text stringlengths 300 320k | source stringlengths 52 154 |
|---|---|
# OpenAI: Preparedness framework
OpenAI released a beta version of their responsible scaling policy (though they don't call it that). See [summary page](https://openai.com/safety/preparedness), [full doc](https://cdn.openai.com/openai-preparedness-framework-beta.pdf), [OpenAI twitter thread](https://twitter.com/OpenAI... | https://www.lesswrong.com/posts/oPbiQfRotHYuC3wfE/openai-preparedness-framework |
# The Shortest Path Between Scylla and Charybdis
**tl;dr:** There's two diametrically opposed failure modes an alignment researcher can fall into: engaging in *excessively concrete* research whose findings won't timely generalize to AGI, and engaging in *excessively abstract* research whose findings won't timely conne... | https://www.lesswrong.com/posts/Zn3iF8qfFFWRrrZNR/the-shortest-path-between-scylla-and-charybdis |
# The 'Neglected Approaches' Approach: AE Studio's Alignment Agenda
*Many thanks to Samuel Hammond, Cate Hall, Beren Millidge, Sumner Norman, Steve Byrnes, Lucius Bushnaq, Joar Skalse, Kyle Gracey, Gunnar Zarncke, Ross Nordby, David Lambert, Simeon Campos, Bogdan Ionut-Cirstea, Ryan Kidd, and Eric Ho for critical comm... | https://www.lesswrong.com/posts/qAdDzcBuDBLexb4fC/the-neglected-approaches-approach-ae-studio-s-alignment |
# When scientists consider whether their research will end the world
Although there is currently no consensus about the future impact of AI, [many](https://aiimpacts.org/how-bad-a-future-do-ml-researchers-expect/) experts are [warning](https://wiki.aiimpacts.org/arguments_for_ai_risk/views_of_ai_developers_on_risk_fro... | https://www.lesswrong.com/posts/xoSh4MhtFarLYZ7eM/when-scientists-consider-whether-their-research-will-end-the |
# The Dark Arts
It is my understanding that you won all of your public forum debates this year. That's very impressive. I thought it would be interesting to discuss some of the techniques you used.
Of course! So, just for a brief overview for those who don't know, **public forum** is a 2v2 debate format, usually on a... | https://www.lesswrong.com/posts/djWftXndJ7iMPsjrp/the-dark-arts |
# Constellations are Younger than Continents
At the Bay Area Solstice, I heard the song [Bold Orion](https://humanistculture.bandcamp.com/track/bold-orion) for the first time. I like it a lot. It does, however, have one problem:
> He has seen the rise and fall of kings and continents and all,
> Rising silent, bold... | https://www.lesswrong.com/posts/YMakfmwZsoLdXAZhb/constellations-are-younger-than-continents |
# Assessment of AI safety agendas: think about the downside risk
*This is a post-response to* [*The 'Neglected Approaches' Approach: AE Studio's Alignment Agenda*](https://www.lesswrong.com/posts/qAdDzcBuDBLexb4fC/the-neglected-approaches-approach-ae-studio-s-alignment)*.*
* * *
You evidently follow a variant of [80... | https://www.lesswrong.com/posts/ETLq9ijtmM9xMtqaC/assessment-of-ai-safety-agendas-think-about-the-downside |
# A Universal Emergent
Decomposition of Retrieval Tasks in Language Models
*This work was done as a Master's thesis project at Conjecture, independent from the primary agenda of the organization. Paper available* [*here*](https://arxiv.org/abs/2312.10091)*, thesis* [*here*](https://drive.google.com/file/d/1PKQQnQH2pyX... | https://www.lesswrong.com/posts/uCuvFKnvzwh34GuX3/a-universal-emergent-decomposition-of-retrieval-tasks-in |
# Effective Aspersions: How the Nonlinear Investigation Went Wrong
### The New York Times
Picture a scene: the New York Times is releasing an article on Effective Altruism (EA) with an express goal to dig up every piece of negative information they can find. They contact Émile Torres, David Gerard, and Timnit Gebru, ... | https://www.lesswrong.com/posts/2vNHiaTb4rcA8PgXQ/effective-aspersions-how-the-nonlinear-investigation-went |
# Monthly Roundup #13: December 2023
I have not actually forgotten that the rest of the world exists. As usual, this is everything that wasn’t worth an entire post and is not being saved for any of the roundup post categories.
(Roundup post categories are currently AI, Medical and Health, Housing and Traffic, Dating,... | https://www.lesswrong.com/posts/sZ8f5NhGPCSkdKqm5/monthly-roundup-13-december-2023 |
# Chording "The Next Right Thing"
One of the songs we're planning for this year's Boston Secular Solstice ( [more](https://www.facebook.com/events/238479572316403/)) is [The Next Right Thing](https://en.wikipedia.org/wiki/The_Next_Right_Thing) from [Frozen II](https://en.wikipedia.org/wiki/Frozen_II). It's sad and dar... | https://www.lesswrong.com/posts/HCw2cgoLCqpqHcLXd/chording-the-next-right-thing |
# SociaLLM: proposal for a language model design for personalised apps, social science, and AI safety research
I first proposed this model [here](https://www.lesswrong.com/posts/aRwBw8W7G9gvLFJSW/a-proposal-for-improving-the-global-online-discourse#Self_supervised_pre_training), as a base model for a [proposed app to ... | https://www.lesswrong.com/posts/YN7PHizHnxinLsKvy/sociallm-proposal-for-a-language-model-design-for |
# Goal-Completeness is like Turing-Completeness for AGI
Turing-completeness is a useful analogy we can use to grasp why AGI will inevitably converge to “goal-completeness”.
