Long pages
Showing below up to 100 results in range #101 to #200.
View (previous 100 | next 100) (20 | 50 | 100 | 250 | 500)
- (hist) Will there be significant changes to the world prior to some critical AI capability threshold being reached? [1,682 bytes]
- (hist) Finding the right primitives for spaced repetition responses [1,647 bytes]
- (hist) Nobody understands what makes people snap into AI safety [1,628 bytes]
- (hist) Big cards can be good for mathematical discovery [1,595 bytes]
- (hist) Spaced writing inbox [1,584 bytes]
- (hist) Bury effortful cards to speed up review [1,583 bytes]
- (hist) Fractally misfit [1,569 bytes]
- (hist) Prosaic AI [1,552 bytes]
- (hist) Pivotal act [1,546 bytes]
- (hist) Politicization of AI [1,474 bytes]
- (hist) Spaced repetition and cleaning one's room [1,468 bytes]
- (hist) Coherence and goal-directed agency discussion [1,457 bytes]
- (hist) Card sharing allows less valuable cards to be created [1,456 bytes]
- (hist) Late singularity [1,455 bytes]
- (hist) Laplace's rule of succession argument for AI timelines [1,451 bytes]
- (hist) AI safety lacks a space to ask stupid or ballsy questions [1,446 bytes]
- (hist) AI safety technical pipeline does not teach how to start having novel thoughts [1,446 bytes]
- (hist) AI will solve everything argument against AI safety [1,444 bytes]
- (hist) Tinkering in math requires loading the situation into working memory [1,428 bytes]
- (hist) Use paper during spaced repetition reviews [1,415 bytes]
- (hist) Simple core of consequentialist reasoning [1,404 bytes]
- (hist) Agent foundations [1,400 bytes]
- (hist) Unbounded working memory assumption in explanations [1,380 bytes]
- (hist) Thinking Mathematics [1,378 bytes]
- (hist) Kanzi [1,371 bytes]
- (hist) Quotability vs ankifiability [1,352 bytes]
- (hist) Different senses of claims about AGI [1,318 bytes]
- (hist) Setting up Windows [1,314 bytes]
- (hist) Feeling like a perpetual student in a subject due to spaced repetition [1,309 bytes]
- (hist) Do an empty review of proof cards immediately after adding to prevent backlog [1,307 bytes]
- (hist) Braid [1,301 bytes]
- (hist) Exhaustive quizzing allows impatient learners to skip the reading [1,298 bytes]
- (hist) Snoozing epicycle [1,298 bytes]
- (hist) My take on RAISE [1,298 bytes]
- (hist) Distillation is not enough [1,283 bytes]
- (hist) Tutoring heuristic for prompt writing [1,280 bytes]
- (hist) Iterated amplification [1,274 bytes]
- (hist) Creative forgetting [1,273 bytes]
- (hist) Managing micro-movements in learning [1,270 bytes]
- (hist) Encoding dependence problem [1,255 bytes]
- (hist) Dual ratings for spaced inbox [1,252 bytes]
- (hist) The mathematics community has no clear standards for what a mathematician should know [1,249 bytes]
- (hist) Busy life periods and spaced inbox [1,226 bytes]
- (hist) Ongoing friendship and collaboration is important [1,216 bytes]
- (hist) Changing selection pressures argument [1,215 bytes]
- (hist) Equivalence classes of prompts [1,210 bytes]
- (hist) Spaced repetition response as chat message or chat reaction [1,209 bytes]
- (hist) Emotional difficulties of spaced repetition [1,202 bytes]
- (hist) Newcomers in AI safety are silent about their struggles [1,196 bytes]
- (hist) Importance of knowing about AI takeoff [1,195 bytes]
- (hist) Potpourri hypothesis for math education [1,192 bytes]
- (hist) Central node trick for remembering equivalent properties [1,184 bytes]
- (hist) AI safety is harder than most things [1,156 bytes]
- (hist) Big card [1,153 bytes]
- (hist) Using spaced repetition to improve public discourse [1,138 bytes]
- (hist) Dealing with bad problems in spaced proof review [1,121 bytes]
- (hist) Different mental representations of mathematical objects is a blocker for an exploratory medium of math [1,120 bytes]
- (hist) Online question-answering services are unreliable [1,112 bytes]
- (hist) Pascal's mugging and AI safety [1,103 bytes]
- (hist) The Witness [1,103 bytes]
- (hist) One wrong number problem [1,102 bytes]
- (hist) Sudden emergence [1,101 bytes]
- (hist) The Sequences vs evergreen notes [1,097 bytes]
- (hist) List of critiques of iterated amplification [1,097 bytes]
- (hist) Discontinuities in usefulness of whole brain emulation technology [1,092 bytes]
- (hist) Spaced repetition is useful because most knowledge is sparsely applicable [1,091 bytes]
- (hist) Science argument [1,087 bytes]
- (hist) Explosive aftermath [1,081 bytes]
- (hist) Spaced inbox review should not be completionist or obligatory [1,078 bytes]
- (hist) Tao Analysis I exercise count [1,062 bytes]
- (hist) What counts as good motivation? [1,057 bytes]
- (hist) Flag things to fix during review [1,057 bytes]
- (hist) Giving advice in response to generic questions is difficult but important [1,026 bytes]
- (hist) Deck options for small cards [1,020 bytes]
- (hist) Christiano's operationalization of slow takeoff [1,018 bytes]
- (hist) Explanation science [1,009 bytes]
- (hist) Spaced repetition prevents unrecalled unrecallables [970 bytes]
- (hist) Emotional difficulties of AI safety research [950 bytes]
- (hist) Duolingo does repetition at the lesson level [938 bytes]
- (hist) You don't need to eat your own dogfood in explanation science [936 bytes]
- (hist) Fractional progress argument for AI timelines [933 bytes]
- (hist) My current thoughts on the technical AI safety pipeline (outside academia) [931 bytes]
- (hist) Open-ended questions are common in real life [916 bytes]
- (hist) Mass shift to technical AI safety research is suspicious [910 bytes]
- (hist) Braid for math [908 bytes]
- (hist) Spaced repetition as soft alarm clock [898 bytes]
- (hist) AI safety is not a community [894 bytes]
- (hist) Short-term preferences-on-reflection [890 bytes]
- (hist) Spaced proof review as a way to invent novel proofs [890 bytes]
- (hist) Scaling hypothesis [886 bytes]
- (hist) We still don't know how to systematically write great word explanations [875 bytes]
- (hist) How meta should AI safety be? [873 bytes]
- (hist) List of timelines for futuristic technologies [871 bytes]
- (hist) Spaced repetition as generator of questions [868 bytes]
- (hist) Resource overhang [867 bytes]
- (hist) Mixed messaging regarding independent thinking [866 bytes]
- (hist) Anki reviews are more fun on mobile [857 bytes]
- (hist) Comparison of sexually transmitted diseases [851 bytes]
- (hist) Discursive texts are difficult to ankify [851 bytes]
- (hist) Aligning smart AI using slightly less smart AI [846 bytes]