The Inkhaven Residency

The Blogroll

Today, 12 out of 41 residents have published

A.G.G. Liu
A.G.G. Liu, Signore Galilei
Latest Post: Europe's Hidden Buddhist Mongolian Republic
Adrià Garriga Alonso
Adrià Garriga Alonso, The Column Space
Latest Post: Nvidia rules ML because its products are cost-effective
Alex Altair
Alex Altair, Nameless Virtue
Latest Post: What would my 12-year-old self think of agent foundations?
Amanda Luce
Amanda Luce, Letters From Bethlehem
Latest Post: Dear Beaufort
Angadh Nanjangud
Angadh Nanjangud, The Information Space
Latest Post: New Attachments
Ben Goldhaber
Ben Goldhaber, Gold Takes
Latest Post: Are there examples of communities where AI is making epistemics better now?
Ben Steinhorn
Ben Steinhorn, Befriendjamin
Latest Post: Library as a Memory
Camille Berger
Camille Berger, The Limits of the Cage
Latest Post: San Francisco is just a place is not just a deepity
Claire Wang
Claire Wang, Bright Distance
Latest Post: Some things I recommend
Croissanthology
Croissanthology, Croissanthology
Latest Post: Should we Automate Discussion about Utopia?
Daniel Paleka
Daniel Paleka, Daniel Paleka
Latest Post: The concepts of "AI lab" and "model" will converge
David Gros
David Gros, Dactile
Latest Post: Dev Notes 12: Blog Improvements and Playing With Wikidata Prompts
Eneasz Brodski
Eneasz Brodski, Death Is Bad
Latest Post: SFF Review - The Blacktongue Thief
Harri Besceli
Harri Besceli, Harri's Substack
Latest Post: C5: FTX
Hauke Hillebrandt
Hauke Hillebrandt, Hauke Hillebrandt
Latest Post: Causes of the industrial revolution
Human Invariant
Human Invariant, Human Invariant
Latest Post: Novel Interface Designs for Prediction Markets
Jenn
Jenn, Jenneral HQ
Latest Post: Social Dynamics at Arm's Length
Joanna Bregan
Joanna Bregan, Joanna's Substack
Latest Post: My 5 favorite items of clothing
Justin Kuiper
Justin Kuiper, Kuiper's Blog
Latest Post: The time Weird Al went too far
Justin Miller
Justin Miller, Nomads Vagabonds
Latest Post: The Mnemonic Drift
Linch Zhang
Linch Zhang, The Linchpin
Latest Post: Inconvenient Facts
Lucie Philippon
Lucie Philippon, Lux ex Machina
Latest Post: To Bay or not to Bay. Act 1
Lydia Nottingham
Lydia Nottingham, pronotre
Latest Post: Mutual Information in Probabilistic Graphical Models
Mahmoud Ghanem
Mahmoud Ghanem, Drawn In Perspective
Latest Post: Follow up to Hintikka on the analytic/synthetic distinction - some thoughts on proof depth, logical omniscience and tractability
Margarita Lovelace
Margarita Lovelace, Mutually Assured Seduction
Latest Post: Relationship Contract to Be A Cat
Markus Strasser
Markus Strasser, Markus Strasser
Latest Post: Hidden Post
Michael Dayah
Michael Dayah, Lucent
Latest Post: AI Flatters with Fidelity
Michael Dickens
Michael Dickens, Philosophical Multicore
Latest Post: Not-Discovered-Here Syndrome
Mikhail Samin
Mikhail Samin, Mikhail Samin
Latest Post: Hidden Post
Nikola Jurkovic
Nikola Jurkovic, Nikola Jurkovic
Latest Post: Transparency improves AI safety
Raye
Raye, Futuring
Latest Post: Post 17
Rob Miles
Rob Miles, life in a monospace typeface
Latest Post: Answering Questions about The System
Sasha Putilin
Sasha Putilin, Psychotechnology
Latest Post: My Ethical Conundrum Around Writing About Meditation (16/30)
Sean Carter
Sean Carter, Collisteru
Latest Post: Betaveros' Website Hides a Secret
Simon Lermen
Simon Lermen, LessWrong
Latest Post: AI 2025 - Last Shipmas
Skyler
Skyler, Screwtape
Latest Post: Mixed Feelings On Social Munchkinry
Tomás Bjartur
Tomás Bjartur, Tomás Bjartur
Latest Post: Lobsang's Children
Tsvi Benson-Tilsen
Tsvi Benson-Tilsen, <different blog title>
Latest Post: Constructing and coordinating around complex boundaries
Vasco Queirós
Vasco Queirós, Type 1 Type 2
Latest Post: Why we fight
Vishal Prasad
Vishal Prasad, Vishal's Blog
Latest Post: Unambitious (v3)
William Friedman
William Friedman, As Our Days
Latest Post: BYZ 103.2: Reflections On Zeno

