rationalnewsletter

Issue #223: Full List

Study finds that 80% of the “most transformative” medicines approved for clinical use by the FDA could be traced back to one or several basic discoveries, which occurred on average three decades before approval. // drive.google.com

# Instrumental

Focusing // CFAR 2017, 16 min

Hiring Programmers in Academia // jkaufman, 1 min

# Epistemic

Levels of Pluralism // adamShimi, 17 min

Bucket Errors // CFAR 2017, 12 min

Questions for a Theory of Narratives // Marv K, 5 min

Double Crux // CFAR 2017, 13 min

Forer Statements As Updates And Affirmations // Scott Alexander, 5 min

# Ai

Reward is not the optimization target // TurnTrout, 13 min

AGI ruin scenarios are likely (and disjunctive) // So8res, 7 min

Conjecture: Internal Infohazard Policy // NPCollapse, 21 min

Brainstorm of things that could force an AI team to burn their lead // So8res, 14 min

Principles of Privacy for Alignment Research // johnswentworth, 8 min

Abstracting The Hardness of Alignment: Unbounded Atomic Optimization // adamShimi, 20 min

Distillation Contest - Results and Recap // Aris, 8 min

AGI-level reasoner will appear sooner than an agent; what the humanity will do with this reasoner is critical // Roman Leventov, 1 min

AGI-level reasoner will appear sooner than an agent; and the humanity will do with this reasoner is critical // Roman Leventov, 1 min

AI ethics vs AI alignment // Wei_Dai, 1 min

Finding Skeletons on Rashomon Ridge // David Udell, 8 min

How transparency changed over time // ViktoriaMalyasova, 8 min

Alignment being impossible might be better than it being really difficult // martinsq, 2 min

[ASoT] Humans Reflecting on HRH // leogao, 2 min

Incoherence of unbounded selfishness // emmab, 1 min

Comparing Four Approaches to Inner Alignment // Lucas Teixeira, 11 min

Seeking beta readers who are ignorant of biology but knowledgeable about AI safety // Holly_Elmore, 1 min

How optimistic should we be about AI figuring out how to interpret itself? // oh54321, 1 min

Quantum Advantage in Learning from Experiments // dennis-towne, 1 min

Active Inference as a formalisation of instrumental convergence // Roman Leventov, 3 min

Defining Optimization in a Deeper Way Part 4 // Jemist, 6 min

How much should we worry about mesa-optimization challenges? // sudo, 2 min

Information theoretic model analysis may not lend much insight, but we may have been doing them wrong! // D0TheMath, 12 min

Does agent foundations cover all future ML systems? // Jonas Hallgren, 1 min

ELK And The Problem Of Truthful AI // Scott Alexander, 23 min

# Meta-ethics

Moral strategies at different capability levels // ricraz, 6 min

# Anthropic

Cook’s Critique of Our Earliness Argument // Robin Hanson, 4 min

# Decision theory

Unifying Bargaining Notions (1/2) // Diffractor, 20 min

«Boundaries», Part 1: a key missing concept from utility theory // Andrew_Critch, 8 min

The Reader's Guide to Optimal Monetary Policy // ege-erdil, 17 min

The Most Important Century: The Animation // Writer, 24 min

Unifying Bargaining Notions (2/2) // Diffractor, 29 min

«Boundaries» Sequence (Index Post) // Andrew_Critch, 1 min

My Bitcoin Thesis @2022 - Part 1 // aysajan, 15 min

Utility functions and probabilities are entangled // thomas-kwa, 1 min

Protectionism in One Country: How Industrial Policy Worked in Canada // Davis Kedrosky, 20 min

Index Post for the «Boundaries» Sequence // Andrew_Critch, 1 min

Who Should Be Our “Adults”? // Robin Hanson, 6 min

# Math and cs

Eavesdropping on Aliens: A Data Decoding Challenge // anonymousaisafety, 3 min

The generalized Sierpinski-Mazurkiewicz theorem. // donald-hobson, 2 min

# Books

Your Book Review: Viral // Scott Alexander, 26 min

Not Sickening Enough // Josh Mitteldorf, 19 min

# Community

Opening Session Tips & Advice // CFAR 2017, 17 min

For Better Commenting, Stop Out Loud // AllAmericanBreakfast, 1 min

# Misc

Technocracy and the Space Age // jasoncrawford, 3 min

Drexler’s Nanotech Forecast // PeterMcCluskey, 4 min

Mistakes as agency // pchvykov, 4 min

Relationship between subjective experience and intelligence? // Q Home, 11 min

“Fanatical” Longtermists: Why is Pascal’s Wager wrong? // yitz, 1 min

How Promising is Theoretical Research on Rationality? Seeking Career Advice // Aspirant223, 3 min

Beware Upward Reference Classes // Robin Hanson, 3 min

Hating On Personal Equity // Robin Hanson, 3 min

Beware Sacred Cows // Robin Hanson, 4 min

# Podcasts

166 – Jugaad Ethics // The Bayesian Conspiracy, 102 min

Currents 067: Zak Stein on Ending Nihilistic Design // The Jim Rutt Show, 66 min

YANSS 228 – Jennifer Shahade on how we can learn about passion, obsession, representation, sexism, strategy, tactics, and sideways thinking by exploring the world of professional chess players //

# Videos of the week

Why TeamSeas Doesn't Work: Their Interceptors // Simon Clark, 21 min