text stringlengths 300 320k | source stringlengths 52 154 |
|---|---|
# Slide deck: Introduction to AI Safety
This is from a presentation that I've been giving over the past few months to various groups within the Johns Hopkins University Applied Physics Laboratory (APL), where I work. I'd be interested to hear people's comments, critiques, suggestions, etc.
Note that this is a somewha... | https://www.lesswrong.com/posts/8MXwoxNuicKtwmZm2/slide-deck-introduction-to-ai-safety |
# [AN #84] Reviewing AI alignment work in 2018-19
Find all Alignment Newsletter resources [here](http://rohinshah.com/alignment-newsletter/). In particular, you can [sign up](http://eepurl.com/dqMSZj), or look through this [spreadsheet](https://docs.google.com/spreadsheets/d/1PwWbWZ6FPqAgZWOoOcXM8N_tUCuxpEyMbN1NYYC02a... | https://www.lesswrong.com/posts/6Rv9kLGmXrkqRrcK9/an-84-reviewing-ai-alignment-work-in-2018-19 |
# Towards deconfusing values
_NB: Kaj recently said [some similar and related things](https://www.lesswrong.com/posts/2yLn8iTrvHoEgqXcJ/the-two-layer-model-of-human-values-and-problems-with) while I was on hiatus from finishing this post. I recommend reading it for a different take on what I view as a line of thin... | https://www.lesswrong.com/posts/WAqG5BQMzAs34mpc2/towards-deconfusing-values |
# Value uncertainty
_Epistemic status: This is basically a new categorisation scheme for, and analysis of, ideas that other people have proposed previously (both in relation to moral philosophy and in relation to AI alignment). I’m not an expert on the topics I cover here, and I’d appreciate feedback or comments in re... | https://www.lesswrong.com/posts/s6BGofzFbEr4Tmxkj/value-uncertainty |
# If brains are computers, what kind of computers are they? (Dennett transcript)
*In 2013, Daniel Dennett gave this talk at an Oxford conference, and* [*it was uploaded*](https://www.youtube.com/watch?v=OlRHd-r2LOw) *to FHI's YouTube channel. I remember this being one of my favourite talks by Dennett. I rewatched it y... | https://www.lesswrong.com/posts/fuGNHdgYWBkA5Fi22/if-brains-are-computers-what-kind-of-computers-are-they |
# Artificial Intelligence, Values and Alignment
I haven't yet had a chance to read the whole thing, but DeepMind has a [new paper](https://deepmind.com/research/publications/Artificial-Intelligence-Values-and-Alignment) up from Iason Gabriel titled "Artificial Intelligence, Values and Alignment". Here's the abstract:
... | https://www.lesswrong.com/posts/LDMqaq7JmtPiHkjA5/artificial-intelligence-values-and-alignment |
# how has this forum changed your life?
Hello! I am totally new here so please bear that in mind in the event I commit faux pas! I'm a writer who has written a LOT about rationality and when rationality eludes us. These are books like So You've Been Publicly Shamed, Them, The Psychopath Test and The Men Who Stare At ... | https://www.lesswrong.com/posts/Ds4h6yiufcoCaACyX/how-has-this-forum-changed-your-life |
# High-precision claims may be refuted without being replaced with other high-precision claims
There's a common criticism of theory-criticism which goes along the lines of:
> Well, sure, this theory isn't *exactly* right. But it's the best theory we have right now. Do you have a better theory? If not, you can't reall... | https://www.lesswrong.com/posts/YcTAs8TZB8ps9TxDr/high-precision-claims-may-be-refuted-without-being-replaced |
# Book Review: Human Compatible
**I.**
[Clarke’s First Law](https://en.wikipedia.org/wiki/Clarke%27s_three_laws) goes: When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.
Stuart Russell is ... | https://www.lesswrong.com/posts/irbREZtZzPi7WEYex/book-review-human-compatible-1 |
# Existing work on creating terminology & names?
I'm in the process of coming up with terminology for various theories, similar to lots of other work on LessWrong and The EA Forum.
Naming things is a bit of a unilateralist action. While community members don't have to accept a specific naming proposal, they are likel... | https://www.lesswrong.com/posts/iTqz2jdzCWFv8cwYT/existing-work-on-creating-terminology-and-names |
# [ELDR Tactics] Consider switching to (mostly) decaf.
_Related post: [Everyday belief in diminishing returns is subject to diminishing returns](https://www.lesswrong.com/posts/x88ijGK2DnerBpRAM/everyday-belief-in-diminishing-returns-is-resistant-to)_
Growing up as a first-generation American, and generally considere... | https://www.lesswrong.com/posts/uLXB4rLZ9WoNKRuuj/eldr-tactics-consider-switching-to-mostly-decaf |
# [Link] Ignorance, a skilled practice
[https://carcinisation.com/2020/01/27/ignorance-a-skilled-practice/](https://carcinisation.com/2020/01/27/ignorance-a-skilled-practice/)
Excerpt:
The Global Knowledge Game
To illustrate that global knowledge is a game, consider a story about Alexander Luria, who studied il... | https://www.lesswrong.com/posts/oWyQurNGoHfCpAH7Y/link-ignorance-a-skilled-practice |
# Create a Full Alternative Stack
[Last time](https://thezvi.wordpress.com/2020/01/28/potential-ways-to-fight-mazes/) I proposed nine strategies for fighting mazes. All of them were either systemic solutions requiring coordinated and/or political action, or cultural shifts that happen one person at a time.