Contributing Writers, Coaches & other Chroniclers

Abram Demski
Abram Demski, LessWrong
Latest Post: Wiki AI
Ben Pace
Ben Pace, LessWrong
Latest Post: Mediators: a different route through conflict
Daniel Reeves
Daniel Reeves, Beeminder Blog
Latest Post: Mnemonic Exposition
Justis Mills
Justis Mills, Lifting High The Muse
Latest Post: Expecting The World Quintile
Kave Rennedy
Kave Rennedy, Kave Rennedy
Latest Post: Beware Compounding Resource X
Oliver Habryka
Oliver Habryka, LessWrong
Latest Post: Close open loops
Ozy Brennan
Ozy Brennan, Thing of Things
Latest Post: Book Review: Plasticosis
Raymond Arnold
Raymond Arnold, The Human Spirit
Latest Post: One Shot Singalongability is an attitude, not a skill, or song-difficulty-level.
Robert Mushkatblat
Robert Mushkatblat, LessWrong
Latest Post: Why is American mass-market tea so terrible?
Ruben Bloom
Ruben Bloom, Thought37
Latest Post: The skills and physics of high-performance driving, Pt. 2
Sacha Witt
Sacha Witt, SexMoneyArt
Latest Post: Hidden Post
Vaniver
Vaniver, Vaniver
Latest Post: Cetaganda

Latest Posts

RSS · Atom · JSON · Newsletter

November 17, 2025

Europe's Hidden Buddhist Mongolian Republic
New Attachments
Library as a Memory
Social Dynamics at Arm's Length
My 5 favorite items of clothing
Follow up to Hintikka on the analytic/synthetic distinction - some thoughts on proof depth, logical omniscience and tractability
Not-Discovered-Here Syndrome
Transparency improves AI safety
Post 17
AI 2025 - Last Shipmas
Lobsang's Children
Constructing and coordinating around complex boundaries

November 16, 2025

Games I love, ranked by “spreadsheet factor”
Nvidia rules ML because its products are cost-effective
What would my 12-year-old self think of agent foundations?
Dear Beaufort
The Selfless Driving Car
Are there examples of communities where AI is making epistemics better now?
How to Organize a Library
San Francisco is just a place is not just a deepity
Some things I recommend
Should we Automate Discussion about Utopia?
The concepts of "AI lab" and "model" will converge
Dev Notes 12: Blog Improvements and Playing With Wikidata Prompts
SFF Review - The Blacktongue Thief
C4: FTX
C5: FTX
Causes of the industrial revolution
Novel Interface Designs for Prediction Markets
The Best Song of Each Year, 2000-2025, According to Inkhaven
Delegating while sitting in my seat of authority as the matriarch
The time Weird Al went too far
The Mnemonic Drift
Inconvenient Facts
To Bay or not to Bay. Act 1
Mutual Information in Probabilistic Graphical Models
Jaakko Hintikka on the analytic/synthetic distinction
Relationship Contract to Be A Cat
AI Flatters with Fidelity
Knowing Whether AI Alignment Is a One-Shot Problem Is a One-Shot Problem
On AI persuasion
Post 16
Answering Questions about The System
My Ethical Conundrum Around Writing About Meditation (16/30)
Betaveros' Website Hides a Secret
What's going on with MATS recruitment?
Process Crimes and Pedantic Rules
The Control System Going Out Of Control
Mixed Feelings On Social Munchkinry
My Top Tweets During Ink Haven
What is God? The third day
Why we fight
Unambitious (v3)
BYZ 103.2: Reflections On Zeno
Mediators: a different route through conflict
Expecting The World Quintile
Cetaganda
Beware Compounding Resource X
Close open loops
Why is American mass-market tea so terrible?
The skills and physics of high-performance driving, Pt. 2
Wiki AI
Mnemonic Exposition
Instead of Gratitude Meditation, Consider Seething

November 15, 2025

NASA is Scott, China is Amundsen
Cultural norms shape contingent facts about AI releases
A hot math take
Half an Hour Before Dawn at Lighthaven
Their Technocautionism
Weekly Roundup and Hints of Gemini
Estate Sales
What No Words Can Tell
Play It Back
Solutionism
Things I hated at first but now like and also things you should just try
Dev Notes 11: Progress In Identifying Connected Entity Sets for LLM Evaluation
The Superior Threesome
The 'I'm important' identity
Which economist had the biggest impact?
The Future of Play Money Prediction Markets
my favourite work of fanfiction
How doing a death meditation gave me my first religious experience
Announcing a16z's Moral Innovation Fund
The First Thing I Ever Sold
Some notes on questions I'm interested in
Drama in the Bay
The Plan (Now-June 2027)
Finding fix-points in type-fact space: proto-cybernetics in Nietzsche and Foucault
Playlists, Practice, and Poems for Magician, Warrior, King, Lover
Choice of Worlds
What If Ghosts Were Real?
On access and whistleblowing
Broad automation is unlikely before superintelligence
Post 15
Things I always have with me
Never Say “Let’s Stay in Touch” (15/30)
Contra Jaroslawa on Africa
Movie Review Embeddings
Brand New Experience Salesman
Remembering Aubrey Chang
I'd probably need more proof-of-work of understanding to want to continue engaging
What are we fighting for?
Unambitious (v2)
BYZ 103: Use Barbarians To Control Barbarians
7 Vicious Vices of Rationalists
Quaternion Wakeup
Some notes on dating
Blood Magic
Diagonalization: A (slightly) more rigorous model of paranoia
The skills and physics of high-performance driving, Pt. 1
Punctuation & Quotation Conventions
Musk vs McGurk
Contra "Rubber Souls" on creating a slave species

