rationalnewsletter

Issue #152

0-shot prompts and Metaprompts: Is GPT-3 learning from the prompts or just accessing existing knowledge? Training samples might not even be necessary. // arxiv.org

# Coronavirus

Covid 2/18: Vaccines Still Work // Zvi, 38 min

Covid: CDC Issues New Guidance on Opening Schools // Zvi, 14 min

How poor is US vaccine response by comparison to other countries? // aphyer, 3 min

How Long Will We Distance? // Robin Hanson, 2 min

# Instrumental

Second Citizenships, Residencies, and/or Temporary Relocation // joshjacobson, 41 min Favorite

The Problem with Giving Advice // Rossin, 2 min

Neck abacus // KatjaGrace, 1 min

Remembering people's name with Anki // ArthurRainbow, 4 min

# Epistemic

The feeling of breaking an Overton window // AnnaSalamon, 1 min

Overconfidence is Deceit // Duncan_Sabien, 13 min

Heuristic: Replace "No Evidence" with "No Reason" // yair-halberstadt, 2 min

On the Nature of Reputation // sustrik, 11 min

How Strong is Our Connection to Truth? // anorangic, 3 min

# Ai

Formal Solution to the Inner Alignment Problem // cocoa, 2 min

Google’s Ethical AI team and AI Safety // magfrump, 7 min

Mathematical Models of Progress? // abramdemski, 3 min

Internet Encyclopedia of Philosophy on Ethics of Artificial Intelligence // Kaj_Sotala, 4 min

Scott and Rohin doublecrux on AI with human models // Raemon, 1 min

# Meta-ethics

Oliver Sipple // KatjaGrace, 2 min

Utilitarian doppelgangers vs. making everything smell like bananas // RobbBB, 3 min

Experientialist Theories of Well-Being // andzuck, 13 min

Notes on Amiability // David_Gross, 7 min

# Longevity

Semaglutide is cool, but no one wants to talk about b. animalis ssp. lactis? // CraigMichael, 2 min

# Anthropic

The Future of Nuclear Arms Control? // Gentzel, 7 min

Are we prepared for Solar Storms? // Self-Embedded Agent, 1 min

Contact with reality // joekc, 25 min

# Decision theory

The Prototypical Negotiation Game // johnswentworth, 5 min

Remember that to value something infinitely is usually to give it a finite dollar value // KatjaGrace, 2 min

The Kelly Criterion in 3D // lsusr, 2 min

Reasonable ways for an average LW retail investor to get upside risk? // BossSleepy, 1 min

The Ultimatum Game // thetwelve, 6 min

# Math and cs

Utility Maximization = Description Length Minimization // johnswentworth, 7 min

Shortcuts With Chained Probabilities // jkaufman, 1 min

Generalised models as a category // Stuart_Armstrong, 5 min

# Books

Chinese History // lsusr, 1 min

Notes on Henrich's "The WEIRDest People in the World" (2020) // MichaelA, 9 min

# Ea

Societal Self-control // aaronb50, 11 min

# Community

Curing Sleep: My Experiences Doing Cowboy Science // jayterwahl, 3 min

This is my 100ᵗʰ post on Less Wrong // lsusr, 1 min

Bedtime reminiscences // Kaj_Sotala, 1 min

Most kinds of noise stress me out a lot, and I have a strong preference for silence. I tend to quickly notice quiet environments. Does this occur to anyone else? Any idea why? // TheTrueSquidward

Other utopias or searching the reference class for the rationality hub project // gogishvilli, 3 min

# Culture war

KAnon // jkaufman, 1 min

What are the most powerful lotuses? // ForensicOceanography, 1 min

# Misc

“PR” is corrosive; “reputation” is not. // AnnaSalamon, 1 min Favorite

How my school gamed the stats // srdjan-miletic, 4 min

The Median is Less than the Average // lsusr, 1 min

Five examples // adamzerner, 6 min

Coffee trucks: a brilliant idea that someone should do? // KatjaGrace, 1 min

# Podcasts

AXRP Episode 4 - Risks from Learned Optimization with Evan Hubinger // DanielFilan, 108 min

# Rational fiction

The Singularity War - Part 3 // lsusr, 3 min

The Singularity War - Part 4 // lsusr, 5 min

What didn’t happen // KatjaGrace, 2 min

# Videos of the week

Bill Gates: The 2021 60 Minutes interview // 60 Minutes, 13 min