text stringlengths 300 320k | source stringlengths 52 154 |
|---|---|
# Which textbook would you recommend to learn decision theory?
Eliezer talks a lot about decision theory in his sequences, e.g. the Aumann agreement theorem or the von Neumann-Morgenstern utility theorem. From what I've seen so far, decision theory looks extremely interesting.
Which textbook in decision theory would ... | https://www.lesswrong.com/posts/EpecJQg7oMcxpgS7L/which-textbook-would-you-recommend-to-learn-decision-theory |
# Alignment Newsletter #43
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_tUCuxpEyMbN1NYYC02aM/edit?usp=sharing) of ... | https://www.lesswrong.com/posts/Dm2mXk94PATehdr9J/alignment-newsletter-43 |
# Book Trilogy Review: Remembrance of Earth’s Past (The Three Body Problem)
Epistemic Status: Stuff that keeps not going away so I should write it up I suppose. Brief spoiler-free review, then main talk will be spoilerific.
Spoiler-Free Review
===================
I read the trilogy a few months ago, on general stron... | https://www.lesswrong.com/posts/XgGwQ9vhJQ2nat76o/book-trilogy-review-remembrance-of-earth-s-past-the-three |
# Deconfusing Logical Counterfactuals
This post aims to provide a strong philosophical foundation for [logical counterfactuals](https://www.lesswrong.com/posts/6t5vRHrwPDi5C9hYr/motivating-a-semantics-of-logical-counterfactuals#comments), while sketching out an informal scheme that will hopefully be formalised further... | https://www.lesswrong.com/posts/BRuWm4GxcTNPn4XDX/deconfusing-logical-counterfactuals |
# Masculine Virtues
Cross-posted from [Putanumonit](https://putanumonit.com/2019/01/26/masculine-virtues/) (where there's already [a good discussion](https://putanumonit.com/2019/01/26/masculine-virtues/#comments) going).
* * *
Boys Will Be Boys
-----------------
Have you seen the [Gillette ad](https://www.yout... | https://www.lesswrong.com/posts/oA5dxBFWJkRxTdXrK/masculine-virtues |
# Wireheading is in the eye of the beholder
tl;dr: there is no natural category called "wireheading", only wireheading relative to some desired ideal goal.
Suppose that we have a built an AI, and have invited a human H to help test it. The human H is supposed to press a button B if the AI seems to be behaving well. T... | https://www.lesswrong.com/posts/BvctuKocyWR4YYea3/wireheading-is-in-the-eye-of-the-beholder |
# Applied Rationality podcast - feedback?
My goal is to help non-rationalists become rationalists, and provide exercises for existing rationalists to hone their skills, in the form of audio/video episodes.
'Rationally Speaking' and 'The Bayesian Conspiracy' are lovely, but they don't serve the function I have in mind... | https://www.lesswrong.com/posts/izjgvZzo9DCCzb3fH/applied-rationality-podcast-feedback |
# The role of epistemic vs. aleatory uncertainty in quantifying AI-Xrisk
(Optional) Background: what are epistemic/aleatory uncertainty?
===============================================================
Epistemic uncertainty is uncertainty about **which model of a phenomenon is correct.** It can be reduced by learning ... | https://www.lesswrong.com/posts/SwPQmJD2Davq3hQgn/the-role-of-epistemic-vs-aleatory-uncertainty-in-quantifying |
# Reliability amplification
In a recent post I talked about [capability amplification](https://medium.com/ai-control/policy-amplification-6a70cbee4f34#.wmeq2iqwv), a putative procedure that turns a large number of fast weak agents into a slower, stronger agent.
If we do this in a naive way, it will decrease reliabil... | https://www.lesswrong.com/posts/6fMvGoyy3kgnonRNM/reliability-amplification |
# Why is this utilitarian calculus wrong? Or is it?
Suppose that I value a widget at $30. Suppose that the widget costs the widget-manufacturer $20 to produce, but, due to monopoly power on their part, they can charge $100 per widget.
The economic calculus for this problem is as follows. $30 (widget valuation) - $100... | https://www.lesswrong.com/posts/dGGC4cqcWtR589bdC/why-is-this-utilitarian-calculus-wrong-or-is-it |
# Drexler on AI Risk
Eric Drexler has published a book-length [paper on AI
risk](https://www.fhi.ox.ac.uk/reframing/), describing an approach that
he calls Comprehensive AI Services (CAIS).
His primary goal seems to be reframing AI risk discussions to use a
rather different paradigm than the one that Nick Bostrom and... | https://www.lesswrong.com/posts/bXYtDfMTNbjCXFQPh/drexler-on-ai-risk |
# Who wants to be a Millionaire?
tldr; Using Toy models, the Kelly criterion, prospect theory, Bayes and amateur psychology in an unnecessarily detailed analysis of Who wants to be a Millionaire. I strongly suspect there will be errors but hopefully not too many.
Motivating example
==================
I was watching ... | https://www.lesswrong.com/posts/Svaj4nSJCzoRZkvjT/who-wants-to-be-a-millionaire |
# Urgent & important: How (not) to do your to-do list
The Eisenhower Box is a well-known, simple decision matrix for dealing with tasks such as a to-do list, based on whether they’re urgent or important.