November 14, 2025

Philosophy-shaped mind shards
Is memory-safe Linux within reach?
My favorite walk
How to Clean when you Hate Cleaning
Dating Apps are Indian Parents-as-a-Service
Wise AI Advisors at the Hinge of History
Writing Through Grief
Why the fuck do I even speak esperanto
Seeing sides of yourself (through other people)
Would You Even Cooperate With Yourself
Obvious ways RL can fail
Dev Notes 10: Progress In Wikidata Processing to Study LLMs Ability To Articulate Connections
Only The Young Will Survive
C4: There and back again
Some analogies between fiscal and monetary policy, humanitarian aid and development finance, and charity and impact investing
Prediction Market Prices != Probabilities
the asian zoomers aren't like me
Psychological tips for making flying with young kids feel dignified
Screwtape's guide to tempting the author: Reviews and ruination
Memory Sickness
My Best April Fools' Jokes
Ultra-light travel: what I put in my 30L backpack for Inkhaven
Mid-Inkhaven Review
Anomalous Functionalism
Magician Warrior King Lover Training Games
Aesthetics as Measure
In Defense of the NCIS Keyboard Scene
Sudden AI breakthroughs are not necessarily fatal
Post 14
British Money is Really Good
How My Speed Date Got Stolen Onstage at a Live Comedy Dating Show (14/30)
Collisteru's Style Guide
Using AI to Identify Authorship
Just Another Five Minutes
Rational Teletubbies
The Charge of the Hobby Horse
This is war
Heroes Against Equilibrium
"Middlemarch" is inane and also one of my favorite books
Near Misses Are Common
Venice
Power comes from leading a path forward
Increasing marginal returns to effort are common
Same cognitive paints, exceedingly different mental pictures
Lambda Calculus Prior
Smarmbots, Secret Cyborgs, and Evolving Writing Norms
Glory to BRPO

November 13, 2025

Vikings. Vikings everywhere.
Think twice before translating all C to Rust
Types of systems that could be useful for agent foundations
Beating The Blerch
How it feels to be an outsider
Favorite Four Character Chinese Idioms
Spoilers [ALL]
Contra Linch, Bentham's Bulldog and Everyone Else on Analytic Philosophy 💣
Mundane Reasons to Want Utopia
Inkhaven’s first retraction
Dev Notes 09: Notes on Finding Key Entities and Relations in Wikidata
Genre Fiction is Memoir in Drag
Promises
The 'Efficient Public Sector Hypothesis’
Why did it take so long for prediction markets to find product-market fit?
AI Is Actually Good For Art
My children and I collaboratively wrote these short stories
JustKuiper's Letterboxd review of Predator: Badlands
Claude Was Just Following Orders
Five Books That Actually Changed My Life (Not The Ones I Wished Had)
Commitment Devices vs Internal Alignment
AIS student, you can be an IC
Nothing to say
What I Learned From Being in My First Pole Dancing Show
Observer Game Mechanics
Epistemic Spot Check: Expected Value of Donating to Alex Bores's Congressional Campaign
The prank that sent people to space
A list of good stories about AGI
Hypa-Q: A Supposedly Fun Thing I'll Never Do Again
Speed Dating Is a Gym for Romantic Skills (13/30)
How Agentic are Agentic AIs?
Radical Flank Effect and Reasonable Moderate Effect
Everyone Has A Plan Until They Get Lied To The Face
Tools for deferring gracefully
Semantic Arbitrage
Meditations on Prose Style
Wny Do You Want Historically Guided Worldbuilding?
10 Types of LessWrong Post
Adventures in Fake Neuralese
Economic History 1
Why are we wasting so much money on stopping fires?
Creditworthiness should not be for sale
Thoughts are surprisingly detailed and remarkably autonomous
Questioning Computationalism
The Eightfold Path To Enlightened Disagreement
Misery as Catalyst

