nearer". This is my review.
First off, if you've read The Singularity Is Near, which was published 19 years ago in 2005, you should be aware that the sequel book is a lot less technical. The Sing... |
59dac211-15b0-4739-92a4-9f202ca57897 | trentmkelly/LessWrong-43k | LessWrong | Proposal: Butt bumps as a default for physical greetings
As Tim Urban explains in The Great Perils of Social Interaction, greetings are hard.[1]
> I’ll be 90 and I still won’t have figured this out. There are different rules for everyone and nothing’s clear—Do I shake my grandfather’s hand or go for the hug? How ab... |
a1851ad2-c7e2-4c9b-aec3-b1be6f35af0b | trentmkelly/LessWrong-43k | LessWrong | You can now log in with your LW1 credentials on LW2
I just finished an update that ported over the old passwords from LW1 while still preserving security (see my previous meta post). We only ported passwords over from sometime in last May, and will run the same import on the new database if the vote completes and we a... |
d357f0c8-e186-4837-8616-25ffcbf87255 | trentmkelly/LessWrong-43k | LessWrong | Why we need a *theory* of human values
There have been multiple practical suggestions for methods about how we should extract the values of a given human. Here are four common classes of such methods:
* Methods that put high weight on human (bounded) quasi-rationality, or revealed preferences. For example, we can as... |
5e50087c-1b44-4b76-ada1-255fa69dd310 | trentmkelly/LessWrong-43k | LessWrong | Instrumental Convergence? [Draft]
This is a draft written by J. Dmitri Gallow, Senior Research Fellow at the Dianoia Institute of Philosophy at ACU, as part of the Center for AI Safety Philosophy Fellowship. This draft is meant to solicit feedback. Here is a PDF version of the draft.
Abstract
The thesis of instrumen... |
2078eb5a-949d-470e-af88-7eea9d47e968 | trentmkelly/LessWrong-43k | LessWrong | Silicon Valley AI/Machine Learning study group
I am organizing a study group for people in the Silicon Valley area who are taking the Stanford online AI or Machine Learning classes. I posted this to the Tortuga LW meetup google group, but in case there are people here who are interested in such a study group but are n... |
fa4b20c1-8085-438c-9d69-5c557251c051 | trentmkelly/LessWrong-43k | LessWrong | Two kinds of population ethics, and Current-Population Utilitarianism
There are two different kinds of questions that could be considered to fall under the subject of population ethics: “What sorts of altruistic preferences do I have about the well-being of others?”, and “Given all the preferences of each individual, ... |
a00948f0-309d-4a82-879c-68fd1c054bbd | trentmkelly/LessWrong-43k | LessWrong | The Other Earth (Chapter 2)
In this post, Chapter 2 of the sequence, I share the fable of The Other Earth from Olaf Stapledon’s Star Maker. It is abridged and lightly edited. I have also added a brief introduction titled “Learning to Travel” to provide context found elsewhere in Star Maker. It is within the fable of T... |
36d588e0-9f35-49b9-a950-539a5d7f4da7 | trentmkelly/LessWrong-43k | LessWrong | [Hammertime Final Exam] Quantum Walk, Oracles and Sunk Meaning
One and a half years ago, alkjash published his Hammertime sequence. Admittedly a bit late to the party I recently went through the 30 days together with a few other aspiring rationalists, who too may post their "final exams" in the coming weeks.
The task... |
40acb124-0bb4-49ab-b78f-9c5361349f67 | trentmkelly/LessWrong-43k | LessWrong | Distinctions when Discussing Utility Functions
An agent considering the relative costs and benefits of a long list of options
Epistemic Status: Early. The categories mentioned come mostly from experience and reflection, as opposed to existing literature.
On its surface, a utility function is an incredibly simple and ... |
0de1912d-ad48-4026-a159-9006bb1addf6 | trentmkelly/LessWrong-43k | LessWrong | Time isn't real man
> "There is nothing more deceptive than an obvious fact.''