I reckon it has multiple flaws. But by fixing each flaw in turn, we end up with a better matrix, which I'll call *... | https://www.lesswrong.com/posts/dJQ7BFz9ZPqstP3an/urgent-and-important-how-not-to-do-your-to-do-list |
# What kind of information would serve as the best evidence for resolving the debate of whether a centrist or leftist Democratic nominee is likelier to take the White House in 2020?
The public debate for who should be the Democratic Party presidential candidate for the 2020 electoral race hinges on whether the Dems ar... | https://www.lesswrong.com/posts/cqGw6yfbM29zazitm/what-kind-of-information-would-serve-as-the-best-evidence |
# How does Gradient Descent Interact with Goodhart?
I am confused about how gradient descent (and other forms of local search) interact with Goodhart's law. I often use a simple proxy of "sample points until I get one with a large U value" or "sample n points, and take the one with the largest U value" when I think a... | https://www.lesswrong.com/posts/pcomQ4Fwi7FnfBZBR/how-does-gradient-descent-interact-with-goodhart |
# How to notice being mind-hacked
_Epistemic status: quite sure, but likely nothing new, I have not done the requisite literature search._
Human mind is not designed with security in mind. It has some defenses against basic adversaries that would have prevented our survival as a species, but not much more than that. ... | https://www.lesswrong.com/posts/akZ9qGrBv4vQmXrtq/how-to-notice-being-mind-hacked |
# Rationality: What's the point?
_This post is part of my_ _[Hazardous Guide To Rationality.](https://www.lesswrong.com/s/pKxyvLyhEFxHgogdD)_ _I don't expect this to be new or exciting to frequent LW people, and I would super appreciate comments and feedback in light of intents for the sequence, as outlined in the abo... | https://www.lesswrong.com/posts/DTLtroPqin7XqrHJ6/rationality-what-s-the-point |
# What are some of bizarre theories based on anthropic reasoning?
Doomsday argument and simulation argument are quite bizarre for most people. But these are not the only strange theories one can come up with when employing anthropics.
Some examples:
**It is likely to have an unusual high IQ.**
Perhaps brain works i... | https://www.lesswrong.com/posts/8uTfyAZ6sMamSn6Z3/what-are-some-of-bizarre-theories-based-on-anthropic |
# AI Safety Prerequisites Course: Revamp and New Lessons
Previous post: [Fundamentals of Formalisation Level 7: Equivalence Relations and Orderings](https://www.lesswrong.com/posts/BkL78sXpNXuHcR64Q/fundamentals-of-formalisation-level-7-equivalence-relations). First post: [Fundamentals of Formalisation level 1: Basic ... | https://www.lesswrong.com/posts/QnZjZvh4SxTP3bHS8/ai-safety-prerequisites-course-revamp-and-new-lessons |
# Conclusion to the sequence on value learning
_This post summarizes the sequence on value learning. While it doesn’t introduce any new ideas, it does shed light on which parts I would emphasize most, and the takeaways I hope that readers get. I make several strong claims here; interpret these as my [impressions, not ... | https://www.lesswrong.com/posts/TE5nJ882s5dCMkBB8/conclusion-to-the-sequence-on-value-learning |
# Constructing Goodhart
A [recent question from Scott Garrabrant](https://www.lesswrong.com/posts/pcomQ4Fwi7FnfBZBR/how-does-gradient-descent-interact-with-goodhart) brought up the issue of formalizing [Goodhart’s Law](https://en.wikipedia.org/wiki/Goodhart%27s_law). The problem is to come up with some model system wh... | https://www.lesswrong.com/posts/NwaNPHYhXDc9LkK8J/constructing-goodhart |
# (notes on) Policy Desiderata for Superintelligent AI: A Vector Field Approach
_**Meta:** I thought I'd spend a little time reading the policy papers that Nick Bostrom has written. I made notes as I went along, so I spent a little while cleaning them up into a summary post. These are my notes on Bostrom, Dafoe and Fl... | https://www.lesswrong.com/posts/hyfedqhgCQriBB9wT/notes-on-policy-desiderata-for-superintelligent-ai-a-vector |
# When to use quantilization
In 2015, Jessica introduced [quantilization](https://intelligence.org/files/QuantilizersSaferAlternative.pdf) as a countermeasure for [Goodhart's Law](https://arxiv.org/pdf/1803.04585.pdf) and specification-gaming. Since these are such central problems in AI safety, I consider quantilizati... | https://www.lesswrong.com/posts/Rs6vZCrnQFWQ4p37P/when-to-use-quantilization |
# Philosophy as low-energy approximation
In 2015, Scott Alexander wrote a post originally titled [High Energy Ethics](https://slatestarcodex.com/2015/03/26/high-energy-ethics/). The idea is that when one uses an extreme thought experiment in ethics (people dying, incest, the extinction of humanity, etc.), this is like... | https://www.lesswrong.com/posts/aNaP8eCiKW7wZxpFt/philosophy-as-low-energy-approximation |
# Automated Nomic Game 2
I've been playing a game of [Nomic](https://legacy.earlham.edu/~peters/nomic.htm), where rules are interpreted by a [script](https://github.com/jeffkaufman/nomic/blob/master/validate.py) we edit instead of by the players. This morning Chelsea won our second game of Nomic with a timing attack:... | https://www.lesswrong.com/posts/Q4shHDqkMcDt7YmN9/automated-nomic-game-2 |
# Complexity Penalties in Statistical Learning
I am currently taking a course on statistical learning at the Australian Mathematical Sciences Institute Summer School. One idea that has appeared many times in the course is that a more complicated model is likely to have many short comings. This is because complicated m... | https://www.lesswrong.com/posts/boBSTYL3K4KSbh4ec/complexity-penalties-in-statistical-learning |
# If Rationality can be likened to a 'Martial Art', what would be the Forms?
In Martial Arts, we have 'forms' that allow us to practice our skills when not with a partner or in adversity.