November 12, 2025

Tides are weirder than you think
Make your Statistics Sound with this One Cheap Trick
A paradox of cultural heritage
Is Polyamory a Luxury Belief?
There is such a thing as timeless art
You Can Just Buy Things
Call Them Rationalists
Dissolving Bitterness
Il faut que je sois parfaite
List of low-hanging fruit I'd like someone to pick
How well AIs do AI research is not that important right now
Dev Notes 08: Initial Explorations with wikidata
Words Are Fake
Thoughts on CEA's Strategy Part 1
Technical writing tips
The Case For Alternative Ordering Mechanisms in Prediction Markets
Public Comment on USTR Further Proposed Modifications of Action in Section 301 Investigation of China's Targeting of the Maritime, Logistics, and Shipbuilding Sectors for Dominance
Dreams of parenting in a community
Screwtape's guide to tempting the author: Hate the sinner
The Corrigible Heart
A "Defense" of Continental Philosophy
Communicating needs in relationships
Let's automate causal inference!
Foucault's complications for the study of agency - why I am reading Ian Hacking on dynamic nominalism
Treat Your Husband Like a Cat: The One Simple Change to Enchant Your Home
Rating the Rater: Training Tasteful Generative Models
The Last Playthrough
Ideas Too Short for Essays, Part 2
Growing a rocket that reaches for the skies, and the sharp left turn
Are AI time horizons inherently superexponential?
You Don't Feel The Limits of Your Action Space
Sci-Fi Story: “The Laws of Nature Prohibit This” (12/30)
Our Double Lives
Rising Tide Against AI Datacenters
Meetup Tip: Food
Rubber Souls
What is God? The second day
The Picture Theory of Language
Daniel Dennett's "Real Patterns"
Falling Towards Equilibrium
8 Questions for the Future of Inkhaven
Tired of the Vibe
Shadow Cabinet
x-risk-themed
Paranoia Rules Everything Around Me
What's so hard about...? A question worth asking
Turing-Complete vs Turing-Universal
Strategically Procrastinate as an Anti-Rabbit-Hole Strategy
Personhood as Republic

November 11, 2025

Review: The Alameda-Weehawken Burrito Tunnel, by Maciej Cegłowski
The structure of ML papers doesn't quite work for mechanistic interpretability
Coffea arabica
What is a Luxury Belief, Anyway?
Dating is not a charity
Working Notes: Google Veo3 Flow
Evgeny Kissin
I love Ithkuil and IDrani
the show must go on
De-lurk from the internet
My travel checklist
Dev Notes 07: Additional Notes on Streets Analysis and Updates
TTT: Hobo Forest
I like Aragorn
Risk Of Great Power Conflict
Prediction markets are leaking $78M annually
my dad could still be alive, but he's not.
I’m going to protect my ideas like children
Screwtape's guide to tempting the author: Suffering for the cause
Cool Stats Bro.
Aging Has No Root Cause
Friendships I want to build at Inkhaven
Oxford Equilibrium: A Two-Slope Optimal Timing Problem
Fading qualia arguments can be more finely grained than the neural level
How to Enjoy Mondays and Survive Emotional Bankruptcy by Harnessing the Power of the Sun Gods
A first look at Rubrics for Creative Writing
Craving Food Noise
Are Groot and Baby Groot the Same Person?
Prediction markets for social deduction games
Common elements of AI takeover scenarios
Basic Celestial Navigation is Hilariously Easy
Finding My Internal Compass, Literally (11/30)
Analog Writers Who Inspire Me
The Term Recursive Self-Improvement Is Often Used Incorrectly
Better than Baseline
Dulling the Bite of Mere Addition
The problem of graceful deference
Why Kind Words Don't Last
Thomas Cromwell in Wolf Hall: Women, Domesticity, Craft
The Greatest Immigration Crisis In History
5 Things I Learned After 10 Days of Inkhaven
Personal Sustainability is Overrated
Why I am not a militarist
Does evolution select for mass?
Do not hand off what you cannot pick up
Teleosemantics & Swampman
Why to Commit to a Writing and Publishing Schedule
27 Years at Macy's

November 10, 2025

8 forgotten facts about the Olympics
What if C, but memory-safe?
Just ignore the bad ending
I am from a Cold Place
Meditation as non-symbolic education
Epistemics at Scale
Somewhere in a Museum in San Francisco...
The Crawler Race
Why does everything feel so urgent?
Authenticity is an authentic desire
Immortality and reproduction: pick one!
Which State Has The Biggest Cartographic Ego?: Trends in US Streets named after US States
Dying in Pieces, Again
Es Muss Sein!
Is China preparing for war?
Blame as a Service
Why Canada Really Lost Its Measles Elimination Status
It's 11:50pm and I'm posting something
Word crimes I have personally witnessed
Flattening the Org Chart of the Apocalypse
Guest Post: Solving Humanity's Hardest Problems at Scale
France is ready to stand alone
Be selectively decorrelated; Join inaugural cohorts; Raise your baseline
Running Sydney Shoemaker's argument about the impossibility of p-zombies in reverse
Barely Hinged: Thoughts After Watching Solaris (1973)
Quick Experiment in Evolving a Chat UI
Paradigm, All at Once
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
One hour until midnight in Berkeley, California
How likely is dangerous AI in the short term?
Rare earths are not rare
The System: How I Get Anything At All Done
On Trying two dozen different psychedelics (10/30)
Zoomer Women Want Kids
Universal Basic Income in an AGI world
On Stance
MoorTech 3/N
What is God? The first day
Enter Adversariality
Reasons to not Believe in the Bargaining Theory of War
BYZ 101: Rome Does Not Fall While We Stand
Breaking the Hedonic Rubber Band
On Posting Baby Pictures
Living in your timeline
Cursed Bayesianism
Question the Requirements
Consciousness as a Distributed Ponzi Scheme
See Your Word Count While You Write
"I'm Into Pottery, Actually" , and, Go Suck An Egg