By way of definition: An AI whose input is an arbitrary goal, which outputs actions to effectively steer the future toward that goal, is **goal-c... | https://www.lesswrong.com/posts/iFdnb8FGRF4fquWnc/goal-completeness-is-like-turing-completeness-for-agi |
# How does a toy 2 digit subtraction transformer predict the sign of the output?
====
Summary
=======
I examine a toy 1-layer transformer language model trained to do two-digit addition of the form $a-b=\pm c$. This model must first predict if the outcome is positive (+) or negative (-) and output the associate... | https://www.lesswrong.com/posts/pbj6tTZyakodxC9Ho/how-does-a-toy-2-digit-subtraction-transformer-predict-the |
# Interview: Applications w/ Alice Rigg
Alice Rigg, a mechanistic interpretability researcher from Ottawa, Canada, joins me to discuss her path, and we both reflect on the applications process for research/mentorship programs (*e.g.,* MATS, ASTRA Fellowship, ARENA, AISC).
She also is one of the moderators for a [Mech... | https://www.lesswrong.com/posts/gKeq7y9aqdnWG8GKf/interview-applications-w-alice-rigg |
# Paper: Tell, Don't Show- Declarative facts influence how LLMs generalize
This post is the abstract and introduction and figures from this [paper](https://arxiv.org/abs/2312.07779) ([Tweets](https://twitter.com/OwainEvans_UK/status/1737171321858240740)) by Alexander Meinke and myself.
Abstract
--------
We examine h... | https://www.lesswrong.com/posts/DR67cnvgCXw7CkXbd/paper-tell-don-t-show-declarative-facts-influence-how-llms |
# Don't Share Information Exfohazardous on Others' AI-Risk Models
(An "exfohazard" is information that's not dangerous to know for individuals, but which becomes dangerous if *widely* known. Chief example: AI capability insights.)
Different alignment researchers have a wide array of different models of AI Risk, and [... | https://www.lesswrong.com/posts/4XN2nJXABo6d5htZY/don-t-share-information-exfohazardous-on-others-ai-risk |
# Meaning & Agency
The goal of this post is to clarify a few concepts relating to AI Alignment under a common framework. The main concepts to be clarified:
* Optimization. Specifically, this will be a type of [Vingean agency](https://www.lesswrong.com/posts/fWiYdyicEaCHCcAKx/vingean-agency). It will split into [Sel... | https://www.lesswrong.com/posts/bnnhypM5MXBHAATLw/meaning-and-agency |
# What are the best Siderea posts?
I think the [Asshole Filters](https://www.lesswrong.com/posts/n86TNJS8k2Shd84kn/the-asshole-filter) post is underrated and everyone should read and share it. I also like the [post](https://siderea.dreamwidth.org/1237182.html) about economic class vs social class. Do the posts only ge... | https://www.lesswrong.com/posts/jWDSuSHuqzSAKpZoX/what-are-the-best-siderea-posts |
# Ronny and Nate discuss what sorts of minds humanity is likely to find by Machine Learning
Context: somebody at some point floated the idea that Ronny might (a) understand the argument coming out of the Quintin/Nora camp, and (b) be able to translate them to Nate. Nate invited Ronny to chat. The chat logs follow, lig... | https://www.lesswrong.com/posts/9x8nXABeg9yPk2HJ9/ronny-and-nate-discuss-what-sorts-of-minds-humanity-is |
# Escaping Skeuomorphism
In discussions about AI capabilities, there's a human aspect of *how* people access those capabilities which I think is important regardless of whether or not you're an AI doomer. Interface design can seem banal, but I think it's fundamental to how people interact with the singularity.
I.
==
... | https://www.lesswrong.com/posts/tHRCtKePGukHFozRD/escaping-skeuomorphism |
# Heuristics for preventing major life mistakes
*Edit: If you are interested in this subject, I'd like to interview you to learn about your own experience with making difficult life mistakes. Hearing people's stories will help validate my framework and give me material for future essays. And I might be able to offer s... | https://www.lesswrong.com/posts/WLyXSdPxaxCYB8sPY/heuristics-for-preventing-major-life-mistakes |
# Inside View, Outside View... And Opposing View
**Summary: I argue that the very useful concepts of** [**inside view and outside view**](https://www.lesswrong.com/tag/inside-outside-view) **are even more useful when completed with a third, equally fundamental, opposing view.**
Epistemic status: I am very sure this i... | https://www.lesswrong.com/posts/5SbwfQgHCoGRG9LQ9/inside-view-outside-view-and-opposing-view |
# A short dialogue on comparability of values
**Q:** Are all of a person's values comparable with each other? For example, is a candlelit dinner comparable to a sunset walk on a beach?
