Pages with the most revisions
Showing below up to 413 results in range #1 to #413.
View (previous 500 | next 500) (20 | 50 | 100 | 250 | 500)
- List of disagreements in AI safety (182 revisions)
- List of men by number of sons, daughters, and wives (159 revisions)
- Secret sauce for intelligence (78 revisions)
- List of AI safety projects I could work on (46 revisions)
- Future planning (44 revisions)
- Something like realism about rationality (44 revisions)
- Deliberate practice for learning proof-based math (43 revisions)
- What makes a word explanation good? (42 revisions)
- Spaced proof review (41 revisions)
- List of arguments against working on AI safety (39 revisions)
- People watching (35 revisions)
- Missing gear for intelligence (33 revisions)
- UDASSA (32 revisions)
- Missing gear vs secret sauce (29 revisions)
- Architecture (29 revisions)
- AI prepping (26 revisions)
- List of experiments with Anki (24 revisions)
- How doomed are ML safety approaches? (23 revisions)
- Late 2021 MIRI conversations (23 revisions)
- My understanding of how IDA works (23 revisions)
- Comparison of pedagogical scenes (23 revisions)
- Medium that reveals flaws (22 revisions)
- There is room for something like RAISE (21 revisions)
- The Hour I First Believed (21 revisions)
- Soft-hard takeoff (21 revisions)
- Comparison of AI takeoff scenarios (21 revisions)
- Credit card research 2021 (20 revisions)
- List of success criteria for HRAD work (20 revisions)
- AlphaGo (20 revisions)
- Explorable explanation (19 revisions)
- SuperMemo shortcuts (19 revisions)
- My current thoughts on the technical AI safety pipeline (outside academia) (18 revisions)
- Spaced proof review is not about memorizing proofs (18 revisions)
- Convergent evolution of values (18 revisions)
- Analyzing disagreements (18 revisions)
- Human safety problem (17 revisions)
- Lumpiness (17 revisions)
- List of teams at OpenAI (17 revisions)
- Spaced proof review routine (17 revisions)
- Improvement curve for good people (16 revisions)
- Spoiler test of depth (16 revisions)
- Asymmetric institution (16 revisions)
- MIRI vs Paul research agenda hypotheses (15 revisions)
- Main Page (14 revisions)
- Interaction reversal between knowledge-to-be-memorized and ideas-to-be-developed (14 revisions)
- Duolingo (13 revisions)
- Hardware argument for AI timelines (13 revisions)
- Anki deck philosophy (13 revisions)
- Jelly no Puzzle (12 revisions)
- Duolingo for math (12 revisions)
- Science argument (11 revisions)
- Goalpost for usefulness of HRAD work (11 revisions)
- Incremental reading in Anki (11 revisions)
- Narrow vs broad cognitive augmentation (11 revisions)
- Add all permutations of a card to prevent pattern-matching (10 revisions)
- AI safety field consensus (10 revisions)
- Finiteness assumption in explorable media (10 revisions)
- Spaced repetition is not about memorization (10 revisions)
- Emotional difficulties of spaced repetition (10 revisions)
- List of terms used to describe the intelligence of an agent (10 revisions)
- Probability and statistics as fields with an exploratory medium (10 revisions)
- Deck options for proof cards (10 revisions)
- Why ain'tcha better at math (10 revisions)
- Tips for reviving a spaced repetition practice (10 revisions)
- Desiderata for dissolving the question (10 revisions)
- List of reasons something isn't popular or successful (10 revisions)
- 3Blue1Brown (10 revisions)
- Reverse side card for everything (9 revisions)
- Progress in self-improvement (9 revisions)
- Popularity symbiosis (9 revisions)
- Braid for math (9 revisions)
- Spaced writing inbox (9 revisions)
- Spaced repetition allows graceful deprecation of experiments (9 revisions)
- Expert response heuristic for prompt writing (9 revisions)
- Switching costs of various kinds of software (9 revisions)
- Narrow window argument against continuous takeoff (9 revisions)
- Ongoing friendship and collaboration is important (9 revisions)
- Evolution (8 revisions)
- Can spaced repetition interfere with internal sense of relevance? (8 revisions)
- Choosing problems for spaced proof review (8 revisions)
- List of timelines for futuristic technologies (8 revisions)
- Coherence and goal-directed agency discussion (8 revisions)
- Tinkering in math requires loading the situation into working memory (8 revisions)