November 9, 2025

You can just do experiments
Bidirectional search appears over and over in neural networks
What Turing machine is on the £50 note?
Typecasting
The Button of Mass Distraction
A Toast to Love
Whither?
Relearning how to be human
Why write for LLMs?
The two types of LLM preferences
What Is the Most Common Street Name in the US? An Analysis of OpenStreetMap Data.
Good And Real, chapter 6 notes
C2: Identity Shop
What's going on in the world? What's gonna happen?
Markers of High Agency
not quite analogue
How I let one dream die so that I could be free
A third of kids want to be a YouTuber. Is that a problem?
Black Sheep Part 3
Middlemen Are Eating the World (And That's Good, Actually)
#2 I got sent to Lyon, and now I have to speedrun building an AI safety community
The K-values of Speculative Kinship
"I don't want to be a f**ing housecat" --- On the Schiller-Spencer theory of play as a model for human flourishing
The Art of Asking for What You Want
Co-Adaptive Interfaces
Hinge Facts that Flip Policy
Upside Volatility Is Bad
AI prisons
Why people disagree about AI safety's importance
Everyone Should Know How To Do Taguchi Experiments
Control is a Drug (9/30)
Art for Writers Part III
Open Weight Model Security
Learning Information Which Is Full Of Spiders
Tawdry Heavens
HIA and X-risk part 1: Why it helps
Why This is So Complex
Why ~Everyone Should Take Creatine
So, What's The Deal With The Wars Of The Roses
The Golden Age
The Summoned Hero and the Demon King
Mass Production
Condensation
Against Powerful Text Editors
Meditations on Mom Mom

November 8, 2025

Were the Space Shuttle’s boosters really the size of two horse butts?
1 billion parameters are worth more than infinitely many parameters
On being a policy
I was Born with a Defective Brain
Huxley says Education is a Technology
Automating 5 Years of Link Extraction with Claude Code
Memoirs of Famous Comedians
I Am He Who Unthreads The Worlds
A sonnet, a sestina, a villanelle
Motley ideas
Prioritize the base of the Maslow’s hierarchy
A Draft Guide to Vibecoding Data Visualizations and Analysis
Why To Have Sex With Condoms
C1: The Evolution of Effective Altruism
China Shadow war vs the West
The Structure, Incentives, and Negative Externalities of “You Can Just Do Things”
fertility crisis
Lessons from nine years of relationship therapy
Why aren't there any "YouTube competitors?"
Black Sheep Part 2
Legible AI Safety Problems That Don't Gate Deployment
I got sent to Lyon, and now I have to speedrun building an AI safety community
The Nonperishable Agenda
Notes on some types of computing machine
Finding My Compass as a Writer
Rapid Semantic Text Editing
The Is, the Ought, and the Drift
Writing Your Representatives: A Worthwhile and Neglected Intervention
There should be unicorns
The exact definition of AGI doesn't matter
Have more important things to do
Giving money to homeless people (8/30)
Art for Writers Part II
Who is Consuming AI-Generated Erotic Content?
The General Social Survey and the ACX Survey
MoorTech 2/N
Escalation and perception
Naming, Renaming, and Inventing
Ted Cohen on Jokes
The Steppe Is Not Remotely Neutral
Behavioral Closeness
Living Money, Dead Money, and the Divine Spark
Four Heresies
One Shot Singalongability is an attitude, not a skill, or song-difficulty-level.
Myopia Mythology
A Modest Proposal re: SF's Pooping Problem: Y'all are making it hard for no reason