**A:** Of course. You can ask the person to choose between these two things. Their answer will give you information about what they... | https://www.lesswrong.com/posts/dfdozjnPdrXD9yiZ8/a-short-dialogue-on-comparability-of-values |
# Legalize butanol?
## ethanol
Alcoholic drinks are popular in most of the world. Excessive consumption of them is also a major public health problem. Bans have been attempted, sometimes [successfully](https://en.wikipedia.org/wiki/Saudi_Arabia), sometimes [unsuccessfully](https://en.wikipedia.org/wiki/Prohibition_in... | https://www.lesswrong.com/posts/ni2sLHJDySErL3pGt/legalize-butanol |
# What's the minimal additive constant for Kolmogorov Complexity that a programming language can achieve?
This is a question that I got from these posts, which essentially talked about how the constant issue impacts discussions of simplicity, and thus I wanted to ask a question about Kolmogorov Complexity:
https://ww... | https://www.lesswrong.com/posts/QSyKwJdxRj27gJGWv/what-s-the-minimal-additive-constant-for-kolmogorov |
# Matrix completion prize results
Earlier this year ARC [posted a prize for two matrix completion problems](https://www.alignment.org/blog/prize-for-matrix-completion-problems/). We received a number of submissions we considered useful, but not any complete solutions. We are closing the contest and awarding the follow... | https://www.lesswrong.com/posts/WiQ8pm2hpNFF3v8je/matrix-completion-prize-results |
# [Valence series] Appendix A: Hedonic tone / (dis)pleasure / (dis)liking
A.1 Post summary / Table of contents
====================================
*Part of the *[*Valence series*](https://www.lesswrong.com/s/6uDBPacS6zDipqbZ9)*.*
I talked about hedonic tone, a.k.a. pleasure or displeasure, a.k.a. liking or dislikin... | https://www.lesswrong.com/posts/xLmzMjxgZDaLyxZKb/valence-series-appendix-a-hedonic-tone-dis-pleasure-dis |
# On the future of language models
1\. Introduction
====================
1.1 Summary of key claims
-----------------------------
* Even without further breakthroughs in AI, language models will have big impacts in the coming years, as people start sorting out proper applications
* The early important... | https://www.lesswrong.com/posts/pnaZajFFWuXAnetpz/on-the-future-of-language-models |
# Gaia Network: a practical, incremental pathway to Open Agency Architecture
Introduction
------------
The Open Agency Architecture (OAA) [proposal](https://www.lesswrong.com/posts/pKSmEkSQJsCSTK6nH/an-open-agency-architecture-for-safe-transformative-ai) by Davidad has been gaining traction as a conceptual architectu... | https://www.lesswrong.com/posts/AKBkDNeFLZxaMqjQG/gaia-network-a-practical-incremental-pathway-to-open-agency |
# Brighter Than Today Versions
One of a small number of consistent elements in the many secular solstices descending from on Ray Arnold's is singing his song Brighter Than Today, usually to begin the transition from darkness back to light, and I think it serves this role well.
Here's what we sounded like [in 2022](ht... | https://www.lesswrong.com/posts/hk4jht48udreDSrMC/brighter-than-today-versions |
# Succession
*“A table beside the evening sea*
*where you sit shelling pistachios,*
*flicking the next open with the half-*
*shell of the last, story opening story,*
*on down to the sandy end of time.”*
### V1: Leaving
Deceleration is the hardest part. Even after burning almost all of my fuel, I’m still comi... | https://www.lesswrong.com/posts/CAzntXYTEaNfC9nB6/succession |
# How Would an Utopia-Maximizer Look Like?
When we talk of aiming for the good future for humanity – whether by aligning AGI or any other way – it's implicit that there are some futures that "humanity" as a whole would judge as good. That in some (perhaps very approximate) sense, humanity could be viewed as an agent w... | https://www.lesswrong.com/posts/okkEaevbXCSusBoE2/how-would-an-utopia-maximizer-look-like |
# My best guess at the important tricks for training 1L SAEs
*TL;DR: this quickly-written post gives a list of my guesses of the most important parts of training a *[*Sparse Autoencoder*](https://transformer-circuits.pub/2023/monosemantic-features) *on a*[*1L Transformer*](https://huggingface.co/NeelNanda/GELU_1L512W_... | https://www.lesswrong.com/posts/fifPCos6ddsmJYahD/my-best-guess-at-the-important-tricks-for-training-1l-saes |
# Why is capnometry biofeedback not more widely known?
What capnometry biofeedback is and why it seems important
=========================================================
A capnometer is a device that measures levels of carbon dioxide in your blood in a non-invasive way (think of it sort of like a pulse oximeter for ... | https://www.lesswrong.com/posts/xRBLn4LWLh7Sty876/why-is-capnometry-biofeedback-not-more-widely-known |
# Prediction Markets aren't Magic
One common theme that I come across quite a bit in the prediction market space is:
> Prediction markets would solve \[x\][^p8xa4gkpgbj]
And the proposal for "solving" \[x\] is:
1. Set up a prediction market
2. ???