- Competence gap (8 revisions)
- Linked list proof card (8 revisions)
- Deck options for small cards (8 revisions)
- Unbounded working memory assumption in explanations (8 revisions)
- Christiano's operationalization of slow takeoff (8 revisions)
- Online question-answering services are unreliable (8 revisions)
- Optimal unlocking mechanism for booster cards is unclear (8 revisions)
- Nobody understands what makes people snap into AI safety (8 revisions)
- Hyperbolic growth (8 revisions)
- Self-graded prompts made for others must provide guidance for grading (8 revisions)
- Simple core of consequentialist reasoning (8 revisions)
- The Precipice notes (8 revisions)
- Booster card (7 revisions)
- Spaced repetition experiments take months to complete (7 revisions - redirect page)
- Pivotal act (7 revisions)
- Importance of knowing about AI takeoff (7 revisions)
- Cognitive biases that are opposites of each other (7 revisions)
- List of people who have thought a lot about spaced repetition (7 revisions)
- Agent foundations (7 revisions)
- The Secret of Psalm 46 outline (7 revisions)
- The Secret of Psalm 46 (7 revisions)
- AI safety lacks a space to ask stupid or ballsy questions (7 revisions)
- Different senses of claims about AGI (7 revisions)
- AI safety is harder than most things (7 revisions)
- List of breakthroughs plausibly needed for AGI (7 revisions)
- Is AI safety no longer a scenius? (7 revisions)
- Emotional difficulties of AI safety research (7 revisions)
- Creative forgetting (7 revisions)
- Spaced inbox ideas (7 revisions)
- Snoozing epicycle (7 revisions)
- Sudden emergence (6 revisions)
- Unintended consequences of AI safety advocacy argument against AI safety (6 revisions)
- Mass shift to technical AI safety research is suspicious (6 revisions)
- The Witness (6 revisions)
- Fractally misfit (6 revisions)
- AI safety technical pipeline does not teach how to start having novel thoughts (6 revisions)
- Card sharing allows less valuable cards to be created (6 revisions)
- Empty review (6 revisions)
- Feeling like a perpetual student in a subject due to spaced repetition (6 revisions)
- Quotability vs ankifiability (6 revisions)
- Summary of my beliefs (6 revisions)
- Stream of low effort questions helps with popularity (6 revisions)
- Jonathan Blow (6 revisions)
- Big card (6 revisions)
- Laplace's rule of succession argument for AI timelines (6 revisions)
- Thinking Mathematics (6 revisions)
- Spaced inbox review should not be completionist or obligatory (6 revisions)
- Counterfactual of dropping a seed AI into a world without other capable AI (6 revisions)
- Comparison of terms related to agency (6 revisions)
- AlphaGo as evidence of discontinuous takeoff (6 revisions)
- Richard Ngo (6 revisions)
- Do an empty review of proof cards immediately after adding to prevent backlog (6 revisions)
- Anki deck options (6 revisions)
- Spaced repetition and cleaning one's room (6 revisions)
- Video games comparison to math (6 revisions)
- Tao Analysis I exercise count (6 revisions)
- Instruction manuals vs giving the answers (6 revisions)
- Iterated amplification (6 revisions)
- Spaced repetition (6 revisions)
- Proof card (5 revisions)
- Paperclip maximizer (5 revisions)
- Discontinuities in usefulness of whole brain emulation technology (5 revisions)
- Use paper during spaced repetition reviews (5 revisions)
- Resource overhang (5 revisions)
- Will there be significant changes to the world prior to some critical AI capability threshold being reached? (5 revisions)
- Changing selection pressures argument (5 revisions)
- Simple core (5 revisions)
- My beginner incremental reading questions (5 revisions)
- Discursive texts are difficult to ankify (5 revisions)
- Reference class forecasting on human achievements argument for AI timelines (5 revisions)
- List of techniques for managing working memory in explanations (5 revisions)
- Hardware-driven vs software-driven progress (5 revisions)
- Giving advice in response to generic questions is difficult but important (5 revisions)
- AI timelines (5 revisions)
- Spaced proof review as a way to invent novel proofs (5 revisions)
- Textbook test for AI theory (5 revisions)
- It is difficult to get feedback on published work (5 revisions)
- Open-ended questions are common in real life (5 revisions)
- Card sharing (5 revisions)