Now for th... | https://www.lesswrong.com/posts/NerqNs3nfAkcJauWE/create-a-full-alternative-stack |
# REVISED: A drowning child is hard to find
Substantial revisions to clarify the post's core claim, including but not limited to this summary at the end:
> # Summary
>
> * Effective Altruism claims that there is a large funding gap for cheap well-understood developing-world interventions.
> * Even the most aggr... | https://www.lesswrong.com/posts/tWuNefz8LF39FLn66/revised-a-drowning-child-is-hard-to-find |
# Instrumental Occam?
In _[The Real Rules have No Exceptions](https://www.lesswrong.com/posts/duxy4Hby5qMsv42i8/the-real-rules-have-no-exceptions),_ Said says (apparently speaking of _instrumental_ rules, not epistemic ones):
> _Prefer simplicity in your rules. Be vigilant that your rules do not grow too complex; mak... | https://www.lesswrong.com/posts/qqG2PdZ7pEcM6ev3S/instrumental-occam |
# Effective Altruism QALY workshop materials & outline (and Jan 13 '19 meetup notes)
What the workshop is & a brief overview
=======================================
Hi all! This is my first post on LessWrong, but I'll be posting here more often with conversation notes & resources from the Effective Altruism Kansas Ci... | https://www.lesswrong.com/posts/gQuWARsPHQ2Jx5rjL/effective-altruism-qaly-workshop-materials-and-outline-and |
# What Money Cannot Buy
[Paul Graham](http://www.paulgraham.com/gh.html):
> The problem is, if you're not a hacker, you can't tell who the good hackers are. A similar problem explains why American cars are so ugly. I call it the design paradox. You might think that you could make your products beautiful just by hirin... | https://www.lesswrong.com/posts/YABJKJ3v97k9sbxwg/what-money-cannot-buy |
# The case for lifelogging as life extension
Those in the cryonics community want to be frozen upon legal death, in order to preserve the information content in their brain. The hope is that, given good protocol, damage incurred during the freezing process will not destroy enough information about you to prevent peopl... | https://www.lesswrong.com/posts/k8mwvvvpjMGcZLAKH/the-case-for-lifelogging-as-life-extension |
# "Memento Mori", Said The Confessor
## Abstract
The fear of death acts as a sort of master key for introductory rationality concepts. Examining the fear of death ties all the rationality basics together into a coherent framework, including:
* Map/Territory Errors
* Something To Protect
* Keeping Your Identity... | https://www.lesswrong.com/posts/RjdvRjmaWwATsbP4J/memento-mori-said-the-confessor |
# [Link] Beyond the hill: thoughts on ontologies for thinking, essay-completeness and forecasting
_This is a link-post for: [https://www.foretold.io/c/1bea107b-6a7f-4f39-a599-0a2d285ae101/n/5ceba5ae-60fc-4bd3-93aa-eeb333a15464](https://www.foretold.io/c/1bea107b-6a7f-4f39-a599-0a2d285ae101/n/5ceba5ae-60fc-4bd3-93aa-e... | https://www.lesswrong.com/posts/xp6hj7WWGjpgQhiTy/link-beyond-the-hill-thoughts-on-ontologies-for-thinking |
# Money isn't real. When you donate money to a charity, how does it actually help?
Money is a shared illusion. Neither paper, gold, nor electron patterns can stop a mosquito or feed a hungry person directly. Transfer of money is a temporary, low friction way to motivate unidentified strangers to preform an action.
... | https://www.lesswrong.com/posts/g6JjPqv4bWGFvdA3g/money-isn-t-real-when-you-donate-money-to-a-charity-how-does |
# A point of clarification on infohazard terminology
*TL;DR: “Infohazard” means any kind of information that could be harmful in some fashion. Let’s use “cognitohazard” to describe information that could specifically harm the person who knows it.*
Some people in my circle like to talk about the idea of information ha... | https://www.lesswrong.com/posts/Rut5wZ7qyHoj3dj4k/a-point-of-clarification-on-infohazard-terminology |
# UML IX: Kernels and Boosting
(This is the ninth post in a sequence on Machine Learning based on [this book](https://www.amazon.com/Understanding-Machine-Learning-Theory-Algorithms/dp/1107057132). Click [here](https://www.lesswrong.com/s/C8wgFMCmoMNbvDSMP/p/Z29etNPef6yvrPteK) for part I.)
