carado.moe Open in urlscan Pro
2001:4b98:dc0:43:f816:3eff:fee9:dedc  Public Scan

Form analysis
0 forms found in the DOM

Text Content

CARADO.MOE

tammy's blog about
AI alignment, utopia, anthropics, and more;

also on twitter, lesswrong, rss, and at


TOP POSTS ★

 * 2023-06-10 · ★ formalizing the QACI alignment formal-goal
 * 2023-06-10 · ★ an Evangelion dialogue explaining the QACI alignment plan
 * 2023-03-20 · ★ the QACI alignment plan: table of contents
 * 2022-12-31 · ★ end of 2022: my life so far
 * 2022-12-14 · ★ all claw, no world — and other thoughts on the universal
   distribution
 * 2022-11-01 · ★ a casual intro to AI doom and alignment
 * 2022-10-03 · ★ my current outlook on AI risk mitigation
 * 2022-08-21 · ★ everything is okay


ALL POSTS

 * 2023-07-10 · epistemic range
 * 2023-06-15 · a short chat about realityfluid
 * 2023-06-10 · ★ formalizing the QACI alignment formal-goal
 * 2023-06-10 · ★ an Evangelion dialogue explaining the QACI alignment plan
 * 2023-05-06 · Orthogonal's Formal-Goal Alignment theory of change
 * 2023-04-22 · the multiverse argument argument against automated alignment
 * 2023-03-25 · opinions on the consequences of AI
 * 2023-03-23 · continue working on hard alignment! don't give up!
 * 2023-03-20 · ★ the QACI alignment plan: table of contents
 * 2023-03-19 · you can't simulate the universe from the beginning?
 * 2023-03-19 · QACI blob location: an issue with firstness
 * 2023-03-13 · your terminal values are complex and not objective
 * 2023-03-12 · the quantum amplitude argument against ethics deduplication
 * 2023-03-11 · problems for formal alignment
 * 2023-03-09 · QACI blobs and interval illustrated
 * 2023-03-08 · QACI blob location: no causality & answer signature
 * 2023-03-06 · before the sharp left turn: what wins first?
 * 2023-03-05 · QACI: the problem of blob location, causality, and
   counterfactuals
 * 2023-03-03 · state of my alignment research, and what needs work
 * 2023-03-01 · on strong/general coherent agents
 * 2023-02-26 · some thoughts about terminal alignment
 * 2023-02-25 · clarifying formal alignment implementation
 * 2023-02-23 · Hello, Elua.
 * 2023-02-16 · don't censor yourself, silly !
 * 2023-02-15 · a narrative explanation of the QACI alignment plan
 * 2023-02-14 · explaining "."
 * 2023-02-13 · is intelligence program inversion?
 * 2023-02-12 · fuzzies & utils: check that you're getting either
 * 2023-02-11 · my takeoff speeds? depends how you define that
 * 2023-02-11 · GPT is dangerous because it is useful at all
 * 2023-02-09 · CEV can be coherent enough
 * 2023-02-07 · tabooing "AGI"
 * 2023-02-07 · so you think you're not qualified to do technical alignment
   research?
 * 2023-02-07 · word report #3
 * 2023-01-29 · formal alignment: what it is, and some proposals
 * 2023-01-29 · a guess at my intrinsic values
 * 2023-01-29 · communicating with successful alignment timelines
 * 2023-01-27 · to me, it's instrumentality that is alienating
 * 2023-01-19 · nostalgia: a value pointing home
 * 2022-12-31 · ★ end of 2022: my life so far
 * 2022-12-28 · making decisions as our approximately simulated selves
 * 2022-12-22 · one-shot AI, delegating embedded agency and decision theory, and
   one-shot QACI
 * 2022-12-22 · being only polynomial capabilities away from alignment: what a
   great problem to have that would be!
 * 2022-12-21 · the scarcity of moral patient involvement
 * 2022-12-19 · our deepest wishes