- Make new cards when you get stuck (5 revisions)
- List of techniques for making small cards (5 revisions)
- Will it be possible for humans to detect an existential win? (4 revisions)
- Personhood API vs therapy axis of interpersonal interactions (4 revisions)
- Distortion of true frequency of piece of knowledge due to spaced repetition (4 revisions - redirect page)
- AI takeoff (4 revisions)
- Andy Matuschak (4 revisions)
- Representing impossibilities (4 revisions)
- Physical vs digital clutter (4 revisions)
- AI will solve everything argument against AI safety (4 revisions)
- Stupid questions (4 revisions)
- Dual ratings for spaced inbox (4 revisions)
- One-sentence summary card (4 revisions)
- Fractional progress argument for AI timelines (4 revisions)
- Busy life periods and spaced inbox (4 revisions)
- Wei Dai (4 revisions)
- Spaced repetition is useful because most knowledge is sparsely applicable (4 revisions)
- Explanation science (4 revisions)
- Asynchronous support (4 revisions)
- Depictions of learning in The Blue Lagoon are awful (4 revisions)
- Make Anki cards based on feedback you receive (4 revisions)
- Duolingo does repetition at the lesson level (4 revisions)
- Iteration cadence for spaced repetition experiments (4 revisions)
- There is pressure to rush into a technical agenda (4 revisions)
- Comparison of sexually transmitted diseases (4 revisions)
- Repetition granularity (4 revisions)
- Combinatorial explosion in math (4 revisions)
- Equivalence classes of prompts (4 revisions)
- Dealing with bad problems in spaced proof review (4 revisions)
- The mathematics community has no clear standards for what a mathematician should know (4 revisions)
- Timeline of my involvement in AI safety (4 revisions)
- Spaced everything (4 revisions)
- Uninsightful articles can seem insightful due to unintentional spaced repetition (4 revisions)
- Pascal's mugging and AI safety (4 revisions)
- Scaling hypothesis (4 revisions)
- Maintaining habits is hard, and spaced repetition is a habit (4 revisions)
- Feynman technique fails when existing explanations are bad (4 revisions)
- Minimal AGI vs task AGI (4 revisions)
- Prosaic AI (4 revisions)
- Scenius (4 revisions)
- Spaced repetition as soft alarm clock (4 revisions)
- Continuous takeoff (4 revisions)
- Statistical analysis of expert timelines argument for AI timelines (4 revisions)
- Flag things to fix during review (4 revisions)
- Using spaced repetition to improve public discourse (4 revisions)
- Kasparov window (4 revisions)
- Cards created by oneself can be scheduled more aggressively (4 revisions)
- Discovery fiction (4 revisions)
- Continually make new cards (4 revisions)
- Website to aggregate solutions to textbook exercises (4 revisions)
- Setting up Windows (4 revisions)
- Rapid capability gain vs AGI progress (4 revisions)
- Tutoring heuristic for prompt writing (4 revisions)
- Spaced proof review as a way to understand key insights in a proof (4 revisions)
- Braid (4 revisions)
- Spaced repetition world (4 revisions)
- Anki reviews are more fun on mobile (4 revisions)
- Late singularity (4 revisions)
- Word explanations are already great (4 revisions)
- Difficulty of AI alignment (3 revisions)
- Should booster cards be marked as new? (3 revisions)
- One wrong number problem (3 revisions)
- Aligning smart AI using slightly less smart AI (3 revisions)
- Deconfusion (3 revisions)
- Vow of silence (3 revisions)
- Central node trick for remembering equivalent properties (3 revisions)
- Politicization of AI (3 revisions)
- The Sequences vs evergreen notes (3 revisions)
- We still don't know how to systematically write great word explanations (3 revisions)
- Hardware overhang (3 revisions)
- Distillation is not enough (3 revisions)
- AI safety is not a community (3 revisions)
- Carl Shulman (3 revisions)
- Live math video (3 revisions)
- Hnous927 (3 revisions)
- Highly reliable agent designs (3 revisions)
- Learning-complete (3 revisions)
- Corrigibility (3 revisions)
- Spaced repetition constantly reminds one of inadequacies (3 revisions)
- Spaced repetition prevents unrecalled unrecallables (3 revisions)
- Finding the right primitives for spaced repetition responses (3 revisions)
- Inter-personal comparison test (3 revisions)
- Video games allow immediate exploration (3 revisions)
- My take on RAISE (3 revisions)