In Rationality, I imagine that this would take the form of "brain teasers", but specifically regarding decision theory, overcoming ... | https://www.lesswrong.com/posts/xs4j9xSDbEntQKTNy/if-rationality-can-be-likened-to-a-martial-art-what-would-be |
# Alignment Newsletter #44
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_tUCuxpEyMbN1NYYC02aM/edit?usp=sharing) of ... | https://www.lesswrong.com/posts/54jDNkonygFheKL9H/alignment-newsletter-44 |
# Security amplification
An apparently aligned AI system may nevertheless behave badly with small probability or on rare “bad” inputs. The [reliability amplification](https://medium.com/ai-control/reliability-amplification-a96efa115687) problem is to reduce the failure probability of an aligned AI. The analogous **sec... | https://www.lesswrong.com/posts/hjEaZgyQ2iprDhkg8/security-amplification |
# Is the World Getting Better? A brief summary of recent debate
* The linked post is a response by Oliver Wiseman of CapX to Jason Hickel's recent [piece in The Guardian,](https://www.theguardian.com/commentisfree/2019/jan/29/bill-gates-davos-global-poverty-infographic-neoliberal) which claimed that Bill Gates' ... | https://www.lesswrong.com/posts/eTMgL7Cx8TsA9nedn/is-the-world-getting-better-a-brief-summary-of-recent-debate |
# Thoughts on Ben Garfinkel's "How sure are we about this AI stuff?"
I liked [this talk by Ben](https://www.youtube.com/watch?v=E8PGcoLDjVk&fbclid=IwAR12bx55ySVUShFMbMSuEMFi1MGe0JeoEGe_Jh1YUaWtrI_kP49pJQTyT40).
I think it raises some very important points. OTTMH, I think the most important one is: **We have no good c... | https://www.lesswrong.com/posts/AE9yM7ZaPiZ662BF8/thoughts-on-ben-garfinkel-s-how-sure-are-we-about-this-ai |
# My use of the phrase "Super-Human Feedback"
I've taken to calling Debate, Amplification, and Recursive Reward Modeling **"Super-human feedback" (SHF)** techniques. The point of this post is just to introduce that terminology and explain a bit why I like it and what I mean by it.
By calling something SHF I mean tha... | https://www.lesswrong.com/posts/naccwaCQEEBXK7hiJ/my-use-of-the-phrase-super-human-feedback |
# Test Cases for Impact Regularisation Methods
_Epistemic status: I’ve spent a while thinking about and collecting these test cases, and talked about them with other researchers, but couldn’t bear to revise or ask for feedback after writing the first draft for this post, so here you are._
A motivating concern in AI a... | https://www.lesswrong.com/posts/wzPzPmAsG3BwrBrwy/test-cases-for-impact-regularisation-methods |
# EA grants available (to individuals)
I'm considering applying for some kind of a grant from the effective altruism community. A quick sketch of the specifics is [here](https://www.lesswrong.com/posts/7XtmBcHMWknKEj4JY/does-the-ea-community-do-basic-science-grants-how-do-i-get#8bs5NuMnmENsDXnzu). Raemon replied there... | https://www.lesswrong.com/posts/NbNzLdAMesP5ejA2H/ea-grants-available-to-individuals |
# Open Thread February 2019
If it’s worth saying, but not worth its own post, you can put it here.
Also, if you are new to LessWrong and want to introduce yourself, this is the place to do it. Personal stories, anecdotes, or just general comments on how you found us and what you hope to get from the site and communit... | https://www.lesswrong.com/posts/JmyZJRfcQT5dDKpkY/open-thread-february-2019 |
# What is learning?
I want to know what learning actually _is._ Ideally you can provide a model that is compact and useful as possible, such that when I encounter the word 'learn' in daily life I could replace it with this model while seeing a complete set of unique gears within it such that I can look at a real-world... | https://www.lesswrong.com/posts/5ooJdcqvfXT8ChHLf/what-is-learning |
# Make an appointment with your saner self
Most people have had the experience of being able to articulate advice that they themselves do not follow, even though it applies to their situation as well. Usually this implies that there's some sort of internal conflict present—a competing commitment that gets in the way o... | https://www.lesswrong.com/posts/DFoNgGDBiiomKSjyu/make-an-appointment-with-your-saner-self |
# The Hamming Question
_This is a stub post, mostly existing so people can easily link to a post explaining what the Hamming question is. If you would like to write a real version of this post, ping me and I'll arrange to give you edit rights to this stub._
_For now, I am stealing the words from Jacobian's [event pos... | https://www.lesswrong.com/posts/P5k3PGzebd5yYrYqd/the-hamming-question |
# HCH is not just Mechanical Turk
HCH, introduced in [Humans consulting HCH](https://ai-alignment.com/humans-consulting-hch-f893f6051455), is a computational model in which a human answers questions using questions answered by another human, which can call other humans, which can call other humans, and so on. Each ste... | https://www.lesswrong.com/posts/4JuKoFguzuMrNn6Qr/hch-is-not-just-mechanical-turk |
# Reinforcement Learning in the Iterated Amplification Framework
When I think about [Iterated](https://blog.openai.com/amplifying-ai-training/) [Amplification](https://ai-alignment.com/iterated-distillation-and-amplification-157debfd1616) (IA), I usually think of a version that uses imitation learning for distillation... | https://www.lesswrong.com/posts/fq7Ehb2oWwXtZic8S/reinforcement-learning-in-the-iterated-amplification |
# The Case for a Bigger Audience
How do people feel like LW 2.0 is going? I'm impressed with the number and quality of posts that are being made, especially relative to the baseline of what LW 1.0 was like right before the relaunch. But I miss the lively discussions in the comments from the "Golden Age" of LW 1.0. ... | https://www.lesswrong.com/posts/2E3fpnikKu6237AF6/the-case-for-a-bigger-audience |