3. Profit
These people need to consider the idea that "predictio... | https://www.lesswrong.com/posts/zLnHk9udC28D34GBB/prediction-markets-aren-t-magic |
# On OpenAI’s Preparedness Framework
Previously: [On RSPs](https://thezvi.substack.com/p/on-responsible-scaling-policies-rsps).
#### **Be Prepared**
[OpenAI introduces their preparedness framework](https://openai.com/safety/preparedness) for safety in frontier models.
A summary of the biggest takeaways, which I wi... | https://www.lesswrong.com/posts/hQPfLsDKWtdvMwyyr/on-openai-s-preparedness-framework |
# AI Safety Chatbot
Hello World! The [AISafety.info](https://aisafety.info/) team is launching a prototype of the [AI Safety Chatbot](... | https://www.lesswrong.com/posts/bT8yyJHpK64v3nh2N/ai-safety-chatbot |
# Rating my AI Predictions
9 months ago [I predicted trends I expected to see in AI over the course of 2023](https://aizi.substack.com/p/whats-left-for-agi-besides-scale). Here's how I did (bold indicates they happened, italics indicates they didn't, neither-bold-nor-italics indicates unresolved):
1. **ChatGPT (or s... | https://www.lesswrong.com/posts/dtrmr6Fn5AyP5GosQ/rating-my-ai-predictions |
# AI #43: Functional Discoveries
We get innovation in functional search. In an even more functional search, we finally get a Nature paper submitted almost two years ago, in which AI discovered a new class of antibiotic. That’s pretty damn exciting, with all the implications thereof.
OpenAI continued its rapid pace of... | https://www.lesswrong.com/posts/WaDFCrd6KEwojLXgj/ai-43-functional-discoveries |
# "Alignment" is one of six words of the year in the Harvard Gazette
To my knowledge, the Gazette's words of the year are nominated and voted on by Harvard faculty. This year's words were Disruption, Combustible, Resilience, Heat, Alignment, and Hope. Here's what they wrote about alignment:
> **Isaac Kohane, Chair, D... | https://www.lesswrong.com/posts/ydJ7zfnsdvdjZMWEb/alignment-is-one-of-six-words-of-the-year-in-the-harvard |
# Attention on AI X-Risk Likely Hasn't Distracted from Current Harms from AI
## Summary
In the past year, public fora have seen growing concern about existential risk (henceforth, x-risk) from AI. The thought is that we could see [transformative AI in the coming years or decades](https://www.cold-takes.com/where-ai-f... | https://www.lesswrong.com/posts/5rexNxtZgkEQBi3Sd/attention-on-ai-x-risk-likely-hasn-t-distracted-from-current |
# Pseudonymity and Accusations
Here's a category of situations I'm not sure how to think about: Avery, writing under a pseudonym ('Alex'), accuses Pat of something, let's say abuse. A major motivation of Avery's is for people to know to consider this information before in their interactions with Pat. Pat claims that a... | https://www.lesswrong.com/posts/HTLd3R9qgvZsThuWW/pseudonymity-and-accusations |
# Most People Don't Realize We Have No Idea How Our AIs Work
This point feels fairly obvious, yet seems worth stating explicitly.
Those of us familiar with the field of AI after the deep-learning revolution know perfectly well that we have no idea how our ML models work. Sure, we have an understanding of the dynamics... | https://www.lesswrong.com/posts/CpjTJtW2RNKvzAehG/most-people-don-t-realize-we-have-no-idea-how-our-ais-work |
# The LessWrong 2022 Review: Review Phase
[This year's LessWrong review](https://www.lesswrong.com/reviewVoting/2022?sort=needsReview) nomination phase ended a few days ago, with 339 posts nominated. For comparison, 291 posts were nominated in the 2021 review.
Nomination Phase Results
------------------------
Here ... | https://www.lesswrong.com/posts/Lqf7H9zREnRbmWL4M/the-lesswrong-2022-review-review-phase |
# Employee Incentives Make AGI Lab Pauses More Costly
Disclaimer: I know approximately nothing about internal governance at AGI labs. My current understanding is that the main AGI labs have basically no plans for how to navigate a conditional pause, and this post is predicated on this assumption.
TL;DR: If an AGI lab... | https://www.lesswrong.com/posts/GXrevJFeGiPkkM2hy/employee-incentives-make-agi-lab-pauses-more-costly |
# The problems with the concept of an infohazard as used by the LW community [Linkpost]
This is going to be a linkpost from Beren on some severe problems that come with the use of the concept of an infohazard on LW.