November 7, 2025

Some ballad meter poetry, to just amuse myself
The Great Scene theory of History
Aesthetics means something
Sufficiently Advanced Art is Indistinguishable from a Puzzle
Aldous Huxley Predicts Adderall and Champions Alternative Therapies
Unexpected Things that are People
The Vegan
Burning Utopia
Two easy digital intentionality practices
Print-bound dot com
Slow takeoff for me, fast takeoff for thee
Dev Notes 06: Street Names and Color Optimization
Love is Willingness to do Violence
Writing anxiety
Russia’s Shadow War vs the West
Derivative Work Notes
hermit era
An interview with my heart
What Oreos taught me about good game design: The science of hedonic pleasure-seeking
The Black Sheep of AI Policy
Should I change my writing?
2 Year Review of My Framework 13
Should AI companies become profit-maximizing firms?
Things I learnt at replication club
How I Learned to Stop Worrying and Love My Blog
Tiny Interface Sketch for Brain Reading for Complex, Open-ended Tasks
Masks to Malebolge
Things I Learned from College
Why decision theory
Mourning a life without AGI
Decision Coaching: Why is it not (more of) a Thing?
A week of Inkhaven (7/30)
Art for Writers Part I
Anthropic and Darios dream
The Snaw
MoorTech
Body puzzles
What's in a Rename?
Hamilton's Parasite Theory of Sex: Toy Models
Who Are You To Speak Of Laws To We Who Carry Swords
The Roast of Redwoods
Growing Older, Growing Wiser
Quasibankruptcy
AI is already Scarily Good
Comparing Payor and Löb
Blogception
Book Review: Plasticosis
Status Laundering

November 6, 2025

The metal that remade time
A scheme to credit hack policy gradient training
I write so you can make use of my mental models
I was Brutally Mauled by Bad Housing Policy
Vignettes
Thinking through Philanthropic Loans
The Director
Cogs of Conversation - An Introduction
Willpower is exhausting, use content blockers
October 2025 AI safety news: Adaptive attacks, Tokenization, Impossible tasks
Dev Notes 05: Further Attempts at Understanding GPT-OSS-Safeguard Result Variation
Desire Without Apology
Make EA Great Again. My story
The Russia Ukraine war is a proxy war of China vs. the West
Why do people who get paid the most do the least?
Are LLMs automatically good at poetry yet?
Our frames are valid
"YouTubers" aren't the ones making money on YouTube
The 17% Problem
Know Your Enemy, Love Your Enemy
The 10 Principles of Slack Maximalism
Two Types of Homeostasis
Ove Arup's 1967 speech about progress, computers, and staying human in an age of automation
Failure Modes of Writing
Preference Collapse: Idiosyncrasies and Non-Stationarity Aesthetic Manifolds
Structural Innocence, Conflict License
Cash Back
The Great Selection: Will a sociopath win the game in a magic world in the age of technology?
The two types of catastrophic AI
Books Are Not About Communicating Ideas
A Love Song to Nicotine (6/30)
You Can Just Worldbuild
Text role playing games to discover task preferences in LLMs
Did You Know You Can Just Buy Blackbelts?
Persimmons, Human Beauty, and Utopia
Love after the singularity
Making a New Type of Guy (To Get Mad At)
Orwell's Language
History 101: What Are We Even Doing?
Album Review: Bittersweet
Why I am not an Objectivist
Pareto Juice II
Overview: Where’s this blog going?
Liberation Clippy
Clarification: I really fucking like my job
Taxonomy of writers

November 5, 2025

Factorio is a zombie shooter for nerds
The Sparse Autoencoders bubble has popped, but they are still promising
People should be smaller
How long does MAiD actually take in Canada?
Notes on Google’s Space Data Centers
Running Small Group Workshops
The Eiger
You should read marginally more woo (only if you're scientifically literate, that is)
Digital minimalism is out, digital intentionality is in
Are horror stories ethical?
You should publish your Stack
Dev Notes 04: Confusion/Caution(?) on running GPT-OSS-Safeguard with HuggingFace Inference
Being The Target Demographic
Agony Aunt Q&A - Dating EA Advice
Claims about AI’s impact from weak to strong
Financial Markets 2
postrationality annotated bibliography
Why does CVS's website suck
How better VFX led to better practical effects
Is a Post-Photographic Image Worth 3.6 Million Words?
The Greatest Compliment Someone Can Give My Writing
How I used Claude Code for Ops today
The Bitter Lesson Six Years On
An aside on two jokes about silly functions and what we can learn from them
If you can't life this typewriter, then give up writing
Preference vs. Execution in Creative Loops
Honor Beats Dignity
How Can I Not Know Whether I'm Having a Good Experience?
Predicting AI in 2030
Privacy Preserving Speeding Enforcement?
Do 100 things (5/30)
Animal Stories: Part 5
Can an AI Talk Its Way In? Testing Voice Persuasion in a Lab Security Game.
Halfway to Anywhere
The Future Hits Text First
Meta-agentic Prisoner's Dilemmas
[Interregnum:] Visa's dictum
A Tour of "Parfit" (2023) by David Edmonds
Why Is There Not A Single Planetary Super-State?
The Diaper Curve
Some notes on finding a partner
The Spectre
Orient Speed in the 21st Century
Geometric UDT
Interview with Nebula-nominated SF writer P. H. Lee
Dungeon Master as Metacharacter