> ― Arthur Conan Doyle
Anyone who' s taken LSD, pulled an all-nighter, been bored, or fell into a YouTube rabbit knows that time is subjective. Brains aren't clocks. They cheat constantly. Compressing
sequences of events, deleting traumati... |
45bf5e27-a961-4924-b10f-f55c81912799 | trentmkelly/LessWrong-43k | LessWrong | Industry and workers
In a previous post I mentioned the Industrial Revolution as a good early example of exploitation of workers for the benefit of industry owners. Naturally, this wasn't the first or only time this happened[1]. The Industrial Revolution, however, is specifically an example of technologies driving out... |
5b726d76-7690-4213-aad4-e01b6e9c6d93 | trentmkelly/LessWrong-43k | LessWrong | Range and Forecasting Accuracy
cross-posted from niplav.site
This text looks at the accuracy of forecasts in relation to the time between forecast and resolution, and asks three questions: First; is the accuracy higher between forecasts; Second; is the accuracy higher between questions; Third; is the accuracy higher ... |
7698effb-9c64-42ae-907b-c5776418ca2a | trentmkelly/LessWrong-43k | LessWrong | A Bi-Modal Brain Model
When I am programming, writing, reading, browsing the internet, watching a movie, or playing a game, my brain is in a different mode of operation compared to when I am sitting in an empty room with nothing to do.
In the empty room, my brain will continuously generate fragments of language and o... |
7610ee13-71ae-44d2-ad98-da6c25d575d1 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Motivated Stopping and Motivated Continuation
Today's post, Motivated Stopping and Motivated Continuation was originally published on 28 October 2007. A summary (taken from the LW wiki):
> When the evidence we've seen points towards a conclusion that we like or dislike, there is a temptation to stop th... |
aee20fea-ac30-438c-9c07-85de885cce3b | trentmkelly/LessWrong-43k | LessWrong | On OpenAI’s Model Spec 2.0
OpenAI made major revisions to their Model Spec.
It seems very important to get this right, so I’m going into the weeds.
This post thus gets farther into the weeds than most people need to go. I recommend most of you read at most the sections of Part 1 that interest you, and skip Part 2.
... |
c092f224-bea3-47bb-8bf7-17529383a11c | trentmkelly/LessWrong-43k | LessWrong | Meetup : Brussels - Neuroatypicality
Discussion article for the meetup : Brussels - Neuroatypicality
WHEN: 14 June 2014 07:40:35PM (+0200)
WHERE: Rue des Alexiens 55 1000 Bruxelles
According to a 2010 unscientific internet survey, over 5% of LessWrong users have Asperger syndrome, which is maybe 25 times more than ... |
1053abd0-284d-4b49-9cea-fd2d492608e9 | StampyAI/alignment-research-dataset/arxiv | Arxiv | A Dual Approach to Scalable Verification of Deep Networks
1 Introduction
---------------
Deep learning has led to tremendous progress in the field of machine learning resulting in remarkable improvements in the solution of a wide-range of complex prediction tasks (Krizhevsky et al., [2012](#bib.bib16); Goodfellow e... |
1b95344a-374c-46df-a712-688e69dc2142 | trentmkelly/LessWrong-43k | LessWrong | On Arrogance
Arrogance is an interesting topic.
Let's imagine we have two people who are having a conversation. One of them is an professor in quantum mechanics and the other person is an enthusiast who has read a few popular science articles online.
The professor always gives his honest opinion, but in an extremely... |
b49fa037-14ef-447a-91c6-4ac774f95e3f | trentmkelly/LessWrong-43k | LessWrong | Prisoner's Dilemma on game show Golden Balls
I found this to be a very interesting method of dealing with a modified Prisoner's Dilemma. In this situation, if both players cooperate they split a cash prize, but if one defects he gets the entire prize. The difference from the normal prisoner's dilemma is that if both d... |
deb0c97c-4a42-4de0-b399-a0563f311e37 | trentmkelly/LessWrong-43k | LessWrong | Pleasure and suffering are not conceptual opposites
Pleasure and suffering are not conceptual opposites. Rather, pleasure and unpleasantness are conceptual opposites. Pleasure and unpleasantness are liking and disliking, or positive affect and negative affect, respectively.[1]
Imagine an experience of suffering that ... |
9b844edc-e6e6-484a-86c1-82bdc8d70740 | trentmkelly/LessWrong-43k | LessWrong | Robert Ettinger, founder of cryonics, now CIs 106th patient
Ben Best: Robert Ettinger deanimated today at around 4pm Eastern Time. He was under hospice care and had an ice bath sitting by his bedside.