# When should we expect the education bubble to pop? How can we short it?
I won't attempt to summarise the case for there being an education bubble here (see links below for some pointers). Rather, my questions are:
1) _assuming_ there is an education bubble, when will it -- as bubbles tend to do -- pop?
(This plau... | https://www.lesswrong.com/posts/EBAccQwDWMiRCWnyk/when-should-we-expect-the-education-bubble-to-pop-how-can-we |
# Dojo on stress
Original post: [http://bearlamp.com.au/dojo-on-stress/](http://bearlamp.com.au/dojo-on-stress/)
* * *
This (thinking) [dojo](http://rationalitydojo.com.au/) came about because someone described their biggest problem as stress, relating to recent job-change events. I ran this dojo in Melbourne and S... | https://www.lesswrong.com/posts/mxv78X93gRses85Za/dojo-on-stress |
# The Argument from Philosophical Difficulty
(I'm reposting [this comment](https://www.lesswrong.com/posts/JbcWQCxKWn3y49bNB/disentangling-arguments-for-the-importance-of-ai-safety#daD7JREPtx2WDe2Wf) as a top-level post, for ease of future reference. The [context](https://www.lesswrong.com/posts/JbcWQCxKWn3y49bNB/dise... | https://www.lesswrong.com/posts/w6d7XBCegc96kz4n3/the-argument-from-philosophical-difficulty |
# Some Thoughts on Metaphilosophy
A powerful AI (or human-AI civilization) guided by wrong philosophical ideas would likely cause astronomical (or [beyond astronomical](https://www.lesswrong.com/posts/Qz6w4GYZpgeDp6ATB/beyond-astronomical-waste)) waste. Solving metaphilosophy is one way in which we can hope to avoid t... | https://www.lesswrong.com/posts/EByDsY9S3EDhhfFzC/some-thoughts-on-metaphilosophy |
# Probability space has 2 metrics
A metric is technically defined as a function from pairs of points to the non negitive reals. d:X×X→\[0,∞) With the properties that d(x,y)=d(y,x) and d(x,y)=0⟺x=y and d(x,y)+d(y,z)≥d(x,z).
Intuitively, a metric is a way of measuring how similar points are. Which points are nearby wh... | https://www.lesswrong.com/posts/63TcB9j7gv8Ampz8u/probability-space-has-2-metrics |
# Structured Concurrency Cross-language Forum
There are structured-concurrency-related efforts going on for different programming languages but the entire effort is kind of scattered, without people being aware of each other and speaking to each other.
If we had a common forum, we could share the use cases, the probl... | https://www.lesswrong.com/posts/TZ6zsypPQutE6ym9C/structured-concurrency-cross-language-forum |
# Fighting the allure of depressive realism
_Epistemic tragic backstory: Personal._
Earlier this week, on my first post on this site under this account, [shminux commented to the value of cognitive-behavioral therapy](https://www.lesswrong.com/posts/sN3fqSwNxxbrKypP3/depression-philosophizing#tLDqjdn6WqiXELtaT) to ge... | https://www.lesswrong.com/posts/H9zQ3CsZvejtBKD64/fighting-the-allure-of-depressive-realism |
# How important is it that LW has an unlimited supply of karma?
Question
--------
LessWrong users can up/downvote posts and comments, which then receive a karma boost (capped by the voters own karma). There is no limit to how many _different_ posts and comments one can do this to. In this sense there is an unlimited ... | https://www.lesswrong.com/posts/wgvwhokzYArtBKxtW/how-important-is-it-that-lw-has-an-unlimited-supply-of-karma |
# Why do you reject negative utilitarianism?
_Note (2022-03-15): For a recent defense of similar axiological views, see my [series on minimalist axiologies](https://forum.effectivealtruism.org/s/MBadsrYLmzLNmYjaj)._
* * *
(Crossposted on the [EA Forum](https://forum.effectivealtruism.org/posts/Wj8JymeiMN28Hrrjo/why-... | https://www.lesswrong.com/posts/8Eo52cjzxcSP9CHea/why-do-you-reject-negative-utilitarianism |
# Coherent behaviour in the real world is an incoherent concept
_Note: after putting this online, I noticed several problems with my original framing of the arguments. While I don't think they invalidated the overall conclusion, they did (ironically enough) make the post much less coherent. The version below has been ... | https://www.lesswrong.com/posts/vphFJzK3mWA4PJKAg/coherent-behaviour-in-the-real-world-is-an-incoherent |
# "Normative assumptions" need not be complex
I've [shown](https://arxiv.org/abs/1712.05812) that, even with simplicity priors, we can't figure out the preferences or rationality of a potentially irrational agent (such as a human $H$).