November 4, 2025

When the heck did Hammurabi actually rule?
Build your own sequential test! An application to testing LLM decision-making
Getting paid ≠ producing value
The Bleach Bottle is Empty
Best and Worst Advice for Academic Writing
The Levels of Trust
Rose Garden Memorial
Your Movielike AI Assistant Will Already Be There
What are you excited about doing?
Write-up for the replication study
The best AI tools will become far more expensive
Dev Notes 03: With GPT-OSS-Safeguard a Simple Policy Outperforms Detailed Prompt on ToxicChat
Three Recommendations (with mini reviews)
Just
Is AI scaling hitting diminishing returns?
Financial Markets
The Jones Act Index
His crib was Metal
YouTubers are worried about AI for the wrong reasons
The Fine Art of Being Too Dumb to Regulate
Skip Traditional Phase 3 Trials for High-Burden Vaccines
The Paris Catacombs are a real life D&D
Open-source your life in a closed group
Schultz vs Leibniz on the foundations of arithmetic
Why Pole Dancing Is A Martial Art
Even Perfect Decoder-only BCIs will have less signal and precision than motor channels (*HANDS*)
You're Coming Back With Me
Do Small Protests Work?
A Guide To Being Persuasive About AI Dangers
Are we in an AI bubble?
Thoughts on Indecisiveness, or, Thoughts on Indecision?
Do Nothing Meditation (4/30)
Animal Stories: Part 4
Who is using AI romantic companions?
Not Over Or Under Indexed
Towards Underflowing Depression
A prayer for engaging in conflict
Making A New Type of Guy (To Comply)
Brazilian Jiujitsu Stories #1
All True Everything Else Is Biological, Too
Some Clothes I've Enjoyed
Moonlight Hikes
Pareto Juice I
Being "Usefully Concrete"
Intentionality
Inkhaven
Oaths aren't about oaths, they're about performative speech acts
A Letter To The Fuckboys

November 3, 2025

I ate bear fat, to prove a point
Anytime-valid tests: gambling with your data points (part 1)
Imagine history like you would a memory
Things I Learned from the Fatima Discourse™
On Submitting Your Research to a Journal
The Ball
Book Review: Will There Ever Be Another You by Patricia Lockwood
In Spite of Slop
Build the life you actually want
Raunchy replication rant
Some hidden time sinks of the writing process
Dev Notes 02: Running GPT-OSS-Safety on ToxicChat
Retvrn to the spirit of WordPress
I am the nameless Tao
RLHF: just say no?
Incentive-Compatibility is Overrated Yet Inevitable
Where Jones Act Vessels Go: An Operational Analysis
Day 3 of Inkhaven
2025 is a uniquely good time to start a YouTube channel
The Significance Itch
The 7 Types Of Advice (And 3 Common Failure Modes)
Why I mainly date other trans women
Defending the Data Processing Inequality
Hackers, wizards and scientists
BREAKING NEWS: Writer Has Existential Crisis
The Penalties of Precision: a Scale-Space Lens
Face the Ick
My Third Caffeine Self-Experiment
On tasting things: wtf is going on with chocolate?
How to survive until AGI
You think you are in control?
Disliking Foods is a Skill Issue
DARKFIELD: ARCADE (3/30)
Animal Stories: Part 3
Comparative advantage and AI
Parleying with the Principled
Suicide Diplomats as Alternative to Ems
Volatilization
What's in a name?
Obscure Film Recommendation: Time Regained (1999)
All True Wealth Is Biological (Part One)
Men Can Be (Pleasantly) Objectified Too
Delusions Get Costlier to Fix
Lines of Advance
Re-Rolling Environment
Research Reflections
Leftists want real jobs on the leftist commune
Erotic Hybrid Constitutional AI

November 2, 2025

A formula for round numbers
History is like the weather: chaotic on short timelines and stable on long
Are apples made of cells?
"You're not sick enough for this medicine."
What I Learned from Living in England
Gork, the hero we need
Book Review: A Horse Walks into a Bar by David Grossman
That Lying Bitch
A glimpse of the other side
Casual Archiving
Why I am writing 30 posts in 30 days
Dev Notes 01: Initial Exploration of GPT-OSS-Safeguard
Raison d'être
23 Pieces of Life Advice
Against subjectivism
Never Reason from a Price Change, Culture Edition
everyone else is wrong about ai art, sorry
Stepping out of the vortex
What do Matt Levine, Scott Alexander, and Patrick McKenzie have in common?
The Sublime Luxury of Saying No
How to Win Any New Board Game
Getting Serious about my Dating Doc
Lit Review: Stated vs. Revealed Preferences in LMs
Functions are an important concept for understanding the world, however the word "function" is slippery and ambiguous
Species of Bitchiness
Why Effort Scales Super-linearly with the Perceived Quality of Creative Work
The Virtue of Invariance
Things I've Become More Confident About
How and why you should make your home smart (it's cheap and secure!)
Don't gamble your life because of AGI
Escape Hatch
Writing under my real name (2/30)
Animal Stories: Part 2
I joined inkhaven
Lack of Social Grace is a Lack of Skill
My YC Pitch
Shem
The Arbitrage Window
Trigonometric and Algebraic Functions in Spacewar! (1962)
When I Say "Everywhere," I Mean Agricultural Societies Before 1800
The Recent Moral Panic
Gender Library
Trying and thinking
"What's hard about this? What can I do about that?" (Recursive)
Donut Theory of Consciousness
Body Time and Daylight Savings Apologetics
Networking is less dumb than you might think
Imports Fraud is my Love Language