The main problem I see that are relevant to infohazards are that it encourages a "Great Man Theory" of... | https://www.lesswrong.com/posts/AM38ydkG8qJE2NEGW/the-problems-with-the-concept-of-an-infohazard-as-used-by |
# Review Report of Davidson on Takeoff Speeds (2023)
I have spent some time studying [Tom Davidson](https://www.lesswrong.com/users/tom-davidson-1?from=post_header)’s Open Philanthropy report on [what a compute-centric framework says about AI takeoff speeds](https://www.lesswrong.com/posts/Gc9FGtdXhK9sCSEYu/what-a-com... | https://www.lesswrong.com/posts/ikWvhgGFYnwY5nWyk/review-report-of-davidson-on-takeoff-speeds-2023 |
# Synthetic Restrictions
Introduction
============
As a child, I have read many stories, delving into everything from geography to Spanish literature, crime novels, and global history books. Each day, before or after school, I'd go to the local library, eagerly choosing the books that would be my companions for the w... | https://www.lesswrong.com/posts/WXG6v6pDfP6HpfP72/synthetic-restrictions |
# AI safety advocates should consider providing gentle pushback following the events at OpenAI
AI safety advocates for the most part have taken an exclusively–and potentially excessively–friendly and cooperative approach with AI firms–and especially OpenAI. I am as guilty of this as anyone[^4tf39bqy9em]–but after the ... | https://www.lesswrong.com/posts/55aJfiSYhCn4Phgib/ai-safety-advocates-should-consider-providing-gentle |
# Idealized Agents Are Approximate Causal Mirrors (+ Radical Optimism on Agent Foundations)
**Epistemic status:** I'm currently unsure whether that's [a fake framework](https://www.lesswrong.com/posts/wDP4ZWYLNj7MGXWiW/in-praise-of-fake-frameworks), [a probably-wrong mechanistic model](https://www.lesswrong.com/posts/... | https://www.lesswrong.com/posts/hx5wTeBSdf4bsYnY9/idealized-agents-are-approximate-causal-mirrors-radical |
# Thoughts on Max Tegmark's AI verification
In [Max Tegmark - How to keep AI under control - TED Talk](https://www.ted.com/talks/max_tegmark_how_to_keep_ai_under_control) Max presents a method to make AI systems safe (all images from that talk's slides):
 studying various techniques for fine-tuning strong models on downstream tasks using labels produced by weak models. They call this pro... | https://www.lesswrong.com/posts/4KLCygqTLsMBM3KFR/measurement-tampering-detection-as-a-special-case-of-weak-to |
# Fact Finding: Attempting to Reverse-Engineer Factual Recall on the Neuron Level (Post 1)
_If you've come here via [3Blue1Brown](https://www.youtube.com/watch?v=9-Jl0dxWQs8), hi! If want to learn more about interpreting neural networks in general, here are some resources you might find useful:_
* _[My getting starte... | https://www.lesswrong.com/posts/iGuwZTHWb6DFY3sKB/fact-finding-attempting-to-reverse-engineer-factual-recall |
# Fact Finding: Simplifying the Circuit (Post 2)
_This is the second post in the Google DeepMind mechanistic interpretability team’s [investigation into how language models recall facts](https://www.alignmentforum.org/s/hpWHhjvjn67LJ4xXX). This post focuses on distilling down the fact recall circuit and models a more ... | https://www.lesswrong.com/posts/3tqJ65kuTkBh8wrRH/fact-finding-simplifying-the-circuit-post-2 |
# Fact Finding: Trying to Mechanistically Understanding Early MLPs (Post 3)
_This is the third post in the Google DeepMind mechanistic interpretability team’s investigation into [how language models recall facts](https://www.alignmentforum.org/s/hpWHhjvjn67LJ4xXX). This post focuses on mechanistically understanding ho... | https://www.lesswrong.com/posts/CW5onXm6uZxpbpsRk/fact-finding-trying-to-mechanistically-understanding-early |
# Fact Finding: How to Think About Interpreting Memorisation (Post 4)
_This is the fourth post in the Google DeepMind mechanistic interpretability team’s investigation into [how language models recall facts](https://www.alignmentforum.org/s/hpWHhjvjn67LJ4xXX). In this post, we take a step back and consider the problem... | https://www.lesswrong.com/posts/JRcNNGJQ3xNfsxPj4/fact-finding-how-to-think-about-interpreting-memorisation |
# Fact Finding: Do Early Layers Specialise in Local Processing? (Post 5)
_This is the fifth post in the Google DeepMind mechanistic interpretability team’s investigation into [how language models recall facts](https://www.alignmentforum.org/s/hpWHhjvjn67LJ4xXX). This post is a bit tangential to the main sequence, and ... | https://www.lesswrong.com/posts/xE3Y9hhriMmL4cpsR/fact-finding-do-early-layers-specialise-in-local-processing |
# The Next Right Token
After many repeats of Frozen II's "The Next Right Thing" in preparation for [secular solstice](https://www.jefftk.com/p/secular-solstice-call-for-singers-and-musicans), [figuring out the chords](https://www.jefftk.com/p/chording-the-next-right-thing) and practicing it with Julia, it struck me th... | https://www.lesswrong.com/posts/LvDyEKepLDMbEQb9X/the-next-right-token |
# AI Girlfriends Won't Matter Much
Love and sex are pretty fundamental human motivations, so it’s not surprising that they are incorporated into our vision of future technology, including AI.
[. We discuss developments in AI and AI safety. No technical background required.