Kernels
=======
To motivat... | https://www.lesswrong.com/posts/QExXmkRmsufYQLsoQ/uml-ix-kernels-and-boosting |
# Map Of Effective Altruism
In the spirit of my old [map of the rationalist diaspora](https://slatestarcodex.com/2014/09/05/mapmaker-mapmaker-make-me-a-map/), here’s a map of the effective altruist movement:
[](http://slatestarcodex.com/blog_images/ea_imag... | https://www.lesswrong.com/posts/4Y2J7NtuweW2B8JvB/map-of-effective-altruism |
# Pessimism About Unknown Unknowns Inspires Conservatism
[This](https://mkcohen-hosted-files.s3-us-west-1.amazonaws.com/Pessimism_alignmentforum.pdf) \[EDIT: [final version](http://proceedings.mlr.press/v125/cohen20a/cohen20a.pdf), [presentation](https://www.learningtheory.org/colt2020/virtual/papers/paper_221.html)\]... | https://www.lesswrong.com/posts/RzAmPDNciirWKdtc7/pessimism-about-unknown-unknowns-inspires-conservatism |
# Protecting Large Projects Against Mazedom
**If we wish to accomplish something that would benefit from or require a larger organization or more levels of management and bureaucracy, what should we do in light of the dangers of mazes?**
There are no easy answers. Real tradeoffs with real sacrifice are the order of t... | https://www.lesswrong.com/posts/4inoCWnKrpHt4gCx9/protecting-large-projects-against-mazedom |
# Category Theory Without The Baggage
*If you are an algebraic abstractologist, this post is probably not for you. Further meta-commentary can be found in the “meta” section, at the bottom of the post.*
So you’ve heard of this thing called “category theory”. Maybe you’ve met some smart people who say that’s it’s real... | https://www.lesswrong.com/posts/B4DuwmtqF3HhNwvua/category-theory-without-the-baggage |
# Looking for books about software engineering as a field
I work in the software industry but am not a software developer. My job is to write about software development, and I've learned a whole bucketload of terms: stuff like 'linked lists', 'CI/CD', 'performance optimization', 'deploy to AWS', 'dockerize', 'microse... | https://www.lesswrong.com/posts/9MyoQQMi7LTZc58KH/looking-for-books-about-software-engineering-as-a-field |
# Long Now, and Culture vs Artifacts
I've recently gotten re-interested in [The Long Now Foundation](http://longnow.org/), and had a conversation about it that [seemed worth writing down](https://www.lesswrong.com/posts/BckFvfxXT323aheKW/writing-down-conversations).
**Longtermism vs Pivotal Generationism**
I love t... | https://www.lesswrong.com/posts/be9KKoBx4ke5ugxXz/long-now-and-culture-vs-artifacts |
# Absent coordination, future technology will cause human extinction
(Crossposted from on [Medium](https://medium.com/@landfish/absent-coordination-future-technology-will-cause-human-extinction-b2a9d3c0cce1#d06d) and my [blog](https://jeffreyladish.com/absent-co... | https://www.lesswrong.com/posts/awyDhCjptukxKdwih/absent-coordination-future-technology-will-cause-human |
# Twenty-three AI alignment research project definitions
I came up with these research project definitions when I read the [iterated
amplification sequence](https://www.lesswrong.com/s/EmDuGeRw749sD3GKd). Last
year I put five of them up for voting (see [Which of these five AI alignment
research projects ideas are no
g... | https://www.lesswrong.com/posts/7mBG2izGtZqZYCWTu/twenty-three-ai-alignment-research-project-definitions |
# Philosophical self-ratification
"Ratification" is [defined](https://www.merriam-webster.com/dictionary/ratification) as "the act or process of ratifying something (such as a treaty or amendment) **:** formal confirmation or sanction". Self-ratification, then, is assigning validity to one's self. (My use of the term ... | https://www.lesswrong.com/posts/g9dNMXKX2fqLgW9a9/philosophical-self-ratification |
# Meta-Preference Utilitarianism
There has been a long standing debate between utilitarians about what should be maximized. Most fall on either the side of ‘average utility’ or ‘total utility’. A select group even chooses ‘minimum utility’ or other lesser known methods.
Previously I have tried to solve this by charti... | https://www.lesswrong.com/posts/2NTSQ5EZ8ambPLxjy/meta-preference-utilitarianism |
# Open & Welcome Thread - February 2020
If it’s worth saying, but not worth its own post, here's a place to put it. (You can also make a [shortform post](http://www.lesslong.com/))
And, if you are new to LessWrong, here's the place to introduce yourself. Personal stories, anecdotes, or just general comments on how yo... | https://www.lesswrong.com/posts/jAixPHwn5bmSLXiMZ/open-and-welcome-thread-february-2020 |
# "But that's your job": why organisations can work
It's no secret that corporations, bureaucracies, and governments don't operate at peak efficiency for ideal goals. Economics has literature on [its own version of the problem](https://en.wikipedia.org/wiki/Principal%E2%80%93agent_problem); on this very site, Zvi has ... | https://www.lesswrong.com/posts/GNmDW2NB8M59bsyC6/but-that-s-your-job-why-organisations-can-work |
# The Adventure: a new Utopia story
_For an introduction to why writing utopias is hard, see_ [_here_](/lw/od7/the_challenge_of_writing_utopia/)_. For a previous utopian attempt, see_ [_here_](/lw/8zs/just_another_day_in_utopia/)_. This story only explores a tiny part of this utopia._
**The Adventure**
==============... | https://www.lesswrong.com/posts/Ybp6Wg6yy9DWRcBiR/the-adventure-a-new-utopia-story |
# [AN #85]: The normative questions we should be asking for AI alignment, and a surprisingly good chatbot
[View this email in your browser](https://mailchi.mp/84b4235cfa34/an-85-the-normative-questions-we-should-be-asking-for-ai-alignment-and-a-surprisingly-good-chatbot?e=[UNIQID])
Find all Alignment Newsletter resou... | https://www.lesswrong.com/posts/Mj259G5n5BxXXrZ7C/an-85-the-normative-questions-we-should-be-asking-for-ai |
# Writeup: Progress on AI Safety via Debate
This is a writeup of the research done by the "Reflection-Humans" team at OpenAI in Q3 and Q4 of 2019. During that period we investigated mechanisms that would allow evaluators to get correct and helpful answers from experts, without the evaluators themselves being expert in... | https://www.lesswrong.com/posts/Br4xDbYu4Frwrb64a/writeup-progress-on-ai-safety-via-debate-1 |
# Synthesizing amplification and debate
# Background
One possible way to train an amplification model is to use an auxiliary reinforcement learning objective to help guide the training of the amplification model. This could be done either by training two separate models, an agent and a question-answerer, or a single ... | https://www.lesswrong.com/posts/dJSD5RK6Qoidb3QY5/synthesizing-amplification-and-debate |
# Potential Research Topic: Vingean Reflection, Value Alignment and Aspiration
_Epistemic Status_: Potential research idea. Time-limited, so not as clear as it could have been.