Obituaries: Mike Darwin, Washington Post, Telegraph
Wikipedia |
f372b02c-9c00-45ac-ad66-2c5cd269ef54 | trentmkelly/LessWrong-43k | LessWrong | Announcing Foresight Institute's AI Safety Grants Program
Foresight Institute has received funding to support projects in three AI safety areas we think are potentially under-explored:
1. Neurotechnology, brain computer interface, whole brain emulation, and "lo-fi" uploading approaches to produce human-aligned softwa... |
b1ceb469-1c3f-48f2-9ce3-c220a61d6c4c | trentmkelly/LessWrong-43k | LessWrong | Fake Explanations
Once upon a time, there was an instructor who taught physics students. One day the instructor called them into the classroom and showed them a wide, square plate of metal, next to a hot radiator. The students each put their hand on the plate and found the side next to the radiator cool, and the dista... |
adcb6e79-92d8-4616-a2cc-0aef001a3324 | trentmkelly/LessWrong-43k | LessWrong | [Book review] Gödel, Escher, Bach: an in-depth explainer
[Thanks to Jacob Falkovich, Eric Neyman, and my LessWrong reviewer, among others, for their feedback on earlier drafts.]
Initially, Gödel, Escher, Bach comes across as a perplexingly well-regarded conspiracy theory text. But reading on, you come to see the magi... |
720f8187-d61c-4581-8eef-3afa6a1eba9e | trentmkelly/LessWrong-43k | LessWrong | How do the ivermectin meta-reviews come to so different conclusions?
Ivermectin and outcomes from Covid-19 pneumonia: A systematic review and meta-analysis of randomized clinical trial studies comes to the conclusion: "Our study suggests that ivermectin may offer beneficial effects towards Covid-19 outcomes. More rand... |
d04456fe-6cc6-4d87-bb1f-b03a7f79a12e | trentmkelly/LessWrong-43k | LessWrong | Epistea Workshop Series: Epistemics Workshop, May 2020, UK
Hi all!
We are excited to announce the Epistea Epistemics Workshop! This is the first out of a series of 3 workshops happening in Europe, organized and run by the Epistea team.
Tl,dr:
* What: a 3.5 day workshop focusing on improving our epistemics and habi... |
a8f8d225-85f3-4825-8c72-356e8ad08846 | trentmkelly/LessWrong-43k | LessWrong | Why we may elect our new AI overlords
In which I examine some of the latest development in automated fact checking, prediction markets for policies and propose we get rich voting for robot politicians.
http://pirate.london/2016/09/why-we-may-elect-our-new-ai-overlords/ |
794d5c2d-d833-421b-b30e-eb5152567e32 | trentmkelly/LessWrong-43k | LessWrong | Egoism In Disguise
Originally posted at Living Within Reason
Epistemic status: moderately certain, but open to being convinced otherwise
tl;dr: any ethical system that relies on ethical intuitions is just egoism that's given a veneer of objectivity.
UTILITARIANISM RELIES ON MORAL INTUITIONS
Most rationalists are u... |
3a71ee74-fb5c-4289-a4ba-83fb4de83ef3 | trentmkelly/LessWrong-43k | LessWrong | [Link] My talk about the Future
I recently gave a talk at the IARU Summer School on the Ethics of Technology.