 * 2022-12-15 · how far are things that care?
 * 2022-12-14 · ★ all claw, no world — and other thoughts on the universal
   distribution
 * 2022-12-11 · a rough sketch of formal aligned AI using QACI
 * 2022-11-22 · just enough spoilers for a fire upon the deep to read a
   yudkowsky fanfic
 * 2022-11-20 · CYOAs and futurism
 * 2022-11-20 · let's stick with the term "moral patient"
 * 2022-11-19 · wonky but good enough alignment schemes
 * 2022-11-19 · logical vs indexical dignity
 * 2022-11-18 · "humans aren't aligned" and "human values are incoherent"
 * 2022-11-18 · generalized wireheading
 * 2022-11-13 · a safer experiment than quantum suicide
 * 2022-11-12 · fully aligned singleton as a solution to everything
 * 2022-11-01 · ★ a casual intro to AI doom and alignment
 * 2022-10-31 · publishing alignment research and exfohazards
 * 2022-10-30 · love, not competition
 * 2022-10-28 · counterfactual computations in world models
 * 2022-10-24 · QACI: question-answer counterfactual intervals
 * 2022-10-12 · some simulation hypotheses
 * 2022-10-11 · outer alignment: two failure modes and past-user satisfaction
 * 2022-10-06 · confusion about alignment requirements
 * 2022-10-03 · ★ my current outlook on AI risk mitigation
 * 2022-09-27 · surprise! you want what you want
 * 2022-09-27 · existential self-determination
 * 2022-09-16 · ordering capability thresholds
 * 2022-09-15 · clippy in panpsychia
 * 2022-09-09 · ethics and anthropics of homomorphically encrypted computations
 * 2022-09-08 · AI alignment curves
 * 2022-09-07 · ethics juice and anthropic juice
 * 2022-09-05 · program searches
 * 2022-08-21 · ★ everything is okay
 * 2022-08-20 · PreDCA: vanessa kosoy's alignment protocol
 * 2022-08-17 · alignment researchspace is potentially malign
 * 2022-08-17 · carmack predictions
 * 2022-08-17 · guiding your brain: go with your gut!
 * 2022-08-17 · alignment research is very weird
 * 2022-08-15 · essential inequality vs functional inequivalence
 * 2022-08-15 · trading with superintelligence: a wonky proto-alignment scheme
 * 2022-08-14 · why my timelines are short: all roads lead to doom
 * 2022-08-13 · goal-program bricks
 * 2022-08-13 · anthropic mindfulness
 * 2022-08-13 · the foundation book
 * 2022-08-12 · future paths
 * 2022-08-12 · what does it mean to value our survival?
 * 2022-08-12 · scopes of utopia
 * 2022-08-11 · unviable moral patients
 * 2022-08-11 · the Insulated Goal-Program idea
 * 2022-08-10 · ruling out intuitions about materially acausal things
 * 2022-08-07 · probability under potential hardware failure
 * 2022-08-07 · quantum immortality and local deaths under X-risk
 * 2022-08-04 · tiling the cosmos might be unavoidable
 * 2022-08-03 · isn't it weird that we have a chance at all?
 * 2022-07-27 · an anthropics example
 * 2022-07-02 · generalized values: testing for patterns in computation
 * 2022-06-21 · recommending Hands and Cities
 * 2022-06-18 · generalized computation interpretability
 * 2022-06-18 · anthropic reasoning coordination
 * 2022-06-18 · solonomonoff induction, time penalty, the universal program, and
   deism
 * 2022-06-10 · outer alignment: politics & philosophy
 * 2022-06-10 · where are your alignment bits?
 * 2022-06-10 · "AI risk drone"
 * 2022-06-10 · diversity vs novelty
 * 2022-05-29 · concentric rings of illiberalism
 * 2022-05-28 · say "AI risk mitigation" not "alignment"
 * 2022-05-28 · plausible vs likely
 * 2022-05-23 · implementing the platonic realm