[Vingean reflection](https://intelligence.org/2015/01/15/new-report-vingean-reflection-reliable-reasoning-self-improving-agents/) is the pro... | https://www.lesswrong.com/posts/uvkockJxm3Ku8rxuk/potential-research-topic-vingean-reflection-value-alignment |
# Some quick notes on hand hygiene
I always wash my hands but typically not very well; I never worked in healthcare and never got trained. Given the [current situation](https://en.wikipedia.org/wiki/2019%E2%80%9320_Wuhan_coronavirus_outbreak) I decided to actually research how to do it properly.
Different sources dif... | https://www.lesswrong.com/posts/ztDYsD4v7AaAbWEDM/some-quick-notes-on-hand-hygiene |
# Plausibly, almost every powerful algorithm would be manipulative
I had an interesting debate recently, about whether we could make smart AIs safe just by focusing on their structure and their task. Specifically, we were pondering something like:
* "Would an algorithm be safe if it was a neural net-style image class... | https://www.lesswrong.com/posts/Ez4zZQKWgC6fE3h9G/plausibly-almost-every-powerful-algorithm-would-be |
# Mazes Sequence Roundup: Final Thoughts and Paths Forward
There are still two elephants in the room that I must address before concluding. Then I will discuss paths forward.
#### Moloch’s Army
The first elephant is Moloch’s Army. I _still_ can’t find a way into this without sounding crazy. The result of this is tha... | https://www.lesswrong.com/posts/xc9DbPPtpqQjLx4Wt/mazes-sequence-roundup-final-thoughts-and-paths-forward |
# Bayes-Up: An App for Sharing Bayesian-MCQ
Inspired by [Lê Nguyên Hoang](https://www.lesswrong.com/users/le-nguyen-hoang-1)'s post on [Bayesian Examination](https://www.lesswrong.com/posts/8wzKawHmh4d3h2otw/bayesian-examination), I have been developing (as a hobby) a new app called Bayes-Up (available at: [bayes-up.w... | https://www.lesswrong.com/posts/7KRWCRBmvLhTMZB5Y/bayes-up-an-app-for-sharing-bayesian-mcq |
# More writeups!
A kind of post I would like to read more of is "I did X, and here's how it went." You tend to see this most with research, but I've enjoyed reading them on all sorts of things. This is one of the [main](https://www.jefftk.com/p/preparing-for-our-cd) [kinds](https://www.jefftk.com/p/boston-solstice-201... | https://www.lesswrong.com/posts/LFrWPkMcennzZwrsn/more-writeups |
# On the falsifiability of hypercomputation
*[ED NOTE: see Vanessa Kosoy's comment [here](https://www.alignmentforum.org/posts/yjC5LmjSRD2hR9Pfa/on-the-falsifiability-of-hypercomputation?commentId=C9sw4eBRxG3aetAM5); this post assumes a setting in which the oracle may be assumed to return a standard natural.]*
It is ... | https://www.lesswrong.com/posts/yjC5LmjSRD2hR9Pfa/on-the-falsifiability-of-hypercomputation |
# A Cautionary Note on Unlocking the Emotional Brain
_\[Cross-posted from [Grand Unified Crazy](https://grandunifiedcrazy.com/2020/02/08/cautionary-note-unlocking-emotional-brain/). Relevant to [Kaj's summary of the book](https://www.lesswrong.com/posts/i9xyZBS3qzA8nFXNQ/book-summary-unlocking-the-emotional-brain).\]_... | https://www.lesswrong.com/posts/rkn2LxGEkbeNkCfGH/a-cautionary-note-on-unlocking-the-emotional-brain |
# What can the principal-agent literature tell us about AI risk?
_This work was done collaboratively with Tom Davidson._
_Thanks to Paul Christiano, Ben Garfinkel, Daniel Garrett, Robin Hanson, Philip Trammell and Takuro Yamashita for helpful comments and discussion. Errors our own._
# Introduction
The AI alignmen... | https://www.lesswrong.com/posts/Z5ZBPEgufmDsm7LAv/what-can-the-principal-agent-literature-tell-us-about-ai |
# Relationship Outcomes Are Not Particularly Sensitive to Small Variations in Verbal Ability
After a friendship-ending fight, you feel an impulse to push through the pain to do an exhaustive postmortem of everything you did wrong in that last, fatal argument—you could have phrased that more eloquently, could have anti... | https://www.lesswrong.com/posts/Ciw8DLJiYnBWa4B7f/relationship-outcomes-are-not-particularly-sensitive-to |
# How to Frame Negative Feedback as Forward-Facing Guidance
Your employee Fred always talks too much during your weekly staff meetings. It's been an ongoing issue. Everyone on your team is annoyed, and so are you. At this point, you have no choice but to give Fred some... _negative feedback_.