- Ignore Anki add-ons to focus on fundamentals (3 revisions)
- What would a vow of silence look like for math? (3 revisions)
- Paul Christiano (3 revisions)
- Managing micro-movements in learning (3 revisions)
- HCH (3 revisions)
- Large graduating interval as substitute for putting effort into making atomic cards (3 revisions)
- Incremental reading (3 revisions)
- It is difficult to find people to bounce ideas off of (3 revisions)
- List of big discussions in AI alignment (3 revisions)
- Using Anki for math (3 revisions)
- Spaced repetition as generator of questions (3 revisions)
- Integration card (3 revisions)
- The Uncertain Future (3 revisions)
- Bury effortful cards to speed up review (3 revisions)
- IDA (2 revisions - redirect page)
- Fake motivation (2 revisions)
- Test (2 revisions)
- Anki (2 revisions)
- Soren Bjornstad (2 revisions)
- Existential win (2 revisions)
- Text to speech software (2 revisions)
- Piotr Wozniak (2 revisions)
- List of interesting search engines (2 revisions)
- Use temporary separate Anki decks to learn new cards based on priority (2 revisions)
- Debates shift bystanders' beliefs (2 revisions)
- Interacting with copies of myself (2 revisions)
- Application of functional updateless timeless decision theory to everyday life (2 revisions)
- Michael Nielsen (2 revisions)
- Whole brain emulation (2 revisions)
- Tao Analysis Solutions (2 revisions)
- Paul christiano (2 revisions - redirect page)
- Exhaustive quizzing allows impatient learners to skip the reading (2 revisions)
- Secret sauce for intelligence vs specialization in intelligence (2 revisions)
- Encoding dependence problem (2 revisions)
- Big cards can be good for mathematical discovery (2 revisions)
- How meta should AI safety be? (2 revisions)
- Value learning (2 revisions)
- Page template (2 revisions)
- Task-dependent diversity (2 revisions)
- Corrigibility may be undesirable (2 revisions)
- What counts as good motivation? (2 revisions)
- Spaced repetition response as chat message or chat reaction (2 revisions)
- Different mental representations of mathematical objects is a blocker for an exploratory medium of math (2 revisions)
- Kanzi (2 revisions)
- Steam game buying algorithm (2 revisions)
- Unreliability of online question-answering services makes it emotionally taxing to write up questions (2 revisions)
- Jessica Taylor (2 revisions)
- Selection effect for who builds AGI (2 revisions)
- Existing implementations of card sharing have nontrivial overhead (2 revisions)
- Philosophical difficulty (2 revisions)
- Newcomers in AI safety are silent about their struggles (2 revisions)
- OpenAI (2 revisions)
- Small card (2 revisions)
- Mapping mental motions to parts of a spaced repetition algorithm (2 revisions)
- List of technical AI alignment agendas (2 revisions)
- Robin Hanson (2 revisions)
- Add the complete proof on proof cards to reduce friction when reviewing (2 revisions)
- List of critiques of iterated amplification (2 revisions)
- Large graduating interval as a way to prevent pattern-matching (2 revisions)
- Selection effect for successful formalizations (2 revisions)
- Definitions last (2 revisions)
- Generic advice is difficult to give but also important (1 revision - redirect page)
- Soft takeoff keyhole (1 revision - redirect page)
- Wei (1 revision - redirect page)
- Takeoff speeds (1 revision - redirect page)
- MIRI (1 revision - redirect page)
- Potpourri hypothesis (1 revision - redirect page)
- Permutation trick (1 revision - redirect page)
- Quick review (1 revision - redirect page)
- Understanding is discontinuous (1 revision - redirect page)
- Human safety (1 revision - redirect page)
- Simple core algorithm of agency (1 revision - redirect page)
- Video games vs math (1 revision - redirect page)
- Extrapolation argument for AI timelines (1 revision - redirect page)
- You don't need to eat your own dogfood in explanation science (1 revision)
- List of thought experiments in AI safety (1 revision)
- Spaced inbox (1 revision - redirect page)
- Math vs video games (1 revision - redirect page)
- Discontinuities in whole brain emulation technology (1 revision - redirect page)
- Minimal AGI (1 revision)
- Content sharing between AIs (1 revision)
- Carl (1 revision - redirect page)
- Evolution analogy (1 revision - redirect page)