But we can get around that issue with '[normative assumptions](https://www.lesswro... | https://www.lesswrong.com/posts/QzsCrzGd4zkNwk9cd/normative-assumptions-need-not-be-complex |
# Would I think for ten thousand years?
Some AI safety ideas delegate key decisions to our idealised selves. This is sometimes phrased as "allowing versions of yourself to think for ten thousand years", or similar sentiments.
Occasionally, when I've objected to these ideas, it's been pointed out that any attempt to c... | https://www.lesswrong.com/posts/6zwW9oGaHbuMuvnmX/would-i-think-for-ten-thousand-years |
# Arguments for moral indefinability
_Epistemic status: I endorse the core intuitions behind this post, but am only moderately confident in the specific claims made._
Moral indefinability is the term I use for the idea that there is no ethical theory which provides acceptable solutions to all moral dilemmas, and ... | https://www.lesswrong.com/posts/ACo8Md94aX7qRpPi7/arguments-for-moral-indefinability |
# Functional silence: communication that minimizes change of receiver's beliefs
I just thought of this concept and I'd like to know if it is new or useful.
"Functional silence" is communication intended to carry as little information as possible. It intends for the beliefs of listeners to be changed as little as poss... | https://www.lesswrong.com/posts/g4M3oGGTE8JS6RmTE/functional-silence-communication-that-minimizes-change-of |
# Learning preferences by looking at the world
We’ve written up a [blog post](https://bair.berkeley.edu/blog/2019/02/11/learning_preferences/) about our recent [paper](https://openreview.net/forum?id=rkevMnRqYQ) that I’ve been linking to but haven’t really announced or explained. The key idea is that since we’ve optim... | https://www.lesswrong.com/posts/7f6DNZhracD7RvxMr/learning-preferences-by-looking-at-the-world |
# Nuances with ascription universality
_Part of this post is based on points made by Scott Garrabrant and Abram Demski at a recent AI safety workshop put on by Alex Zhu._
Paul Christiano recently published [a](https://ai-alignment.com/towards-formalizing-universality-409ab893a456) [sequence](https://ai-alignment.com/... | https://www.lesswrong.com/posts/R5Euq7gZgobJi5S25/nuances-with-ascription-universality |
# Layers of Expertise and the Curse of Curiosity
_Epistemic status: oversimplification of a process I'm confident about; meant as proof of concept._
_Related to:_ [_Double-Dipping in Dunning-Kruger_](https://www.lesswrong.com/posts/qS2dnpeQyppetgXqv/double-dipping-in-dunning-kruger)
Expertise comes in different, mos... | https://www.lesswrong.com/posts/7vQT9KwkKrf8i7HN4/layers-of-expertise-and-the-curse-of-curiosity |
# The RAIN Framework for Informational Effectiveness
*[Epistemic status: Medium-Uncertainty. I've only spent a few days thinking about this, but it seems to fit well for some specific problem so far.]*
The following describes one possible framework for understanding the usefulness of different sources of information.... | https://www.lesswrong.com/posts/s4TrCbCXvvWfkT2o6/the-rain-framework-for-informational-effectiveness |
# Individual profit-sharing?
Here's a sketch of an idea:
* Design an open-source legal agreement that two people sign.
* The contract states that each person agrees to give the other 1% of their annual earnings, each year for _X_ years. (Ideally _X_ = several decades; both duration & percentage could be customize... | https://www.lesswrong.com/posts/aAD225BtAa6XD46ce/individual-profit-sharing |
# Anchoring vs Taste: a model
Here I'll develop my [observation](https://www.lesswrong.com/posts/dYsc2EGdvsCjjM4Wt/three-anchorings-number-attitude-and-taste) that [anchoring](https://en.wikipedia.org/wiki/Anchoring) [bias](http://web.mit.edu/ariely/www/MIT/Chapters/CA.pdf) is formally similar to taste based preferenc... | https://www.lesswrong.com/posts/7gMkkEMQa78h8Bpmr/anchoring-vs-taste-a-model |
# Humans interpreting humans
In a [previous post](https://www.lesswrong.com/posts/7gMkkEMQa78h8Bpmr/anchoring-vs-taste-a-model), I showed how, given certain normative assumptions, one could distinguish agents $H$ for whom anchoring was a bias, from those $H'$ for which it was a preference.