In it, I touched on many of the research themes of the FHI: the accuracy of predictions, the limitations and biases of predictors, the huge risks that humanity may face, the huge benefits that we may gain, an... |
d9317457-9fd8-4c7f-b761-e338a5790d54 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | $20K in Bounties for AI Safety Public Materials
TLDR
----
We are announcing a $20k bounty for publicly-understandable explainers of AI safety concepts. We are also releasing the results of the [AI Safety Arguments competition](https://forum.effectivealtruism.org/posts/p3eiBqnijXPv5pCMA/usd20k-in-prizes-ai-safety-argu... |
c44758d3-10f3-4880-8f9b-fed5b707faeb | trentmkelly/LessWrong-43k | LessWrong | Announcing The Roots of Progress Blog-Building Intensive
Today we’re launching The Roots of Progress Blog-Building Intensive, an 8-week program for aspiring progress writers to start or grow a blog.
What: Through this program you will launch (or re-launch) a blog/Substack, get into a regular writing habit, improve yo... |
1c0f8ca9-2baa-4b27-b51d-52eabcad1c3d | trentmkelly/LessWrong-43k | LessWrong | Transformer Circuit Faithfulness Metrics Are Not Robust
When you think you've found a circuit in a language model, how do you know if it does what you think it does? Typically, you ablate / resample the activations of the model in order to isolate the circuit. Then you measure if the model can still perform the task y... |
7465f415-7522-45ac-90a8-68d51162b98e | trentmkelly/LessWrong-43k | LessWrong | Transmute guilt into resolve
A friend of mine came to me and said that he cares about his immediate friends, and he cares about humanity in the abstract, but he has trouble caring for most people. They seemed too shallow, too bitter, too spiteful to be worth an effort.
He'd been a sixth grade teacher, so I asked, "Wh... |
21c2d7bd-46a7-45df-a40f-18df5cb8459b | trentmkelly/LessWrong-43k | LessWrong | Interview with Aubrey de Grey, chief science officer of the SENS Research Foundation
Introduction
In the last post, I explained why I think SENS Research Foundation is probably a great choice for its cost-effectiveness, and I announced that I would interview its chief science officer, Aubrey de Grey, for a further eva... |
dd00377c-f504-4865-88a0-4f28d73dbb9c | StampyAI/alignment-research-dataset/arxiv | Arxiv | Is BERT Really Robust? A Strong Baseline for Natural Language Attack on Text Classification and Entailment
1 Introduction
---------------
In the last decade, machine learning (ML) models have achieved remarkable success in various tasks such as classification, regression and decision making. However, recently they ... |
6264179f-df2b-4499-a159-6ca91d49f089 | trentmkelly/LessWrong-43k | LessWrong | What is Compute? - Transformative AI and Compute [1/4]
Cross-posted here on the EA Forum.
Transformative AI and Compute - A holistic approach - Part 1 out of 4
This is part one of the series Transformative AI and Compute - A holistic approach. You can find the sequence here and the summary here.
This work was condu... |
3c04162c-26d4-4608-ab09-35c2c547fcbd | trentmkelly/LessWrong-43k | LessWrong | What are good topics for literature review prizes?
The last prize, for the best literature review of spaced repetition, was moderately successful, inspiring a pretty good review of the academic literature on spaced repetition. I am interested in experimenting more with prizes, but I would like to get input other peopl... |
fb385ba7-577e-4dd5-95a8-4a4b612360d1 | trentmkelly/LessWrong-43k | LessWrong | PIBBSS Summer Symposium 2023
Tl;dr
We are pleased to invite you to the PIBBSS Summer Symposium where the fellows from the ‘23 fellowship program present their work.
The symposium is taking place online, over several days in the week of September 18th.
* Check out the full program, including brief descriptions fo... |
8fe3b675-837a-4fc2-ba16-c8d5a454abba | trentmkelly/LessWrong-43k | LessWrong | The Fundamental Theorem for measurable factor spaces
I present the fundamental theorem for all finitely factored measurable spaces. The fundamental theorem is that two events are orthogonal if and only if they are independent in all product probability distributions. It tells us that the definition of orthogonality re... |
4f44d95c-1f6e-4add-a984-49648d5d4e28 | trentmkelly/LessWrong-43k | LessWrong | Haskenthetical update: user-defined macros
A while back I wrote about Haskenthetical, a language that I'm implementing for fun. I'm still working on it, at least from time to time.
