Philosophical Multicore

I sleep less when I exercise more

Donation Timing Under Uncertainty About AI Timelines

Thoughts on investing for transformative AI

I'm extremely worried that superintelligent AI will kill everyone

I was wrong: concentrated factor portfolios don't have alpha

Can AI make advancements in moral philosophy by writing proofs?

Pausing AI Is the Best Answer to Post-Alignment Problems

By Strong Default, ASI Will End Liberal Democracy

The Future Will Be Weirder Than That

Which is better for sentient beings: an "ethical" AI or a corrigible AI?

The resource-constraints argument for why aligned ASI wouldn't be bad for animals

List of ideas for improving animal welfare in light of transformative AI

I used to think aligned ASI would be good for all sentient beings; now I don't know what to think

Cost-effectiveness model for AI alignment-to-animals vs. alignment-in-general

Which types of AI alignment research are most likely to be good for all sentient beings?

Worlds where we solve AI alignment on purpose don't look like the world we live in

Value Investing in the Age of AGI

The Structural Return Argument Against Value Investing

Contra "Time Series Momentum: Is It There?"

If AI alignment is only as hard as building the steam engine, then we likely still die

I'm wary of increasing government expertise on AI

Rest in Peace Commento; Long Live Comentario

I need the Writing Style Guide people to figure out how to put a smiley face inside parentheses

I did Inkhaven

How do I know if I'm dreaming?

Prioritizing your objectives is better than grazing past them accidentally

My Carlin-esque list of pet peeves

Not being awkward is NP-hard

Some little things I do to make life easier

Kid me was bad at Magic: The Gathering

Gaming keyboards are not good for gaming

Belief in expert mistakes

TV is better when you trust the writers

I like reborrowed words

Wartime ethics is weird

Alignment Bootstrapping Is Dangerous

Magic: The Gathering Arena decklists for people on a budget

How to Fix Quidditch

I don't like having goals

Some Curiosity Stoppers I've Heard

Where I Am Donating in 2025

An unnecessarily long analysis of one line from The Princess Bride

We won't solve non-alignment problems by doing research

Do Disruptive or Violent Protests Work?

Why would God have a gender?

Not-Discovered-Here Syndrome

Knowing Whether AI Alignment Is a One-Shot Problem Is a One-Shot Problem

What If Ghosts Were Real?

In Defense of the NCIS Two-People-One-Keyboard Scene

Epistemic Spot Check: Expected Value of Donating to Alex Bores's Congressional Campaign

Ideas Too Short for Essays, Part 2

Are Groot and Baby Groot the Same Person?

A Thesis Regarding The Impossibility Of Giving Accurate Time Estimates, Presented As An Experiment On Form In Which The Essay Solely Consists Of A Title; In Which The Thesis States That, If Task Times Follow A Pareto Distribution (With The Right Parameters), Then An Unknown Task Takes Infinite Time In Expectation; And Therefore, In The General Case, You Cannot Provide An Accurate Time Estimate Because Any Finite Estimate Provided Will Not Capture The Expected Value; And, More Precisely, Every Estimate Will Be An Underestimate, Because Every Number Is Smaller Than Infinity; And This Matches With The General Observation That, When People Estimate Task Times, They Usually Underestimate The True Time; However, In Opposition To This Thesis Are At Least Two Observations; First, That Even If Tasks Take Infinite Time In Expectation, The Median Task Time Is Finite, And An Infinite-Expected-Value Task-Time Distribution Does Not Preclude The Possibility That Time Estimates Can Overestimate As Often As They Underestimate, But People Fail To Do This; Second, That Certain Known Biases That Result In People Underestimating The Difficulty Of Tasks, Such As Envisioning The Best-Case Scenario Rather Than The Average Case; However, In Defense Of The Original Thesis, Optimism Bias And The Pareto-Distributed Problem Space May Be Two Perspectives On The Same Phenomenon; But Even If We Reconcile The Second Concern With The Thesis, We Are Still Left With The First Concern, In Which An Unbiased Estimate Of The Median Time Should Still Be Possible, But People Are Overly Optimistic About Median Task Times; Thus, Ultimately Concluding That The Thesis Of This Essay--Or, More Accurately, The Thesis Of This Title--Is A Faulty Explanation Of People's General Inability To Provide Accurate Time Estimates; Then Following Up This Thesis With The Additional Observation That We Can Model Tasks As Turing Machines; And The Halting Problem States That It Is Impossible In General To Say Whether A Turing Machine Will Halt, And As A Corollary, It Is Impossible In General To Predict How Long A Turing Machine Will Run For Even If It Does Halt; So Perhaps The Halting Problem Means That We Cannot Make Accurate Time Estimates In General; However, It Is Not Clear That The Sorts Of Tasks That Human Beings Estimate Are Sufficiently General For This Concern To Apply, And Indeed It Seems Not To Apply Because Some Subset Of People Do In Fact Succeed At Making Unbiased Time Estimates In At Least Some Situations, At Least Where 'Unbiased' Is Defined Relative To The Median Rather Than The Mean; It Is Difficult To Say In Which Real-Life Situations The Halting Problem Is Relevant Because It Is Not Feasible To Construct A Formal Mathematical Proof For Realistic Real-Life Situations Because This Would Require Creating A Sophisticated Model In Which The State Of The Universe Is Translated To A Turing Machine, Which Would Be An Extremely Large Turing Machine And Probably Not Feasible To Reason About; Leading To The Conclusion That This Essay's Speculation Led Nowhere

Upside Volatility Is Bad

Writing Your Representatives: A Worthwhile and Neglected Intervention

Things I Learned from College

Cash Back

How Can I Not Know Whether I'm Having a Good Experience?