You sit down at your com... | https://www.lesswrong.com/posts/mgMA2XK9uSexeKexd/how-to-frame-negative-feedback-as-forward-facing-guidance |
# Who... (or what) designed this site and where did they come from?
Hi there,
I'm brand new here and I have a question for this fine community.
Who designed this site?
The reason I ask is because the design of this site is superb. Every element—everywhere I look—is visually gorgeous to look at. Not only that, but a... | https://www.lesswrong.com/posts/4iti4jZ9G4C7PHpYn/who-or-what-designed-this-site-and-where-did-they-come-from |
# Does there exist an AGI-level parameter setting for modern DRL architectures?
Suppose the architecture includes memory (in the form of a recurrent state) and will act as the policy network for an observation-based RL agent. Evaluating the agent from a reasonable initial state, would you guess that there _exists_ a m... | https://www.lesswrong.com/posts/5pXMThcnZ5PqDeSSz/does-there-exist-an-agi-level-parameter-setting-for-modern |
# State Space of X-Risk Trajectories
*Justin Shovelain developed the core ideas in this article and assisted in writing, David Kristoffersson was the lead writer and editor.*
[*Cross-posted on the EA Forum*](https://forum.effectivealtruism.org/posts/TCxik4KvTgGzMowP9/state-space-of-x-risk-trajectories)
# Abstract
Cu... | https://www.lesswrong.com/posts/P5bczkJAAST9KrkRW/state-space-of-x-risk-trajectories |
# Did AI pioneers not worry much about AI risks?
It's seems noteworthy just how little the AI pioneers from the 40s-80s seemed to care about AI risk. There is no obvious reason why a book like "Superintelligence" wasn't written in the 1950s, but for some reason that didn't happen... any thoughts on why this was the ca... | https://www.lesswrong.com/posts/TDzvtnQ8cpfdp7y4t/did-ai-pioneers-not-worry-much-about-ai-risks |
# A Simple Introduction to Neural Networks
This post will discuss _what_ neural networks are (chapter I), _why_ they work (chapter II), and _how_ they are trained (chapter III). It serves as the tenth entry in a sequence on Machine Learning, but it's written to be accessible without having read any previous part. Most... | https://www.lesswrong.com/posts/Madwb2t79LGrLqWLH/a-simple-introduction-to-neural-networks |
# Attainable Utility Landscape: How The World Is Changed
![](https://i.imgur.com/HPIugVR.p... | https://www.lesswrong.com/posts/fj8eyc7QzqCaB8Wgm/attainable-utility-landscape-how-the-world-is-changed |
# Gricean communication and meta-preferences
There's a bit of an analogy between communication and value learning that I've been thinking about recently.
Communication:
--------------
By "Gricean communication," I mean [Paul Grice's model of communication as actions based on recursive modeling](https://plato.stanfor... | https://www.lesswrong.com/posts/8NpwfjFuEPMjTdriJ/gricean-communication-and-meta-preferences |
# Simulation of technological progress (work in progress)
I've made a model/simulation of technological progress, that you can download and run on your laptop.
My goal is to learn something about intelligence explosions, takeoff speeds, discontinuities, human-level milestones, AGI vs. tools, bottlenecks, or somet... | https://www.lesswrong.com/posts/TiJ76n2J8i6YKryaD/simulation-of-technological-progress-work-in-progress |
# The Relational Stance
There's a concept that's pretty core to my understanding of relationships, which I'm sure must have been written up *somewhere* but I don't know of a good explanation. Here is a first pass.
The concept is, well, 'relating' – the meaning that you ascribe to a relationship, and the stance from w... | https://www.lesswrong.com/posts/5EPzaQoKJinuxaq5L/the-relational-stance-1 |
# Will COVID-19 survivors suffer lasting disability at a high rate?
The case fatality rate of 2019-nCoV (aka Coronavirus, COVID-19) is still uncertain, with estimates floating around ranging from [0.16%-5.7%](https://www.worldometers.info/coronavirus/coronavirus-death-rate/), higher among the elderly and people with p... | https://www.lesswrong.com/posts/h4GFHbhxE2pfiadhi/will-covid-19-survivors-suffer-lasting-disability-at-a-high |
# Demons in Imperfect Search
One day, a gradient descent algorithm ball was happily rolling down a high-dimensional surface hill. All it wanted was to roll as far down as possible. Unbeknownst to the ball, just off to the side was a steep drop-off - but there was a small bump between the ball and the drop-off. No matt... | https://www.lesswrong.com/posts/KnPN7ett8RszE79PH/demons-in-imperfect-search |
# Suspiciously balanced evidence
What probability do you assign to the following propositions?