Subscribe [here](https://newsletter.safe.ai/subscribe?utm_medium=web&utm_source=subscribe-widget-... | https://www.lesswrong.com/posts/HEuDwEk22JfCBHh9o/aisn-28-center-for-ai-safety-2023-year-in-review |
# A Crisper Explanation of Simulacrum Levels
I've read the previous work on [Simulacrum Levels](https://www.lesswrong.com/tag/simulacrum-levels), and I've seen people express some confusion regarding how they work. I'd had some of those confusions myself when I first encountered the concept, and I think they were caus... | https://www.lesswrong.com/posts/KnQpzYRR4ogPNtzem/a-crisper-explanation-of-simulacrum-levels |
# Viral Guessing Game
Julia introduced me to [Metazooa](https://metazooa.com/), a taxonomic guessing game. You're trying to figure out what animal the computer has chosen. Each time you guess the computer will reveal the lowest taxonomic clade that they both belong to. For example, if you guessed platypus and the answ... | https://www.lesswrong.com/posts/pJB6KDD8DjKJPKkkZ/viral-guessing-game |
# align your latent spaces
> I learned very early the difference between knowing the name of something and knowing something.
— Richard Feynman
Here is a simplified diagram of the mind of someone reading English text and writing English text in response:
, do you think it's a good idea to have a baby now? What factors would you take into account (e.g. age)?
Today I saw a tweet by Eliezer Yudkowski t... | https://www.lesswrong.com/posts/voCirwwanPKFBvyBN/would-you-have-a-baby-in-2024 |
# Defense Against The Dark Arts: An Introduction
Epistemic status: I’m a non expert analyzing personal anecdotes and lower level political issues, take with a grain of salt.
Edit: The post is specifically about the Dark Arts as we see them in politics, not in intra or interpersonal relations. Thank you AnniePosting f... | https://www.lesswrong.com/posts/Mf8wmJG6HuYSikwqJ/defense-against-the-dark-arts-an-introduction |
# Freeze Dried Raspberry Truffles
I've historically found a lot of "raspberry flavored" things pretty disappointing. Raspberry lollypops? Raspberry slush? Anything with raspberry extract? Even raspberry jam is a lot worse than just [squishing up some frozen raspberries](https://www.jefftk.com/p/jam-is-obsolete) becaus... | https://www.lesswrong.com/posts/dZmpvtycXsnC5z4zc/freeze-dried-raspberry-truffles |
# Why does expected utility matter?
In the context of decision making under uncertainty we consider the strategy of maximizing the expected monetary revenue and expected utility; we provide an argument to show that under certain hypothesis it is rational to maximize expected monetary revenue, then we show that the arg... | https://www.lesswrong.com/posts/ALtFyrzvrNZPbfKBJ/why-does-expected-utility-matter |
# Anki setup best practices?
I'm currently trying to use Anki to crash course japanese vocab before a trip in a few weeks. I feel like the defaults set a card too far into the future after I press Easy even if I had already gotten it wrong several times ahead of that.
Are there like a set of extensions that you'... | https://www.lesswrong.com/posts/NFuHHrzbgWhiTPhkW/anki-setup-best-practices |
# How Emergency Medicine Solves the Alignment Problem
[Emergency medical technicians](https://en.wikipedia.org/wiki/Emergency_medical_technician) (EMTs) are not licensed to practice medicine. An EMT license instead grants the authority to perform a [specific set of interventions](https://www.ems.gov/national-ems-scope... | https://www.lesswrong.com/posts/WrqjAgpivxGFofECF/how-emergency-medicine-solves-the-alignment-problem |
# Link Collection: Impact Markets
0\. Readme (or don't)
=====================
* This is **not** a literature review. I'll vouch for links with an associated archived link, author name, and summary, but not the others.
* Last updated: Dec 2023
1\. Overviews
=============
* [Impact Markets: The Annoying Details... | https://www.lesswrong.com/posts/ogPEbdmeFmucB8v6e/link-collection-impact-markets |
# Flagging Potentially Unfair Parenting
Recently I made a post in our [kids group](https://www.jefftk.com/news/kidsgroup) about something Lily (9y) had done that was funny, mischievous, and also potentially embarrassing. A friend asked whether Lily knew I was writing about her antics and said they would have felt mort... | https://www.lesswrong.com/posts/M6MjKmp3xZYe2FiAA/flagging-potentially-unfair-parenting |
# Picasso in the Gallery of Babel
tl;dr
-----
When (not if) generative models get so good at faking human artefacts that no human can distinguish between human and synthetic artefacts, Strange Things will happen to the way humans consume and think about and make art.