But agent $H'$ looks clearl... | https://www.lesswrong.com/posts/eXiuk8Yjd7FsunhxM/humans-interpreting-humans |
# Three Kinds of Research Documents: Exploration, Explanation, Academic
*Aug 2020 Edit: Changed "Clarification" to "Exploration", thanks to a comment by Richard_Ngo*
*Epistemic Status: Low. This was a quick idea, but the grouping honesty doesn't work as well as I'd like. I still think it could be useful to some peopl... | https://www.lesswrong.com/posts/b8kRgm5gihovFKnRr/three-kinds-of-research-documents-exploration-explanation |
# Alignment Newsletter #45
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_tUCuxpEyMbN1NYYC02aM/edit?usp=sharing) of ... | https://www.lesswrong.com/posts/psKxyNGH9HuxvpFPB/alignment-newsletter-45 |
# New York Restaurants I Love: Breakfast
Previously (and more generally useful): [Restaurant Guide 1: Restaurants should not look like (most) restaurants](https://thezvi.wordpress.com/2017/03/05/restaurant-guide-1-restaurants-should-not-look-like-most-restaurants/), [Restaurant Guide 2: Pizza](https://thezvi.wordpress... | https://www.lesswrong.com/posts/s4N8XwWhzeYH6R68D/new-york-restaurants-i-love-breakfast |
# The Prediction Pyramid: Why Fundamental Work is Needed for Prediction Work
*Epistemic state: I feel like this post makes fairly intuitive claims, but I have uncertainty on many of the specifics.*
In data science, it is a common mistake for organizations to focus on specific exciting parts like machine learning and... | https://www.lesswrong.com/posts/zuAj86jcB5JB4dgqm/the-prediction-pyramid-why-fundamental-work-is-needed-for |
# Who owns OpenAI's new language model?
OpenAI published about [their very general language-writing & comprehension model](https://blog.openai.com/better-language-models/) ([a](https://web.archive.org/web/20190214171619/https://blog.openai.com/better-language-models/)) today.
Interestingly, they're not releasing the ... | https://www.lesswrong.com/posts/KB5CeQYdq5oeah3nd/who-owns-openai-s-new-language-model |
# How does OpenAI's language model affect our AI timeline estimates?
OpenAI [recently announced](https://blog.openai.com/better-language-models/) progress in NLP, using a large transformer-based language model to tackle a variety of tasks and breaking performance records in many of them. It also generates synthetic sh... | https://www.lesswrong.com/posts/WjxSFmm7GvWEMovzR/how-does-openai-s-language-model-affect-our-ai-timeline |
# Quantifying anthropic effects on the Fermi paradox
_[Crossposted](https://forum.effectivealtruism.org/posts/9p52yqrmhossG2h3r/quantifying-anthropic-effects-on-the-fermi-paradox) to the EA forum._
_Summary:_
I apply the self-indication assumption (a theory of anthropics) and some non-causal decision theories to the... | https://www.lesswrong.com/posts/Cq6iZ9bHPB4Kq85qo/quantifying-anthropic-effects-on-the-fermi-paradox |
# Cooperation is for Winners
Cross-posted from [Putanumonit](https://putanumonit.com/2019/02/12/cooperation-is-for-winners/).
* * *
My last post was about [learning to compete well from sports](https://putanumonit.com/2019/01/26/masculine-virtues/). But why compete at all? After all, I’m the one who wrote that [winn... | https://www.lesswrong.com/posts/yMzAPQDEE6qry82gA/cooperation-is-for-winners |
# So you want to be a wizard
This is a link post to a recent talk by Julia Evans, popular programming blogger, called "So you want to be a wizard". I'm linking it on LW because I think it's one of the better takes I've seen on how to "apply rationality" to programming. A few interesting examples of things she discusse... | https://www.lesswrong.com/posts/8sLmjbRZ4WYpi8Dgc/so-you-want-to-be-a-wizard |
# Some disjunctive reasons for urgency on AI risk
(This has been sitting in my drafts folder since August 2017. Robin Hanson's recent [How Lumpy AI Services?](https://www.overcomingbias.com/2019/02/how-lumpy-ai-services.html) made me think of it again. I'm not sure why I didn't post it back then. I may have wanted to ... | https://www.lesswrong.com/posts/8oSCw3z2dZgWjanqB/some-disjunctive-reasons-for-urgency-on-ai-risk |
# How the MtG Color Wheel Explains AI Safety
Duncan Sabien has a post titled [How the ‘Magic: The Gathering’ Color Wheel Explains Humanity](https://medium.com/s/story/the-mtg-color-wheel-c9700a7cf36d). Without the context of that post, or other experience with the MtG color wheel, this post will probably not make sens... | https://www.lesswrong.com/posts/9CKBtxWtjvminNTmC/how-the-mtg-color-wheel-explains-ai-safety |
# Pedagogy as Struggle
\[See also: [On Giving Advice](https://mindlevelup.wordpress.com/2017/04/08/on-giving-advice/) and [Recognizing vs Generating](https://mindlevelup.wordpress.com/2017/09/24/recognizing-vs-generating/)\]
So this quarter, I’ve been tutoring for an undergraduate computer science course, and one int... | https://www.lesswrong.com/posts/2HFDRtwMBB6gu8GG9/pedagogy-as-struggle |
# Why didn't Agoric Computing become popular?
I remember being quite excited when I first read about Agoric Computing. From the [authors' website](https://e-drexler.com/d/09/00/AgoricsPapers/agoricpapers.html):
>Like all systems involving goals, resources, and actions, computation can be viewed in economic terms. T... | https://www.lesswrong.com/posts/5XzQQHwYtSgrATzMC/why-didn-t-agoric-computing-become-popular |
# Graceful Shutdown
Within a structured concurrency world, graceful shutdown is a little understood, little addressed and confusing topic.