- Hnous (1 revision - redirect page)
- Add easy problems as cards with large graduating interval (1 revision)
- WBE (1 revision - redirect page)
- Doomer argument against AI safety (1 revision)
- Keyhole argument against continuous takeoff (1 revision - redirect page)
- Uncertain Future (1 revision - redirect page)
- Fake review (1 revision - redirect page)
- Tao Analysis Flashcards (1 revision)
- Mixed messaging regarding independent thinking (1 revision)
- Kasparov Window (1 revision - redirect page)
- HRAD (1 revision - redirect page)
- Output curve for good people (1 revision - redirect page)
- AI polarization (1 revision - redirect page)
- AlphaGo Zero (1 revision - redirect page)
- Master How To Learn guy (1 revision - redirect page)
- SuperMemo (1 revision)
- How similar are human brains to chimpanzee brains? (1 revision)
- Paul (1 revision - redirect page)
- Can the behavior of approval-direction be undefined or random? (1 revision)
- People are bad (1 revision)
- Short-term preferences-on-reflection (1 revision)
- Teaching-complete (1 revision - redirect page)
- Simple core algorithm for agency (1 revision - redirect page)
- Continuous AI takeoff (1 revision - redirect page)
- Education in The Blue Lagoon is awful (1 revision - redirect page)
- Are due counts harmful? (1 revision)
- Teaching complete (1 revision - redirect page)
- Goal inference (1 revision - redirect page)
- Lumpy (1 revision - redirect page)
- Explosive aftermath (1 revision)
- Deliberate practice for math (1 revision - redirect page)
- Intra-personal comparison test (1 revision)
- Interaction reversal between knowledge and notes (1 revision - redirect page)
- Pascal's mugging argument against AI safety (1 revision - redirect page)
- Learning complete (1 revision - redirect page)
- Terms used to describe intelligence of agents (1 revision - redirect page)
- Toby Ord questions (1 revision - redirect page)
- Discontinuity to AGI (1 revision - redirect page)
- Secret sauce (1 revision - redirect page)
- Parasitizing on popularity (1 revision - redirect page)
- Tricky examples in basic probability (1 revision)
- Asymmetric weapon (1 revision - redirect page)
- Medium that reveals its flaws (1 revision - redirect page)
- Simple core algorithm (1 revision - redirect page)
- Thinking Physics for math (1 revision - redirect page)
- Discontinuity to HLMI (1 revision - redirect page)
- Desiderata for cognitive reductions (1 revision - redirect page)
- Number of relevant actors around the time of creation of AGI (1 revision)
- Spaced inbox review should not be completionist (1 revision - redirect page)
- Meta-execution (1 revision)
- Slow takeoff (1 revision - redirect page)
- Guided by the beauty of our weapons (1 revision - redirect page)
- Realism about rationality (1 revision - redirect page)
- Discontinuity from AGI (1 revision - redirect page)
- Slow feedback loop for spaced repetition (1 revision - redirect page)
- Weird recursion (1 revision)
- Missing gear (1 revision - redirect page)
- Fast takeoff (1 revision - redirect page)
- Nate (1 revision - redirect page)
- Yudkowskian (1 revision - redirect page)
- Humans consulting HCH (1 revision - redirect page)
- Spaced repetition isn't about memorization (1 revision - redirect page)
- Discontinuity from HLMI (1 revision - redirect page)
- Single-architecture generality (1 revision)
- Slow feedback cycle for spaced repetition (1 revision - redirect page)
- DSA (1 revision - redirect page)
- Hansonian (1 revision - redirect page)
- Hanson (1 revision - redirect page)
- Single-model generality (1 revision)
- RSI (1 revision - redirect page)
- Video games work with finite and discrete objects (1 revision - redirect page)
- List of timelines (1 revision - redirect page)
- Broad augmentation (1 revision - redirect page)
- If you want to succeed in the video games industry (1 revision)
- Bury cards to speed up review (1 revision - redirect page)
- Continuous takeoff keyhole (1 revision - redirect page)
- RCG (1 revision - redirect page)
- Second species argument (1 revision)
- Soft alarm clock (1 revision - redirect page)
- Eliezer (1 revision - redirect page)
- Takeoff speed (1 revision - redirect page)
- Narrow augmentation (1 revision - redirect page)
- Potpourri hypothesis for math education (1 revision)
- Intelligence amplification (1 revision)