Claim 1: Models get better
----------------------... | https://www.lesswrong.com/posts/6GgnrfP4XFsYC29oT/picasso-in-the-gallery-of-babel |
# Environmental allergies are curable? (Sublingual immunotherapy)
I used to have severe environmental allergies to cats, dogs, and the outdoors. As I child, I woke up crying most mornings because of my allergies, and I would occasionally wake up with croup and difficulty breathing. Once I had to be taken the hospital.... | https://www.lesswrong.com/posts/WWJuw5oy44ZQojw2M/environmental-allergies-are-curable-sublingual-immunotherapy |
# METR is hiring!
This is a quick update that [METR (formerly ARC Evals)](http://metr.org) is recruiting for four positions. I encourage you to err on the side of applying to positions that interest you even if you’re unsure about your fit! We’re able to sponsor US visas for all the roles below except Research Assista... | https://www.lesswrong.com/posts/6Tm4RbaXpHcXxDwSD/metr-is-hiring |
# Differential Optimization Reframes and Generalizes Utility-Maximization
Consider the following world: a set $P$ of past nodes, a set of environment nodes $E$ a set of actor/agent nodes $A$, and a set of future nodes $F$. Consider our previously defined function $\mathrm{Op}(A; p, f)$ for $p \in P, f \in F$, which re... | https://www.lesswrong.com/posts/3L6tWDaTpuNLD9paR/differential-optimization-reframes-and-generalizes-utility |
# 5. Moral Value for Sentient Animals? Alas, Not Yet
*Part 5 of* [*AI, Alignment, and Ethics*](https://www.lesswrong.com/s/QypaioYFd3CMFhefn)*. This will probably make more sense if you start with* [*Part 1*](https://www.lesswrong.com/s/QypaioYFd3CMFhefn/p/umnEwbK7sWSAnZ9vN)*.*
**TL;DR** In Parts [1](https://www.less... | https://www.lesswrong.com/posts/FQ4DsmswAK77Ei6pH/5-moral-value-for-sentient-animals-alas-not-yet |
# AGI will be made of heterogeneous components, Transformer and Selective SSM blocks will be among them
This post is prompted by two recent pieces:
First, in the podcast "[Emergency Pod: Mamba, Memory, and the SSM Moment](https://www.cognitiverevolution.ai/emergency-pod-mamba-memory-and-the-ssm-moment/)", Nathan Labe... | https://www.lesswrong.com/posts/Btom6dX5swTuteKce/agi-will-be-made-of-heterogeneous-components-transformer-and |
# Critical review of Christiano's disagreements with Yudkowsky
*This is a review of Paul Christiano's article "*[*where I agree and disagree with Eliezer*](https://www.alignmentforum.org/posts/CoZhXrhpQxpy9xw9y/where-i-agree-and-disagree-with-eliezer)*". Written for the* [*LessWrong 2022 Review*](https://www.lesswrong... | https://www.lesswrong.com/posts/8HYJwQepynHsRKr6j/critical-review-of-christiano-s-disagreements-with-yudkowsky |
# In Defense of Epistemic Empathy
**TLDR:** *Why think your ideological opponents are unreasonable? Common reasons: their views are (1) absurd, or (2) refutable, or (3) baseless, or (4) conformist, or (5) irrational. None are convincing.*
* * *
Elizabeth is skeptical about the results of the 2020 election. Theo thi... | https://www.lesswrong.com/posts/8kF6WZC7XBCwcRmrX/in-defense-of-epistemic-empathy |
# NYT is suing OpenAI&Microsoft for alleged copyright infringement; some quick thoughts
[Unpaywalled article](https://www.nytimes.com/2023/12/27/business/media/new-york-times-open-ai-microsoft-lawsuit.html?mwgrp=a-dbar&unlocked_article_code=1.JE0.6qro.Ll97Re6V8jrh&smid=url-share), [the lawsuit](https://nytco-assets.ny... | https://www.lesswrong.com/posts/uuwLZ9XLoziagfupW/nyt-is-suing-openai-and-microsoft-for-alleged-copyright |
# Natural Latents: The Math
Natural latents are a relatively elegant piece of math which we figured out over the past year, in our efforts to boil down and generalize various results and examples involving natural abstraction. In particular, this new framework handles approximation well, which was a [major piece](http... | https://www.lesswrong.com/posts/dWQWzGCSFj6GTZHz7/natural-latents-the-math |
# Merry Christmas Everyone!
This was my Christmas Card from my parents this year:
AI-generated and not at all terrifying. Happy New Year! | https://www.lesswrong.com/posts/73zARgmMayy2gkvZk/merry-christmas-everyone |
# E.T. Jaynes Probability Theory: The logic of Science I
This book (available as a free pdf [here](http://www.med.mcgill.ca/epidemiology/hanley/bios601/GaussianModel/JaynesProbabilityTheory.pdf)) is one of the foundational texts for Rationalists; Eliezer refers to Jaynes as [The Level Above Mine](https://www.lesswrong... | https://www.lesswrong.com/posts/KN3BYDkWei9ADXnBy/e-t-jaynes-probability-theory-the-logic-of-science-i |
# Gunpowder as metaphor for AI
The exothermic reaction of combusting gunpowder is a continuous but fast process. Sufficiently fast that a human will perceive it as discontinuous.