Do Small Protests Work?

My Third Caffeine Self-Experiment

Things I've Become More Confident About

Will Welfareans Get to Experience the Future?

The Next-Gen LLM Might Pose an Existential Threat

Mechanisms Rule Hypotheses Out, But Not In

How Much Does It Cost to Offset an LLM Subscription?

I Made an Emacs Extension That Displays Magic: the Gathering Card Tooltips

AI Safety Landscape and Strategic Gaps

Healthy Cooking Tips from a Lazy Person

Doctor Strange Didn't See Only One Victory out of 14,000,605 Futures

Is It So Much to Ask for a Nice Reliable Aggregated X-Risk Forecast?

Annual Subscription Discounts Usually Aren't Worth It

LLMs Might Already Be Conscious

In Which I Defend Fruit's Honor

Updates Digest: Inaugural Edition

Against Ergodicity Economics

Let's Take a Moment to Marvel At How Bad the Original USDA Food Pyramid Was

Can You Maintain Lean Mass in a Calorie Deficit?

Why Would AI Companies Use Human-Level AI to Do Alignment Research?

Do Protests Work? A Critical Review

The Triple-Interaction-Effects Argument

You Can Now Read My Reading Notes

There Are Three Kinds of "No Evidence"

Return Stacked Funds: A New Way to Get Leverage

I Was Probably Wrong About HIIT and VO2max

Retroactive If-Then Commitments

The 7 Best High-Protein Breakfast Cereals

Charity Cost-Effectiveness Really Does Follow a Power Law

"You Can't Calculate the Expected Utility of a Communist Revolution"

Thoughts on My Donation Process

Where I Am Donating in 2024

Subjects in Pysch Studies Are More Rational Than Psychologists

My Submission for Worst Argument In The World

Outlive: A Critical Review

I Have Whatever the Opposite of a Placebo Effect Is

Protein Quality (DIAAS) Calculator

Just Because a Number Is a Rounding Error Doesn't Mean It's Not Important

A 401(k) Sometimes Isn't Worth It

Continuing My Caffeine Self-Experiment

Some Things I've Changed My Mind On

Notes on Eat, Drink, and Be Healthy

What's the Healthiest Body Composition?

What's the Healthiest BMI?

Caffeine Cycling Self-Experiment

How Well Did Scott Alexander's List of Social Science Findings Hold Up?

Explicit Bayesian Reasoning: Don't Give Up So Easily

Does Caffeine Stop Working?

Avoiding Caffeine Tolerance

Utilitarianism Isn't About Doing Bad Things for the Greater Good. It's About Doing the Most Good

The United States Is Weird

Should Patient Philanthropists Invest Differently?

Two Types of Scientific Misconceptions You Can Easily Disprove

Most Theories Can't Explain Why Game of Thrones Went Downhill

Philanthropists Probably Shouldn't Mission-Hedge AI Progress

The Review Service Vortex of Death

Index Funds That Vote Are Active, Not Passive

A Preliminary Model of Mission-Correlated Investing

How I Estimate Future Investment Returns

Can Good Writing Be Taught?

Existential Risk Reduction Is Naive (And That's a Good Thing)

Altruistic Investors Care About Other Altruists' Portfolios

Should Earners-to-Give Work at Startups Instead of Big Companies?

Obvious Investing Facts

Future Funding/Talent/Capacity Constraints Matter, Too

Low-Hanging (Monetary) Fruit for Wealthy EAs

Do I Read My Own Citations?

Summaries Are Important

Argumentative Tactics I Would Like to Stop Seeing

My Experience Trying to Force Myself to Do Deep Work

Mission Hedgers Want to Hedge Quantity, Not Price

How Do AI Timelines Affect Giving Now vs. Later?

Metaculus Questions Suggest Money Will Do More Good in the Future

Reverse-Engineering the Philanthropic Discount Rate

How Can We Increase the Frequency of Rare Insights?

Investment Strategies for Donor-Advised Funds

A Comparison of Donor-Advised Fund Providers

More Evidence on Concentrated Stock Selection

The True Cost of Leveraged ETFs

Investors Can Simulate Leverage via Concentrated Stock Selection

Asset Allocation and Leverage for Altruists with Constraints

Uncorrelated Investments for Altruists

If Causes Differ Astronomically in Cost-Effectiveness, Then Personal Fit In Career Choice Is Unimportant

Donor-Advised Funds vs. Taxable Accounts for Patient Donors

The Risk of Concentrating Wealth in a Single Asset

"Disappointing Futures" Might Be As Important As Existential Risks

Mission Hedging via Momentum Investing

Giving Now vs. Later for Existential Risk: An Initial Approach

Should We Prioritize Long-Term Existential Risk?

The Importance of Unknown Existential Risks

Estimating the Philanthropic Discount Rate

Do Theoretical Models Accurately Predict Optimal Leverage?

How Much Leverage Should Altruists Use?

Correction on Giving Now vs. Later

Are All Actions Impermissible Under Kantian Deontology?

New Page: Convert Credences into a Bet

What Are the Best TV Shows (According to IMDb Episode Ratings)?

High School Science Experiments

How Can Donors Incentivize Good Predictions on Important but Unpopular Topics?

Should Global Poverty Donors Give Now or Later?

Why Do Small Donors Give Now, But Large Donors Give Later?

Where Some People Donated in 2017

Newcomb's Problem and Efficient Markets

All Expected Utility Distributions Have One of Two Big Problems

New Comment System

An Idea for How Large Donors Can Support Small Charities