 * 2022-05-22 · AI boxing could be easy
 * 2022-05-19 · predictablizing ethic deduplication
 * 2022-05-19 · generalized adding reality layers
 * 2022-05-16 · smaller X-risk
 * 2022-05-14 · cognitive biases regarding the evaluation of AI risk when doing
   AI capabilities work
 * 2022-05-13 · life refocus
 * 2022-05-12 · hope for infinite compute
 * 2022-05-12 · AI risk plans
 * 2022-04-30 · the uncertainty of 2+2=4
 * 2022-04-30 · a unit for utils
 * 2022-04-13 · finding earth in the universal program
 * 2022-04-12 · The Peerless
 * 2022-04-10 · bracing for the alignment tunnel
 * 2022-03-30 · should we implement free will?
 * 2022-03-22 · goals for emergency unaligned AI
 * 2022-03-21 · the word "syntax" in programming, linguistics and LISP
 * 2022-03-21 · values system as test-driven development
 * 2022-03-21 · are there finitely many moral patients?
 * 2022-03-21 · making the UD and UDASSA less broken: identifying time steps
 * 2022-03-06 · experience/moral patient deduplication and ethics
 * 2022-03-04 · recognition
 * 2022-03-02 · do not hold on to your believed intrinsic values — follow your
   heart!
 * 2022-02-23 · my current pyramid of needs
 * 2022-02-06 · forking bitrate and entropy control
 * 2022-02-04 · balancing utilitarianism
 * 2022-02-04 · political technology
 * 2022-02-03 · hackable multiverse
 * 2022-02-02 · a cognitively hazardous idea
 * 2022-01-11 · how timelines fall
 * 2022-01-11 · uploading people for alignment purposes
 * 2022-01-07 · questions about the cosmos and rich computations
 * 2022-01-05 · brittle physics and the nature of X-risks
 * 2021-12-27 · less quantum immortality?
 * 2021-12-25 · yes room above paperclips?
 * 2021-12-25 · non-scarce compute means moral patients might not get optimized
   out
 * 2021-12-25 · thinking about psi: as a more general json
 * 2021-12-25 · database transactions: you guessed it, it's WASM again
 * 2021-12-09 · psi rewriting
 * 2021-12-09 · unoptimal superintelligence doesn't lose
 * 2021-12-09 · emotionally appreciating grand political visions
 * 2021-12-09 · freedom and diversity in Albion's Seed
 * 2021-12-09 · non-interfering superintelligence and remaining philosophical
   progress: a deterministic utopia
 * 2021-12-05 · the deobfuscation conjecture
 * 2021-12-04 · think in what ?
 * 2021-11-21 · Genuineness, Existential Selfdetermination, Satisfaction: pick 2
 * 2021-11-21 · the two-vtable problem
 * 2021-11-20 · unoptimal superintelligence loses
 * 2021-11-20 · endiannesses
 * 2021-11-20 · rust & wasm, without wasm-pack
 * 2021-11-20 · no room above paperclips
 * 2021-11-08 · against AI alignment ?
 * 2021-11-08 · psi: a universal format for structured information
 * 2021-11-01 · saving the server-side of the internet: just WASM, again
 * 2021-10-25 · lamenting nerds
 * 2021-10-23 · alignment is an optimization processes problem
 * 2021-10-22 · to wasm and back again: the essence of portable programs
 * 2021-10-14 · cosmic missing outs
 * 2021-10-13 · exact minds in an exact world
 * 2021-10-11 · meta-tracking
 * 2021-09-12 · do not form your own opinion
 * 2021-08-31 · ∀V: A Utopia For Ever
 * 2021-08-25 · what happens when you die?
 * 2021-08-23 · right to death, therefore
 * 2021-08-16 · kolmogorov complexity objectivity and languagespace
 * 2021-08-15 · book recommendation: Greg Egan's Axiomatic
 * 2021-07-25 · what is value?
 * 2021-07-21 · culture tribes and legitimacy
 * 2021-07-21 · systems and diversity