This is just refining a metaphor for a subject that's been hashed out many times [before](https://www.lesswrong.com/posts/cHJxSJ4jBmBRGtbaE... | https://www.lesswrong.com/posts/oXsEz3ZGyQbXDEpgc/gunpowder-as-metaphor-for-ai |
# How to develop a photographic memory 1/3
* Mastery of research, memory and logic is key to rebuilding a metaphorical Library of Alexandria. The first asks: how do we find knowledge? The second asks: how do we remember the knowledge we find? The third asks: how do we apply that knowledge accurately? I will be dealing... | https://www.lesswrong.com/posts/EiXgowiPwtg62Ksxf/how-to-develop-a-photographic-memory-1-3 |
# AI #44: Copyright Confrontation
The New York Times has thrown down the gauntlet, suing OpenAI and Microsoft for copyright infringement. Others are complaining about recreated images in the otherwise deeply awesome MidJourney v6.0. As is usually the case, the critics misunderstand the technology involved, complain ab... | https://www.lesswrong.com/posts/3GzRrqLAcdDXzbqc4/ai-44-copyright-confrontation |
# Building Trust in Strategic Settings
In the [previous post](https://www.lesswrong.com/posts/uDojCNZauajeN9vNc/an-ontology-for-strategic-epistemology), we saw some different types of claims that we might want software systems to automatically evaluate. Ideally, we'd like our software systems to incorporate accurate a... | https://www.lesswrong.com/posts/XZmtcxTkAYJm9gF5J/building-trust-in-strategic-settings |
# Distributed Strategic Epistemology
In this post we put all of the [pieces](https://www.lesswrong.com/posts/uDojCNZauajeN9vNc/an-ontology-for-strategic-epistemology) we've discussed [so far](https://www.lesswrong.com/posts/XZmtcxTkAYJm9gF5J/building-trust-in-strategic-settings) together into an network of [epistemolo... | https://www.lesswrong.com/posts/HiRbzpuPrCs2xZpdB/distributed-strategic-epistemology |
# Social Choice Theory and Logical Handshakes
How can we use techniques from [formal verification](https://en.wikipedia.org/wiki/Formal_verification) to enable socially optimal outcomes in situations where individual incentives are misaligned with the social good?
Suppose that we have a good enough implementation of ... | https://www.lesswrong.com/posts/6hgmEvEdrf6SYBiqK/social-choice-theory-and-logical-handshakes |
# Will 2024 be very hot? Should we be worried?
*tl;dr: There are several trends which suggest that global temperatures over the next year will experience a short-term increase, relative to the long-term increase in temperatures caused by man-made global warming. *
*Credits: Most of the information comes from* [*Berkl... | https://www.lesswrong.com/posts/kiACM88mT3tMx5TGx/will-2024-be-very-hot-should-we-be-worried |
# Progress links digest, 2023-12-29: Rayleigh's oil drop experiment and more
Somewhat delayed owing to the holidays. As always, to follow news and announcements in a more timely fashion, follow me on [Twitter](https://twitter.com/intent/follow?screen_name=jasoncrawford), [Threads](https://www.threads.net/@jasoncrawfor... | https://www.lesswrong.com/posts/rqzopPDRzHCqWeiK8/progress-links-digest-2023-12-29-rayleigh-s-oil-drop |
# Dark Skies Book Review
Book review: Dark Skies: Space Expansionism, Planetary Geopolitics, and
the Ends of Humanity, by Daniel Deudney.
Dark Skies is an unusually good and bad book.
Good in the sense that 95% of the book consists of uncontroversial,
scholarly, mundane claims that accurately describe the view... | https://www.lesswrong.com/posts/hp2KoAWB7n69PT8HP/dark-skies-book-review-1 |
# The Plan - 2023 Version
Background: [The Plan](https://www.lesswrong.com/posts/3L46WGauGpr7nYubu/the-plan), [The Plan: 2022 Update](https://www.lesswrong.com/posts/BzYmJYECAc3xyCTt6/the-plan-2022-update). If you haven’t read those, don’t worry, we’re going to go through things from the top this year, and with modera... | https://www.lesswrong.com/posts/HfqbjwpAEGep9mHhc/the-plan-2023-version |
# Optimization Markets
How can we coordinate on solving important computational problems? And how can we do this safely?
Computing Markets
-----------------
We can think of computing power as a resource, and one way to get people to contribute their resources to a project you think is important is to pay them! The g... | https://www.lesswrong.com/posts/pBkoHBYpvfouaMKEY/optimization-markets |
# When Can Optimization Be Done Safely?
When facing a problem in [computer science](https://www.lesswrong.com/tag/computer-science), one can often write a program that solves that problem directly. Usually, a programmer understands why their programs do what they do. [Not always](https://stackoverflow.com/questions/18... | https://www.lesswrong.com/posts/guqG7hmSt49i7LJ6y/when-can-optimization-be-done-safely |
# If Clarity Seems Like Death to Them
> "—but if one hundred thousand [normies] can turn up, to show their support for the [rationalist] community, why can't you?"
>
> I said wearily, "Because every time I hear the word _community_, I know I'm being manipulated. If there is such a thing as _the [rationalist] community... | https://www.lesswrong.com/posts/yFZH2sBsmmqgWm4Sp/if-clarity-seems-like-death-to-them |
# How to develop a photographic memory 2/3
As you may have guessed, these first two (2) posts give you the building blocks of mnemonics. The final post will polish some theories I've developed, refine the process in building personal mnemonic techniques, and finally: *how* to acquire a photographic memory.
As was th... | https://www.lesswrong.com/posts/j7tEDrwnfjcswA46y/how-to-develop-a-photographic-memory-2-3 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.