rationalnewsletter

Issue #239: Full List

Toxoplasma-infected students are 180% more likely start ventures & infected founders have 8% more revenue // royalsocietypublishing.org

# Instrumental

Elastic Productivity Tools // sberens, 2 min

Developer experience for the motivation // adamzerner, 4 min

Preventing, reversing, and addressing data leakage: some thoughts // VipulNaik, 27 min

Strategy of Inner Conflict // JonathanMoregard, 8 min

How do people do remote research collaborations effectively? // Krieger, 1 min

The Echo Principle // JonathanMoregard, 4 min

# Epistemic

Internal communication framework // rosehadshar, 14 min

Moral contagion heuristic // mvolz, 2 min

Miracles and why not to believe them // mruwnik, 3 min

# Ai

By Default, GPTs Think In Plain Sight // Fabien, 13 min

Results from the interpretability hackathon // esben-kran, 7 min

Current themes in mechanistic interpretability research // Lee_Sharkey, 14 min

Update to Mysteries of mode collapse: text-davinci-002 not RLHF // janus, 2 min

Disagreement with bio anchors that lead to shorter timelines // marius-hobbhahn, 7 min

Engineering Monosemanticity in Toy Models // adam-jermyn, 4 min

Don't design agents which exploit adversarial inputs // TurnTrout, 17 min

A Short Dialogue on the Meaning of Reward Functions // leon-lang, 3 min

Unpacking "Shard Theory" as Hunch, Question, Theory, and Insight // Jacy Reese, 3 min

The limited upside of interpretability // Peter S. Park

The Ground Truth Problem (Or, Why Evaluating Interpretability Methods Is Hard) // jessica-cooper, 2 min

generalized wireheading // carado-1, 2 min

Why don't we have self driving cars yet? // Linda Linsefors, 1 min

LLMs may capture key components of human agency // cat-1, 5 min

AI Ethics != Ai Safety // Dentin, 1 min

Massive Scaling Should be Frowned Upon // harsimony, 6 min

Updates on scaling laws for foundation models from ' Transcending Scaling Laws with 0.1% Extra Compute' // Nick_Greig, 1 min

Distillation of "How Likely Is Deceptive Alignment?" // NickGabs, 11 min

The Disastrously Confident And Inaccurate AI // sharat-jacob-jacob, 16 min

A short critique of Vanessa Kosoy's PreDCA // martinsq, 5 min

Is there any policy for a fair treatment of AIs whose friendliness is in doubt? // nahoj, 1 min

AI Forecasting Research Ideas // Jsevillamol

Is there some reason LLMs haven't seen broader use? // tailcalled, 1 min

Will nanotech/biotech be what leads to AI doom? // tailcalled, 1 min

What is our current best infohazard policy for AGI (safety) research? // Roman Leventov, 1 min

Characterizing Intrinsic Compositionality in Transformers with Tree Projections // ulisse-mini, 1 min

AI Safety Microgrant Initiative // Chris_Leong

The economy as an analogy for advanced AI systems // rosehadshar, 6 min

Is the speed of training large models going to increase significantly in the near future due to Cerebras Andromeda? // asta-vista, 1 min

I (with the help of a few more people) am planning to create an introduction to AI Safety that a smart teenager can understand. What am I missing? // Tapatakt, 1 min

AGIs may value intrinsic rewards more than extrinsic ones // cat-1, 4 min

Cognitive science and failed AI forecasts // ea-1, 2 min

The two conceptions of Active Inference: an intelligence architecture and a theory of agency // Roman Leventov, 4 min

How AI Fails Us: A non-technical view of the Alignment Problem // testingthewaters, 2 min

# Meta-ethics

Deontology and virtue ethics as "effective theories" of consequentialist ethics // Jan_Kulveit

In Defence of Temporal Discounting in Longtermist Ethics // DragonGod

Reflective Consequentialism // adamzerner, 5 min

My Deontology Says Narrow-Mindedness is Always Wrong // LVSN, 1 min

# Decision theory

If Professional Investors Missed This... // jkaufman, 3 min

Two New Newcomb Variants // eva_, 3 min

Why bet Kelly? // AlexMennen, 5 min

logical vs indexical dignity // carado-1, 1 min

Why are profitable companies laying off staff? // yair-halberstadt, 1 min

How does acausal trade work in a deterministic multiverse? // benj, 1 min

Value Created vs. Value Extracted // Sable, 8 min

We need better prediction markets // eigen, 1 min

Is Crypto built on a foundation of ignorance about Information Asymmetry? // TrevorWiesinger, 1 min

Decision making under model ambiguity, moral uncertainty, and other agents with free will? // Jobst Heitzig, 1 min

Sacred Money // Robin Hanson, 2 min

Clean Politician Bonds // Robin Hanson, 2 min

# Math and cs

Some research ideas in forecasting // Jsevillamol

# Books

Review: Bayesian Statistics the Fun Way by Will Kurt // matto, 2 min

# Ea

Noting an unsubstantiated communal belief about the FTX disaster // yitz

Estimating the probability that FTX Future Fund grant money gets clawed back // spencerg

Introducing The Logical Foundation, A Plan to End Poverty With Guaranteed Income // michael-simm

Discussion: Was SBF a naive utilitarian, or a sociopath? // NicholasKross

X-risk Mitigation Does Actually Require Longtermism // DragonGod

"You could make a religion out of this!" // TheAspiringHumanist, 4 min

# Community

The Alignment Community Is Culturally Broken // sudo, 2 min

Announcing the Progress Forum // jasoncrawford, 1 min

Kelsey Piper's recent interview of SBF // agucova

SBF x LoL // NicholasKross

# Fun

Improving Emergency Vehicle Utilization // jkaufman, 1 min

# Misc

Sadly, FTX // Zvi, 56 min

The FTX Saga - Simplified // jorge-velez, 9 min

Make the Drought Evaporate! // AnthonyRepetto, 4 min

When should we be surprised that an invention took “so long”? // jasoncrawford, 5 min

The Futility of Status and Signalling // Ape in the coat, 3 min

The Psychopharmacology Of The FTX Crash // Scott Alexander, 25 min

More Or Less Sacred // Robin Hanson, 4 min

# Podcasts

175 – FTX + EA, and Personal Finance // The Bayesian Conspiracy, 105 min

# Rational fiction

Lying to Save Humanity // cebsuv, 1 min

# Videos of the week

My Videos // Robin Hanson, 2 min