rationalnewsletter

Issue #51

Based on neural activity, scientists can predict which of two images a person will choose to imagine up to 11 seconds before they consciously make the choice. // www.nature.com

# Instrumental

A cognitive intervention for wrist pain // rmoehn, 5 min

# Epistemic

Does Reality Drive Straight Lines On Graphs, Or Do Straight Lines On Graphs Drive Reality? // Scott Alexander, 3 min Favorite

Active Curiosity vs Open Curiosity // Unreal, 4 min

Plans are Recursive & Why This is Important // Ruby, 13 min

Understanding information cascades // jacobjacob, Benito, 3 min

How to Understand and Mitigate Risk // mr-hire, 19 min

# Ai

Boeing 737 MAX MCAS as an agent corrigibility failure // shminux, 1 min Favorite

AI Safety Prerequisites Course: Basic abstract representations of computation // RAISE, 1 min

Designing agent incentives to avoid side effects // Vika, TurnTrout, 3 min

A theory of human values // Stuart_Armstrong, 8 min

[Link] OpenAI LP // Alexei, 1 min

# Anthropic

Risk of Mass Human Suffering / Extinction due to Climate Emergency // willfranks, 1 min

Meta-Doomsday Argument: Uncertainty About the Validity of the Probabilistic Prediction of the End of the World // avturchin, 1 min

# Decision theory

Verifying vNM-rationality requires an ontology // jeyoor, 1 min

# Books

Muqaata'a by Fahad Himsi (I.) // Martin Sustrik, 4 min

# Ea

You Have Four Words // Raemon, 1 min

How large is the harm from info-cascades? [Info-cascade series] // jacobjacob, Benito, 1 min

# Community

Question: MIRI Corrigbility Agenda // algon33, 1 min

# Culture war

Privacy // Zvi, 7 min

Blackmailers are privateers in the war on hypocrisy // Benquo, 6 min

Domestic Cozy: 2 // Venkatesh Rao, 2 min

To Oppose Polarization, Tug Sideways // Robin Hanson, 4 min

Predictable Identities: 4 // Jacob Falkovich, 1 min

# Art

Gwern’s AI-Generated Poetry // Scott Alexander, 9 min

# Fun

New York Restaurants I Love: Pizza // Zvi, 10 min

Speculations on Duo Standard // Zvi, 10 min

Puritan Spotting // Scott Alexander, 6 min

# Misc

Humans aren't agents - what then for value learning? // Charlie Steiner, 3 min

Mysteries, identity, and preferences over non-rewards // Stuart_Armstrong, 1 min

What Vibing Feels Like // mr-hire, 2 min

Worlding Raga: 3 — Slouching with God // Venkatesh Rao, 5 min

Ketamine: Now By Prescription // Scott Alexander, 7 min

# Podcasts

#150 — The Map of Misunderstanding // Waking up with Sam Harris, 121 min

Episode 80 – The Reward and Peril of GPT2 // The Bayesian Conspiracy, 132 min

# Rational fiction

[Fiction] IO.SYS // DataPacRat, 27 min

# Videos of the week

OpenAI and the Text Generator Too Dangerous to Exist // TWiT Netcast Network, 8 min