I've been fighting with it myself for almost a year (longer than with any other problem I can recall) and it haven't been a pleasant experience. If felt like running in circles. T... | https://www.lesswrong.com/posts/9ge4dSsFpv2DZRCeY/graceful-shutdown |
# Major Donation: Long Term Future Fund Application Extended 1 Week
(Crossposted to the [EA Forum](https://forum.effectivealtruism.org/posts/5G5xspvM8wro78Gds/major-donation-ltf-ea-fund-application-extended-1-week))
[The Long-Term Future EA Funds](https://app.effectivealtruism.org/funds/far-future) application round ... | https://www.lesswrong.com/posts/ZKsSuxHWNGiXJBJ9Z/major-donation-long-term-future-fund-application-extended-1 |
# Extraordinary ethics require extraordinary arguments
_Previous post: [Fighting the allure of depressive realism](https://www.lesswrong.com/posts/H9zQ3CsZvejtBKD64/fighting-the-allure-of-depressive-realism)_
_My blog entries are about a personal battle against depression and anxiety, from the point of view of someon... | https://www.lesswrong.com/posts/7GjmfXBcgzxn33FoZ/extraordinary-ethics-require-extraordinary-arguments |
# Can We Place Trust in Post-AGI Forecasting Evaluations?
*TLDR*
Think "A prediction market, where most questions are evaluated shortly after an AGI is developed." We could probably answer hard questions more easily post-AGI, so delaying them would have significant benefits.
## Motivation
Imagine that select pre-AG... | https://www.lesswrong.com/posts/HB6aGCPjtvpLk4K9t/can-we-place-trust-in-post-agi-forecasting-evaluations |
# Robin Hanson on Lumpiness of AI Services
Basically, in this post Robin Hanson argues that AI systems will have multiple separate components, rather than be a big uniform lump, arguing by analogy with the multiplicity of firms in the economy. Some excerpts:
> Long ago people like Marx and Engels predicted that the f... | https://www.lesswrong.com/posts/vLEa2kaaQsyDxvPu5/robin-hanson-on-lumpiness-of-ai-services |
# Avoiding Jargon Confusion
Previous discussion on jargon:
* [Against Naming Things](https://www.lesswrong.com/posts/6JrrCK3WDYmQMkgdT/against-naming-things-and-so-on) (by whales)
* [Common vs Expert Jargon](https://www.lesswrong.com/posts/DcRFTx62sTTRQo3Jw/common-vs-expert-jargon) (by me)
If you're proposing a ... | https://www.lesswrong.com/posts/c882tMwMMSWQbWyBC/avoiding-jargon-confusion |
# Is voting theory important? An attempt to check my bias.
You may have seen either or both of my two prior posts here, in which I first said I'm beginning to [seek a grant for some voting theory research](https://www.lesswrong.com/posts/7XtmBcHMWknKEj4JY/does-the-ea-community-do-basic-science-grants-how-do-i-get) aft... | https://www.lesswrong.com/posts/Bb7AYRypzpY7GYkdH/is-voting-theory-important-an-attempt-to-check-my-bias |
# Implications of GPT-2
I was impressed by [GPT-2](https://blog.openai.com/better-language-models/), to the point where I wouldn't be surprised if a future version of it could be used [pivotally](https://arbital.com/p/pivotal/) using existing protocols.
Consider generating half of a [Turing test](https://en.wikipedia... | https://www.lesswrong.com/posts/YJRb6wRHp7k39v69n/implications-of-gpt-2 |
# A Strange Situation
Hello all,
I am a newcomer to the site, but I have been lurking and reading the various posts on this site for about 8 months. I have to tell you, life has started making a lot more sense since I read these posts. However, I am feeling somewhat directionless after reading the archives, as I cann... | https://www.lesswrong.com/posts/F6AXKrMdzzSCHqaam/a-strange-situation |
# Epistemic Tenure
In this post, I will try to justify the following claim (which I am not sure how much I believe myself):
Let Bob be an individual that I have a lot intellectual respect for. For example, maybe Bob had a history of believing true things long before anyone else, or Bob has a discovered or invented so... | https://www.lesswrong.com/posts/7YG9zknYE8di9PuFg/epistemic-tenure |
# Decelerating: laser vs gun vs rocket
In our paper on [exploring the universe](http://www.fhi.ox.ac.uk/wp-content/uploads/intergalactic-spreading.pdf), some of our probes required huge quantities of reaction mass in order to decelerate on arrival.
This is due to the tyranny of the [rocket equation](https://en.wikipe... | https://www.lesswrong.com/posts/WnPa2KyaTcLXRYZp8/decelerating-laser-vs-gun-vs-rocket |
# Blackmail
Epistemic Status: Surprisingly controversial
Response to: [Why Blackmail Should Be Illegal](https://marginalrevolution.com/marginalrevolution/2019/02/why-should-blackmail-be-illegal.html) (Marginal Revolution), [Should Blackmail Be Legal?](https://www.econlib.org/should-blackmail-be-legal/) (David Henders... | https://www.lesswrong.com/posts/EYd63hYSzadcNnZTD/blackmail |
# Kocherga's leaflet
Some context: Kocherga is a [rationalist community hub in Moscow](https://www.lesswrong.com/posts/WmfapdnpFfHWzkdXY/rationalist-community-hub-in-moscow-3-years-retrospective).