The thing I'm pleased to have implemented recently is user-defined macros. This makes it potentially a lot more ergonomic.
(You still ca... |
83d2e899-67f6-40d6-a638-ee5df1859f90 | trentmkelly/LessWrong-43k | LessWrong | Comprehensible Input is the only way people learn languages - is it the only way people *learn*?
Everything I've learned up to this point regarding language and language-learning is that the human first has an initial phase of gathering lots of 'chunks' of information. Then, with time, they learn how those chunks fit ... |
18f0cf09-79d8-4525-8882-ff15a62c8178 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] A New Day
Today's post, A New Day was originally published on 31 December 2008. A summary (taken from the LW wiki):
> Try spending a day doing as many new things as possible.
Discuss the post here (rather than in the comments to the original post).
This post is part of the Rerunning the Sequences se... |
dde5db6b-428e-4d62-8f92-6287b677e915 | StampyAI/alignment-research-dataset/blogs | Blogs | Ensuring smarter-than-human intelligence has a positive outcome
I recently gave a talk at Google on the problem of aligning smarter-than-human AI with operators’ goals:
The talk was inspired by “[AI Alignment: Why It’s Hard, and Where to Start](https://intelligence.org/2016/12/28/ai-alignment-why-its-hard-a... |
8e178ce0-d197-41e8-97fe-65cadba0f741 | trentmkelly/LessWrong-43k | LessWrong | Sandberg, A. and Bostrom, N. (2011): Machine Intelligence Survey
> As some readers may recall, we had a conference this January about intelligence, and in particular the future of machine intelligence. We did a quick survey among participants about their estimates of when and how human-level machine intelligence would... |
c040dfad-b4b8-4bc5-b78e-103af148ca77 | trentmkelly/LessWrong-43k | LessWrong | Maps vs Buttons; Nerds vs Normies
It took me the longest time to see through the illusion that was "rational" discussion.
The setting: A friend and I exchange inquiries about each other's beliefs about X.
The result: the friend would give a number of answers that allowed me to piece together their view of X. An asto... |
a8b675c2-758c-47cf-9bb4-c9c72ae22a18 | trentmkelly/LessWrong-43k | LessWrong | Does UDT *really* get counter-factually mugged?
I'm hoping someone can clear this up for me; I'd be much obliged.
On LessWrong, I think someone made a good point regarding UDT's propensity (or lack there-of) to be counter-factually mugged, see: http://lesswrong.com/lw/3l/counterfactual_mugging/dm3r |
8dbc09a2-7140-4648-901b-61ddf658905d | trentmkelly/LessWrong-43k | LessWrong | Neural Basis for Global Workspace Theory
Epistemic Status: Intense amateur neuroscience here. Hoping to leverage Cunningham's Law to reach enlightenment.
Kaj Sotala has a great sequence on Multiagent Models of the Mind, a sequence that's lead to a lot of fun developments in how I think about minds in general. It also... |
b1b13069-91a1-43d3-ab81-d0ca8b92dfc6 | trentmkelly/LessWrong-43k | LessWrong | Fund Transit With Development
When transit gets better the land around it becomes more valuable: many people would like to live next to a subway station. This means that there are a lot of public transit expansions that would make us better off, building space for people to live and work. And yet, at least in the US, ... |
d3bab0e8-21ed-4ce6-a6af-5faea7c15f24 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | LessWrong is now a book, available for pre-order!
For the first time, you can now read the best new ideas on LessWrong in a physical book set:
> *A Map that Reflects the Territory: Essays by the LessWrong Community*
>
>
It is available for pre-order [here](http://www.lesswrong.com/books).
(And for those of you w... |
cedeb17b-fe87-4e0e-a09e-9f5a0411a37f | trentmkelly/LessWrong-43k | LessWrong | AXRP Episode 38.4 - Shakeel Hashim on AI Journalism
YouTube link
AI researchers often complain about the poor coverage of their work in the news media. But why is this happening, and how can it be fixed? In this episode, I speak with Shakeel Hashim about the resource constraints facing AI journalism, the disconnect b... |
84de1ca2-c27a-4ead-a6f4-e7e1f27335bc | trentmkelly/LessWrong-43k | LessWrong | LW 2.0 Site Update: 09/21/17
Hey! A few quick updates on edits done today (well, yesterday - we have late work hours).