* "Human activity has caused substantial increases in global mean surface temperature over the last 50 years and barring major policy changes will continue to do so over at least the next 50 years -- say, at least half a ... | https://www.lesswrong.com/posts/AZPg5Wac2mMmKKK3K/suspiciously-balanced-evidence |
# [AN #86]: Improving debate and factored cognition through human experiments
Find all Alignment Newsletter resources [here](http://rohinshah.com/alignment-newsletter/). In particular, you can [sign up](http://eepurl.com/dqMSZj), or look through this [spreadsheet](https://docs.google.com/spreadsheets/d/1PwWbWZ6FPqAgZW... | https://www.lesswrong.com/posts/cZqPGDxbJcbShGwDn/an-86-improving-debate-and-factored-cognition-through-human |
# Building and using the subagent
*Overall summary post [here](https://www.lesswrong.com/s/iRwYCpcAXuFD24tHh/p/PmqQKBmt2phMT7YLG).*
When designing a low-impact agent $A$, one typically wants them to maximise $R_0 - \rho$, where $R_0$ is some positive reward function we want $A$ to increase (ie the actual goal) and ... | https://www.lesswrong.com/posts/fqJmZBG5xmvvd8WRi/building-and-using-the-subagent |
# Confirmation Bias As Misfire Of Normal Bayesian Reasoning
From the subreddit: [Humans Are Hardwired To Dismiss Facts That Don’t Fit Their Worldview](https://www.fastcompany.com/90458795/humans-are-hardwired-to-dismiss-facts-that-dont-fit-their-worldview). Once you get through the preliminary Trump supporter and anti... | https://www.lesswrong.com/posts/mbCccXJuuRBZdXdpH/confirmation-bias-as-misfire-of-normal-bayesian-reasoning |
# In theory: does building the subagent have an "impact"?
**EDIT:** *Extensive modifications to this post, sorry, too numerous to list. If you read the original post, it's worth rereading.*
*Overall summary post [here](https://www.lesswrong.com/s/iRwYCpcAXuFD24tHh/p/PmqQKBmt2phMT7YLG).*
This post looks more generall... | https://www.lesswrong.com/posts/jrrZids4LPiLuLzpu/in-theory-does-building-the-subagent-have-an-impact |
# How do you use face masks?
I did buy face masks a while ago, because other rationalists recommended wearing them.
The list time I had a flu and went to the doctors office I wore a face mask, the receptionist then told me to take it off while sitting in the waiting room.
While I do think I have a better idea then t... | https://www.lesswrong.com/posts/PLEia6QRgp2m5Hrq3/how-do-you-use-face-masks |
# Distinguishing definitions of takeoff
I find discussions about AI takeoff to be very confusing. Often, people will argue for "slow takeoff" or "fast takeoff" and then when I ask them to operationalize what those terms mean, they end up saying something quite different than what _I_ thought those terms meant.
To hel... | https://www.lesswrong.com/posts/YgNYA6pj2hPSDQiTE/distinguishing-definitions-of-takeoff |
# Training Regime Day 0: Introduction
# Introduction
The sad truth of life is that there is a difference between knowing how to do something and actually being able to do it. Knowledge about how various instrumental rationality techniques work is not a substitute for actually being able to use them in practice. As th... | https://www.lesswrong.com/posts/r9JPZQ622KThzgmf5/training-regime-day-0-introduction |
# Please Help Metaculus Forecast COVID-19
Metaculus has launched a new Prize. In contrast to [its earlier contest](https://www.lesswrong.com/posts/hzoYPCZfbWczbCjRB/the-bentham-prize-at-metaculus), which was focused on information to improve forecasts of relevance to Animal Welfare, [the Li Wenliang Prize](https://www... | https://www.lesswrong.com/posts/LuL7LLqcdmM7TTYvW/please-help-metaculus-forecast-covid-19 |
# Perceptrons Explained
This is an introduction to the perceptron learning algorithm, a proof of its convergence, and some visualizations. Ideally, in the same vein as Nicky Case's Explorable Explanations, but I don't think it's at that level yet. Only LW-tangential, mostly just an educational post into basic ML. | https://www.lesswrong.com/posts/DFG2SzkMCeTFXFnFr/perceptrons-explained |
# The Reasonable Effectiveness of Mathematics or: AI vs sandwiches
**TLDR:** I try to find the root causes of why math is useful.
**Epistemic status:** Marginally confident in veracity, not at all confident in novelty.
# Background
I recently had a [discussion](https://www.lesswrong.com/posts/ZeE7EKHTFMBs8eMxn/clar... | https://www.lesswrong.com/posts/qpbYwTqKQG8G7mdFK/the-reasonable-effectiveness-of-mathematics-or-ai-vs |
# The Catastrophic Convergence Conjecture
_TLDR: Recommend some posts for a "practice of rationality" sequence I want to curate! Proposing posts that should exist but don't is also cool._
I'v... | https://www.lesswrong.com/posts/wccxMtZdEvHzLRNTZ/a-practice-of-rationality-sequence |
# Bayesian Evolving-to-Extinction
*The present discussion owes a lot to Scott Garrabrant and Evan Hubinger.*
In [Defining Myopia](https://www.alignmentforum.org/s/HeYtBkNbEe7wpjc6X/p/qpZTWb2wvgSt5WQ4H), I formalized *temporal* or *cross-instance* myopia / non-myopia, but I claimed that there should also be some kin... | https://www.lesswrong.com/posts/u9Azdu6Z7zFAhd4rK/bayesian-evolving-to-extinction |
# Reference Post: Trivial Decision Theory Problem
A trivial decision problem is one where there is only a single option that the agent can take. In that case, the most natural answer to the answer to the question, "What action should we take?" would be "The only action that we can take!". We will call this the _Trivia... | https://www.lesswrong.com/posts/XAeWHqQTWjJmzB4k6/reference-post-trivial-decision-theory-problem |
# Exercises in Comprehensive Information Gathering
Looking back, several of the most durably-valuable exercises I’ve done over the years have a general theme of comprehensive information gathering.