 * 2021-07-19 · botched alignment and alignment awareness
 * 2021-07-18 · when in doubt, kill everyone
 * 2021-07-18 · AI alignment timeline codes
 * 2021-07-17 · AI alignment and wolfram physics
 * 2021-07-16 · universal complete
 * 2021-07-10 · estimating the amount of populated intelligence explosion
   timelines
 * 2021-07-09 · purposes for art
 * 2021-06-30 · we're all doomed
 * 2021-06-29 · disclosing subjectivity
 * 2021-06-25 · classifying computational frameworks
 * 2021-06-25 · degrees of runtime metaprogrammability
 * 2021-06-20 · cm21, a pixel art editor
 * 2021-06-18 · categories of knowledge representation
 * 2021-06-16 · the systematic absence of libertarian thought
 * 2021-06-16 · my answer to the fermi paradox
 * 2021-06-16 · the many faces of chaos magick
 * 2021-06-16 · the persistent data structure argument against linear
   consciousness
 * 2021-06-16 · refusing to answer ≠ giving a negative answer
 * 2021-06-05 · I'm creating a world simulation video game
 * 2021-06-04 · Overcoming Narratives
 * 2021-05-16 · Saving The Client-Side Web: just WASM and the DOM
 * 2021-05-03 · The Unsatisfactorily Far Reach Of Property
 * 2021-05-03 · Video Games Needs A Platform
 * 2021-04-28 · Plausible Quantum Suicide
 * 2021-04-24 · CC_ -1
 * 2021-04-24 · Let's not generalize over people
 * 2021-04-01 · Value and Earning
 * 2021-03-31 · "4=5"
 * 2021-03-28 · Cultural and Memetic Hygiene
 * 2021-03-04 · From-above vs Fine-grain diversity
 * 2021-03-04 · Value Crystallization
 * 2021-03-04 · Normies Are in Hell Too
 * 2021-03-04 · Symbology for Topia
 * 2021-03-04 · Growth Doesn't Care About Crises
 * 2021-01-22 · Communicating Clearly
 * 2021-01-14 · A canonical bit-encoding for ranged integers
 * 2021-01-01 · Non-solving ideologies
 * 2020-12-31 · Core values: Defining freedom
 * 2020-12-29 · A canonical and efficient byte-encoding for ints
 * 2020-12-25 · You are your information system
 * 2020-12-24 · CSS for pixeley images
 * 2020-12-23 · Unfair feedback loops
 * 2020-12-22 · Rationalist by necessity
 * 2020-12-21 · Against Unicode
 * 2020-12-20 · Cringe as prejudice
 * 2020-11-18 · A Prototypeness Hierarchy of Realities
 * 2020-11-15 · Two Principles For Topia
 * 2020-10-04 · Gender Bootstrappism
 * 2020-10-04 · Word Report #2
 * 2020-10-04 · Socialism as a conspiracy theory
 * 2020-10-04 · For UBI
 * 2020-10-04 · Real quick, on free will
 * 2020-10-04 · Where next for piracy ?
 * 2020-09-09 · Determining core values & existential self-determination
 * 2020-07-19 · Cool linguistic purisms
 * 2020-07-18 · Progress/decline in fields
 * 2020-06-28 · Song Pairs that can be listened to together
 * 2020-06-02 · Word Report #1
 * 2020-04-30 · the Economic Compass
 * 2020-04-27 · Limiting Real Universes
 * 2020-04-16 · A Collection Of Compasses
 * 2020-03-31 · Book Review: 12 Rules For Life
 * 2020-03-30 · Topia: Layer 0
 * 2020-03-29 · the Belief In Society compass
 * 2020-03-28 · On Economics
 * 2019-12-31 · KOLSITAN, a tiny video game
 * 2019-08-27 · Building The Castle vs Finding The Monolith
 * 2019-07-28 · Some post- words for the future
 * 2019-04-11 · Semantics: Primes and Universals, a book review
 * 2019-03-26 · The Last Global Era
 * 2017-09-26 · Analogpunk

unless otherwise specified on individual pages, all posts on this website are
licensed under the CC_-1 license.
unless explicitely mentioned, all content on this site was created by me; not by
others nor AI.