Thanks Chris_Leong for making a post to document folks' bugs. In the comments ozymandias requested that the page containing every post published to the site organised by day, be more findable. That's... |
6e718db7-1855-4738-a65c-d62af236d202 | trentmkelly/LessWrong-43k | LessWrong | Why death for oneself, suffering for others?
Doctors in a study preferred treatments with more chance of death over those with more chance of suffering for themselves more often than for their patients. Robin hypothesises*:
> Avoiding death is a primary goal of medicine. Avoiding side effects of treatment is a second... |
62f4754c-92fe-4ce0-88a1-2abeef53a2f9 | trentmkelly/LessWrong-43k | LessWrong | Announcing the AI Safety Nudge Competition to Help Beat Procrastination
Have you achieved your goal? Fill in this form to be added to the prize draw!
Have you been postponing doing alignment work? Is your learning progressing slower than you'd like? We've had these experiences too, so we've created the AI Safety Nudg... |
c024c311-245e-4928-a9fc-c60139bbc39e | trentmkelly/LessWrong-43k | LessWrong | The Control Group Is Out Of Control
I.
Allan Crossman calls parapsychology the control group for science.
That is, in let’s say a drug testing experiment, you give some people the drug and they recover. That doesn’t tell you much until you give some other people a placebo drug you know doesn’t work – but which they ... |
86c96038-49ac-4f62-8609-82676f0eb2ab | trentmkelly/LessWrong-43k | LessWrong | Re: Taste
"I read Friendly And Hostile Analogies For Taste by Scott Alexander and I'm confused," said Elsani.
"Why are you reading a blog post that's over 100 years old?" asked her friend Taze, "Society has made so much progress since then."
"Has it?" asked Elsani, "A random walk, in retrospect, looks like like dire... |
04254e5e-dc90-4cc8-aca8-594d95112637 | trentmkelly/LessWrong-43k | LessWrong | Morality vs related concepts
Cross-posted to the EA Forum.
How can you know I’m talking about morality (aka ethics), rather than something else, when I say that I “should” do something, that humanity “ought” to take certain actions, or that something is “good”? What are the borderlines and distinctions between morali... |
7879ec38-eb8f-4b8c-8447-75689a012c7c | trentmkelly/LessWrong-43k | LessWrong | play and spinning plates
There are four levels of ignorance [1]:
* Known Knowns - Knowledge - Things you know and you know you know them.
* Known Unknowns - Lack of Knowledge - Things you know you don’t know. - I don’t know how to juggle, piffle is probably a word, That feeling that there’s something you’ve forgott... |
33e51124-47c0-40f5-96a6-eec931218e90 | trentmkelly/LessWrong-43k | LessWrong | Passing Up Pay
I recently switched from earning money to donate to doing something directly useful. I think this work is very important (or I wouldn't have switched!), and if I were still earning to give it's the kind of work I'd be excited to fund. I also think this about Julia's work; what should this mean for our d... |
687610bc-af72-404c-a791-083508159db3 | trentmkelly/LessWrong-43k | LessWrong | Information on time-complexity prior?
On a recent episode of The Bayesian Conspiracy, the time-complexity prior was mentioned as a plausible alternative to Solomnoff induction. And reading Abram Demski's explanation of logical induction in Toward a New Technical Explanation of Technical Explanation, a prior over time-... |
7f380fb3-d40a-4623-a344-9173c92a9912 | trentmkelly/LessWrong-43k | LessWrong | Humans Can Be Manually Strategic
Summary: Humans don't always take the time to think and plan very well. That does not mean we can’t think about our goals and how we want to go about achieving them. This meetup is an attempt to prompt strategic thinking.
Tags: Repeatable, Investment, Highly Experimental. (Highly Exp... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.