The most recent example involves capital investments. Economists talk about “capital goods” as physical stuff - machines... | https://www.lesswrong.com/posts/9LXxgXySTFsnookkw/exercises-in-comprehensive-information-gathering |
# Why Science is slowing down, Universities and Maslow's hierarchy of needs
I don’t have a very high prior in regards to the correctness of Maslow's hierarchy of needs, but as far as general theories for understanding human needs go, I think it’s a pretty good.
![](https://i.pinimg.com/originals/28/99/25/289925d5... | https://www.lesswrong.com/posts/mRSzaYZ4WcotuA2HA/why-science-is-slowing-down-universities-and-maslow-s |
# Training Regime Day 1: What is applied rationality?
# Introduction
I claim that this series of posts will be a decent training regime for applied rationality. Accordingly, I think that one of the first steps is to tell you what I even think applied rationality is. However, since "applied rationality" is a little har... | https://www.lesswrong.com/posts/xhQxJ7WSZZRkzNZJ8/training-regime-day-1-what-is-applied-rationality |
# On characterizing heavy-tailedness
_\[CONTEXT: For a while I have been meaning to engage with a literature review on heavy tailed distributions. Instead of just indefinitely postponing the project I resolved to write some preliminary thoughts on the topic, so I can get started on understanding the concept better wit... | https://www.lesswrong.com/posts/MkuXMey549KMHJob2/on-characterizing-heavy-tailedness |
# Taking the Outgroup Seriously
_(Author's note/content warning: this post contains politically controversial examples, which are unfortunately perhaps necessary given the subject. I have tried to be relatively even-handed in this matter, my apologies if I erred in doing so.)_
Occasionally, one reads a purported "tak... | https://www.lesswrong.com/posts/DLqjNgEasCr7Wofvr/taking-the-outgroup-seriously |
# Training Regime Day 2: Searching for bugs
# Introduction
In CFAR terminology, a bug is something that systematically goes wrong in your life. It can be anything as small as "it takes me one minute to get out of bed in the morning when it should take me five seconds" to as large as "I hate my job" or even larger. Kn... | https://www.lesswrong.com/posts/babweATFMT9jF6aX7/training-regime-day-2-searching-for-bugs |
# [Link and commentary] The Offense-Defense Balance of Scientific Knowledge: Does Publishing AI Research Reduce Misuse?
This is (partly) a linkpost for [a paper](https://arxiv.org/abs/2001.00463) published earlier this year by Toby Shevlane and Allan Dafoe, both researchers affiliated with the Centre for the Governanc... | https://www.lesswrong.com/posts/H8fTYkNkpYio7XG8L/link-and-commentary-the-offense-defense-balance-of |
# UML XI: Nearest Neighbor Schemes
(This is the eleventh post in a sequence on Machine Learning based on [this book](https://www.amazon.com/Understanding-Machine-Learning-Theory-Algorithms/dp/1107057132). Click [here](https://www.lesswrong.com/s/C8wgFMCmoMNbvDSMP/p/Z29etNPef6yvrPteK) for part I.)
[Last time, I tried ... | https://www.lesswrong.com/posts/eeCReGWj7Qoz8XNvJ/uml-xi-nearest-neighbor-schemes |
# On the falsifiability of hypercomputation, part 2: finite input streams
In [part 1](https://unstableontology.com/2020/02/07/on-the-falsifiability-of-hypercomputation/), I discussed the falsifiability of hypercomputation in a *typed* setting where putative oracles may be assumed to return natural numbers. In this set... | https://www.lesswrong.com/posts/PtaN3oMFPfAAuBNtw/on-the-falsifiability-of-hypercomputation-part-2-finite |
# Attainable Utility Preservation: Concepts
_Posts like this have been written before, but I think it's worth making the point periodically._
[Lurker ratios](https://en.wikipedia.org/wiki/1%25_rule_(Internet_culture)) have likely [increased](https://www.socialmediatoday.com/content/lurkers-are-9... | https://www.lesswrong.com/posts/5XNAPzPwnetop3pLu/how-to-lurk-less-and-benefit-others-while-benefiting |
# Stepwise inaction and non-indexical impact measures
*Overall summary post [here](https://www.lesswrong.com/s/iRwYCpcAXuFD24tHh/p/PmqQKBmt2phMT7YLG).*
In [a previous post](https://www.lesswrong.com/posts/fqJmZBG5xmvvd8WRi/impact-measures-and-subagents), I asked which impact measures were vulnerable to subagents. Vi... | https://www.lesswrong.com/posts/JB6edzY5cccrdbQxP/stepwise-inaction-and-non-indexical-impact-measures |
# Appendix: mathematics of indexical impact measures
*Overall summary post [here](https://www.lesswrong.com/s/iRwYCpcAXuFD24tHh/p/PmqQKBmt2phMT7YLG).*
This post looks at the fascinating situation of indexical impact measures (under the inaction baseline), in the presence of subagent. What happens here in non-trivial... | https://www.lesswrong.com/posts/M9aoMixFLf8JFLRaP/appendix-mathematics-of-indexical-impact-measures |
# Subagents and impact measures: summary tables
These tables will summarise the results of this whole sequence, checking whether subagents can neutralise the impact penalty.