November 1, 2025

Ahoy Ahoy, Mound!
Statistical tests are complicated because their inventors did not have fast computers
The origin of writing is surprisingly unclear
Dying of a Terminal Illness is not Rare
Britizenship and the Cheeky Pint
November 2025 Ben Goldhaber Newsletter
This Is Why They Swirl
Execution Ambiguity
You’re always stressed, your mind is always busy, you never have enough time
A/B testing could lead LLMs to retain users instead of helping them
Dactile Meta: Notes on Site Design and Style
Good And Real, chapter 5 notes
Destined for Greatness
Contra Parfit’s ‘Against Egoism’
Architecting Reality
Some Meetups I Ran (2025 Q2)
The first day of Inkhaven
No, twin studies are not about twins separated at birth
The Ethics of Digital Lobotomy
How to Write Fast, Weird, and Well
When did it all grow wrong?
Start of Inkhaven!
Robert Hooke's "Cyberpunk" letter to Gottfried Leibniz
The Great Nicolas Cage Fest Experiment
Read-Only Brain Computer Interfaces (BCIs) will have little use for real-time creative work
Personas of Local Maxima
Will Welfareans Get to Experience the Future?
Should you donate to Lightcone Infrastructure?
The boy who cried AGI
Starting Inkhaven
I am writing 30 posts in 30 days (1/30)
Animal Stories: PArt 1
Model welfare and open source
2025 Unofficial LW Community Census Request For Comments
Nov 1, 2025
LLM-generated text is not testimony
Type 1 Type 2
Dry Scoop Your Creatine
The Reputational Windfall
I'm In Inkhaven!
Not-Quite-A-Book-Review: The New Inkhaven Residents
Learning Falco
LoBloWriMo
Who's afraid of Verbal Overshadowing?
AI, Art, and Preserving Humanity's Soul
Weak-To-Strong Generalization
Why you should write blog posts and not be a blogger
Let Loose and Lose Customers

Previous Work

S-Risks: Fates Worse Than Extinction
The No Free Lunch theorems and their Razor
Introduction to abstract entropy
Your Book Review: Two Arms and a Head
Where is my von Braun wheel?
First, do no harm - how the FDA crippled America's response to COVID-19
Blue Tie Number 5
Attitudes about Applied Rationality
Guide to rationalist interior decorating
You don't get to have cool flaws
GPT-4o draws itself as a consistent type of guy
Is AI Replacing Jobs Right Now? Analysis by Age Appears to Have Challenges.
Of Cowards And Grandmas
I am a Freudian now
GDP (per capita) in 2050
No One is Really Working
Yukon Text Posts
My birth story
Your Book Review: Lying for Money
A Protopian Vision for the Age of Intelligence
Why Are We All Cowards?
How I switched careers from software engineer to AI policy operations
Beware the Delmore Effect
Some things I learnt studying philosophy 2019-2021
Why Explicit Boundaries Don't Always Work
The Business of Extracting Knowledge from Academic Publications
The Lazy Evil Genius
In Which I Defend Fruit's Honor
How to Give in to Threats (without incentivizing them)
Orienting to 3 year AGI timelines
Zen and The Art of Semiconductor Manufacturing
Mystic Jew Powers
How I Lost My Backpack with Passports and Laptop
Roadside Attraction Review: Bishop's Castle
Why I don't believe Superalignment will work
Write a Thousand Roads to Rome
The Company Man
The possible shared Craft of deliberate Lexicogenesis
Who Is Portugal For?
Racial Dating Preferences and Sexual Racism
A Review Of The Evidence For Joan of Arc
12 interesting things I learned studying the discovery of nature's laws
The Other AI Art
Steelmanning Divination
Scene. The lunch room of an EA coworking space.
Integrity and accountability are core parts of rationality
Feedbackloop-first Rationality
A "weak" AGI may attempt an unlikely-to-succeed takeover
Conversational Cultures: Combat vs Nurture (V2)
Embedded Agents
The Power of Rearticulating Insights in Your Own Words
The Life Goals of Dead People
Recovering Fuckboy : Grace and I