There was a large crowdfunding campaign for printing the Russian HPMoR edition last year. Kocherga was the main local pic... | https://www.lesswrong.com/posts/nduYR6toaP5ySYFmz/kocherga-s-leaflet |
# [Link] OpenAI on why we need social scientists
[https://distill.pub/2019/safety-needs-social-scientists/](https://distill.pub/2019/safety-needs-social-scientists/)
> We believe the AI safety community needs to invest research effort in the human side of AI alignment. Many of the uncertainties involved are empirical... | https://www.lesswrong.com/posts/xkRRRZ7Pdny7AQK5r/link-openai-on-why-we-need-social-scientists |
# Impact Prizes as an alternative to Certificates of Impact
**I posted this one to the EA Forum, because it seemed a bit more relevant there. I think that LessWrong users will also find it interesting.**
*TLDR Example*
An EA donor puts up a $50k prize for distribution in 2022. In 2022, several projects that have sta... | https://www.lesswrong.com/posts/bqaeYRABLXKGt89FG/impact-prizes-as-an-alternative-to-certificates-of-impact |
# First steps of a rationality skill bootstrap
Epistemic status: slightly more advanced than mixing stuff in a pot and throwing it at a wall to see what sticks
This is approximately the first step in a rationality bootstrap system I threw together for myself which I am sharing because I think it's a useful exercise i... | https://www.lesswrong.com/posts/JEeZoSRCE5thXM4qM/first-steps-of-a-rationality-skill-bootstrap |
# Is there an assurance-contract website in work?
I bet I'm not the only one who, after finishing inadequate equilibria, thought in excitement "Ok, so where's the KickStarter for better Nash equlibria?".
I didn't find any existing site that does something like it, but i wonder if maybe someone in this community is wo... | https://www.lesswrong.com/posts/98HihPxpoRFja2JNF/is-there-an-assurance-contract-website-in-work |
# Pavlov Generalizes
Sarah Constantine recently wrote about [the Pavlov strategy](https://www.lesswrong.com/posts/3rxMBRCYEmHCNDLhu/the-pavlov-strategy) for iterated prisoner's dilemma. She points out that it deserves to be better-known, given that it has similar virtues to Tit-for-Tat (TfT). I strongly agree.
* Mu... | https://www.lesswrong.com/posts/XTgkhjNTEi97WHMi6/pavlov-generalizes |
# What i learned giving a lecture on NVC
Recently, I gave a lecture on [Nonviolent Communication](https://en.wikipedia.org/wiki/Nonviolent_Communication) (NVC) at a democratic school in Israel, the attendees (or at least the active ones), were mostly 11-12 year old girls, and a teacher.
The first thing i found was a ... | https://www.lesswrong.com/posts/cCcX4ZNHePoAcacHs/what-i-learned-giving-a-lecture-on-nvc |
# When does introspection avoid the pitfalls of rumination?
A theme I see being explored in the LW-sphere is the importance of connecting to system 1 information in various ways.
There's lots of stuff about meditation, [Focusing](https://medium.com/@ThingMaker/focusing-for-skeptics-6b949ef33a4f), [connecting to sub-a... | https://www.lesswrong.com/posts/FbadXDmwStjNenofp/when-does-introspection-avoid-the-pitfalls-of-rumination |
# Predictive Reasoning Systems
*Meta: This is meant to be a succinct and high-level overview. I expect to go more into detail on parts of the system in future posts. This work builds on [this previous post](https://www.lesswrong.com/posts/zuAj86jcB5JB4dgqm/the-prediction-pyramid-why-fundamental-work-is-needed-for).*
... | https://www.lesswrong.com/posts/FGJKj6sxx9q4FAmbS/predictive-reasoning-systems |
# Two Small Experiments on GPT-2
I did two small experiments on the GPT-2 small model. First experiment: can GPT-2-small answer sentiment analysis questions? (It can't.) Second experiment: When GPT-2 writes continuations of Howl, is it picking up the "Moloch in X!" template from its priming, or from a copy of Howl in ... | https://www.lesswrong.com/posts/s9sDyZ9AA3jKbM7DY/two-small-experiments-on-gpt-2 |
# Thoughts on Human Models
Human values and preferences are hard to specify, especially in complex domains. Accordingly, much AGI safety research has focused on approaches to AGI design that refer to human values and preferences _indirectly_, by learning a model that is grounded in expressions of human values (via sta... | https://www.lesswrong.com/posts/BKjJJH2cRpJcAnP7T/thoughts-on-human-models |
# What's your favorite LessWrong post?
This was last asked in 2014 on the [annual survey](https://www.lesswrong.com/posts/oGezscrQvPDgGvrbt/2014-less-wrong-census-survey). Since we have strong upvotes, you may vote for 2 posts. Each answer should link to a single post, and I will delete repeats. Please use a strong an... | https://www.lesswrong.com/posts/8EYtjxPXx8bjFRWoG/what-s-your-favorite-lesswrong-post |
# Ideas for Next Generation Prediction Technologies
2021 Note: This was written and posted in December, 2016. The date it shows on LessWrong is 2019; I believe this refers to a time the post was (very minorly) updated, as part of moving it to the Prediction-Driven Collaborative Reasoning Systems sequence.
Prediction ... | https://www.lesswrong.com/posts/vAzWfi5kDjZb27zD2/ideas-for-next-generation-prediction-technologies |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.