First of all, given a subagent, here are the results for various impact penalties and baselines, and various "value difference summary functions... | https://www.lesswrong.com/posts/PmqQKBmt2phMT7YLG/subagents-and-impact-measures-summary-tables |
# Jan Bloch's Impossible War
**Epistemic Status:** Endorsed
**Content Warning:** Neuropsychological Infohazard, Evocation Infohazard, World War I
**Recommended Prior Reading:** [Blueprint for Armageddon Part I](https://www.dancarlin.com/product/hardcore-history-50-blueprint-for-armageddon-i/)
**Part of the Serie... | https://www.lesswrong.com/posts/hmai5Lru5kWXpH7Ju/jan-bloch-s-impossible-war |
# Training Regime Day 3: Tips and Tricks
# Introduction
When you're trying to learn martial arts, usually you go to a physical location and get instructed by a real live person in real time. However, there are also books on martial arts. It is unknown to me whether or not anyone can actually learn anything useful abo... | https://www.lesswrong.com/posts/NeGC3TFXMX8sxLNdY/training-regime-day-3-tips-and-tricks |
# [Productivity] How not to use "Important // Not Urgent"
_Epistemic status: Personal experience._
This is an **Eisenhower matrix**, named after President Dwight Eisenhower.
Eisenhower matrices have a very useful, _very specific_ role in m... | https://www.lesswrong.com/posts/oacA4nTz42H4F8xuX/productivity-how-not-to-use-important-not-urgent |
# Set Ups and Summaries
Part of the [research process](https://acesounderglass.com/2019/12/11/hows-that-epistemic-spot-check-project-coming/) I’m developing involves reading and thinking through the first and last chapter of a book, and first and last few paragraphs of a chapter, to get an expectation of what’s to com... | https://www.lesswrong.com/posts/N25mBedGXPrrNqD4E/set-ups-and-summaries |
# Wireheading and discontinuity
**Outline**: After a short discussion on the relationship between wireheading and reward hacking, I show why checking the continuity of a sensor function could be useful to detect wireheading in the context of continuous RL. Then, I give an example that adopts the presented formalism. I... | https://www.lesswrong.com/posts/KLNDgqQLfpFXbhQak/wireheading-and-discontinuity |
# Counterfactuals versus the laws of physics
*Overall summary post [here](https://www.lesswrong.com/s/iRwYCpcAXuFD24tHh/p/PmqQKBmt2phMT7YLG).*
I've [claimed](https://www.lesswrong.com/posts/JB6edzY5cccrdbQxP/stepwise-inaction-and-non-indexical-impact-measures#xmJBmgCpdYkgrQTS7) that, given a subagent:
>[...] restric... | https://www.lesswrong.com/posts/bqyCd38tACvKgqmXG/counterfactuals-versus-the-laws-of-physics |
# (In)action rollouts
*Overall summary post [here](https://www.lesswrong.com/s/iRwYCpcAXuFD24tHh/p/PmqQKBmt2phMT7YLG).*
I've previously looked at [subagents in the context of stepwise inaction baselines](https://arxiv.org/pdf/1806.01186.pdf). But there have been improvements to the basic stepwise inaction design, to ... | https://www.lesswrong.com/posts/z9MfmF8gA7SBxGSmb/in-action-rollouts |
# Training Regime Day 4: Murphyjitsu
# Introduction
Recall that applied rationality is being able to [properly decide which advisors to use during decision making](https://www.lesswrong.com/posts/xhQxJ7WSZZRkzNZJ8/training-regime-day-1-what-is-applied-rationality).
Imagine that I am throwing a ball up into the air ... | https://www.lesswrong.com/posts/zZkMm6ySRJKhHJHfq/training-regime-day-4-murphyjitsu |
# Big Yellow Tractor (Filk)
_To the tune of "Big Yellow Taxi", with apologies to Joni Mitchell_
we'll pave paradise and put up a parking lot
if that's what it takes to make all the wild suffering stop
the pain it only grows and grows
until we boil it up in our pot
we'll pave paradise and put up a parking lot ... | https://www.lesswrong.com/posts/3zzzFy5KGA9Qcsn4k/big-yellow-tractor-filk |
# Blog Post Day (Unofficial)
_TL;DR: You are invited to join us [online](https://discord.gg/j5TzpVF) on Saturday the 29th, to write that blog post you've been thinking about writing but never got around to. Comment if you accept this invitation, so I can gauge interest._
The Problem:
------------
Like me, you are to... | https://www.lesswrong.com/posts/q972DHxCwZKFbEPMn/blog-post-day-unofficial |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.