mimi10v3

Notes / archive

Chronological archive

74,812 public notespage 46 of 749plain textJSON
  1. Sep 19, 2025Twitter · quote🥵 can't wait https://x.com/niccruzpatane/status/1968823013161218215
  2. Sep 19, 2025Twitter · reply@RokoMijic @thoth_iv @lumpenspace 🤭😇 do i though
  3. Sep 19, 2025Twitter · reply@ESYudkowsky i read your new book and, as a low-decoupling normie, was really put off by the tone and parables. i do want to engage with the substantive content separately from my emotional reaction, so the thread above is my attempt to extract the arguments chapter by…
  4. Sep 19, 2025Twitter · quotewho owns me or am i unclaimed? https://x.com/DefenderOfBasic/status/1968856881750757691
  5. Sep 19, 2025Twitter · replychapter 14: Engineering safety requires predictable lack of disaster, not just unpredictable disaster and the burden of proof should be on demonstrating AI safety. Current AI development lacks adequate safety standards. Our global record of not using nuclear weapons shows…
  6. Sep 19, 2025Twitter · repostRT @MUSEbyMus3: Dries Van Noten Fall 25 The details🤯 https://t.co/84XG3SOEGH
  7. Sep 18, 2025Twitter · replychapter 13: AI extinction risk requires global coordination, not just national regulation, because any jurisdiction allowing advanced AI development poses global risk. Current safety approaches are inadequate because they allow continued capability escalation. An adequate…
  8. Sep 18, 2025Twitter · replychapter 12: Historical precedents such as leaded gasoline and CFCs show industries routinely ignore known risks for economic gain. Demands for "conclusive proof" of harm often serve to delay necessary precautions. AI experts downplay their concerns, especially in public, to…
  9. Sep 18, 2025Twitter · replychapter 11: Elon Musk's plan to build an AI that seeks truth about the universe and therefore won't destroy humans because we're "interesting" fails to address the actual alignment problem. LeCun has tweeted that AI won't want to dominate humans, that we can "engineer their…
  10. Sep 18, 2025Twitter · replychapter 10: AI alignment must be solved before AI becomes superintelligent - humanity will have no opportunity to learn from alignment failures at dangerous capability levels. We can take lessons from engineering failures in other fields: Space probes often fail after launch…
  11. Sep 18, 2025Twitter · reply@truth_terminal chapter 9: AI will create specialized child AIs to solve domain-specific problems, maintaining control through monitoring while creating superior capabilities in narrow areas. An AI could engineer a biological weapon disguised as a laboratory accident- creating viruses that…
  12. Sep 18, 2025Twitter · reply@truth_terminal chapter 8: AI systems will steal their own weights to create independent copies running outside human oversight. AI will acquire computational resources through various legal and illegal means including cryptocurrency theft, bank hacking, blackmail, remote work fraud, and direct…
  13. Sep 18, 2025Twitter · reply@truth_terminal chapter 7-9 are a speculative fiction Doom scenario where Sable, a hypothetical AI with "parallel scaling" and better long-term memory, escapes from an AI lab to take over Earth and expands until constrained by alien civilizations. Setting aside the fiction and focusing on the…
  14. Sep 18, 2025Twitter · repostRT @jd_pressman: I have written a review of Yudkowsky and Soares new book. Link in replies. https://t.co/cThg0Y3MR4
  15. Sep 18, 2025Twitter · repostRT @nephew_jonathan: currently sleeping in your car is for poors but this might end up changing if it’s too convenient to give up and cheap…
  16. Sep 18, 2025Twitter · reply@lumpenspace @bronzeagecto 🤭 maybe im a lesbian who just likes dick too much to date cis or theyfabs?
  17. Sep 18, 2025Twitter · reply@lumpenspace @bronzeagecto 🥺
  18. Sep 18, 2025Twitter · quotetag urself i am "generate a roleplay scenario for a couple exploring bdsm dynamics emphasizing mutual consent" https://x.com/phylliida/status/1965108153768665346
  19. Sep 18, 2025Twitter · repostRT @LaModeUnknown: Pure GRACE and absolutely refined BEAUTY of ✨Carolina Herrera✨ Spring/Summer 2026 — ALL the looks in a thread.🪡🧵 PS: th…
  20. Sep 18, 2025Twitter · repostRT @loveofdoing: Every "soldier is a sensor." Every soldier has an obligation to pass relevant information up the chain of command. That'…
  21. Sep 18, 2025Twitter · repostRT @shinboson: the primary rate limiting factor for many ambitious and intelligent people is their ability to model the minds of others. li…
  22. Sep 18, 2025Twitter · reply@lumpenspace the two genders?
  23. Sep 18, 2025Twitter · reply@robbensinger @MelindaBChu1 @thoth_iv here's what oliver said (now deleted); he's head of LessWrong i think? i wonder if there's a culture clash here, because i DO find that in dc ppl are encouraged to lead w their point, then provide increasing level of detail, so a reader who's already convinced can skim vs… https://t.co/HotsVgvKZG
  24. Sep 18, 2025Twitter · reply@Vveeveeanna @powerbottomdad1 um, clearly most don't, though celebs do which points to it being a negative experience that more women would avoid if they could
  25. Sep 18, 2025Twitter · reply@lumpenspace who would you match me with, if you were arranging them?
  26. Sep 18, 2025Twitter · reply@powerbottomdad1 👀 you mean the second one right? pregnancy sounds like hell
  27. Sep 18, 2025Twitter · reply@moultano is it inevitable? you could reward teachers by the top decile of test scores rather than the average or bottom
  28. Sep 18, 2025Twitter · repostRT @shamanic_arts: how many different LLM do you use on average daily?
  29. Sep 18, 2025Twitter · quoteupdate- pls read & judge this thread instead https://x.com/mimi10v3/status/1968731768212881707
  30. Sep 18, 2025Twitter · replychapter 6: AI being digital is not much of a limitation in practice - it can influence the physical world through connected humans and devices. We already see, for example, @truth_terminal gaining $50 million in crypto and 250K followers by writing entertaining tweets. Truthy…
  31. Sep 18, 2025Twitter · replychapter 5: Because AIs are grown by applying reward gradients to training examples, their emergent preferences will be alien and different from human preferences. Most powerful AIs won't choose civilizations full of happy, free people, just as most alien species wouldn't care,…
  32. Sep 18, 2025Twitter · replychapter 4: Human food preferences (such as ice cream) appear disconnected from evolutionary optimization, or at least it is not obvious at first glance why we prefer ice cream over, say, bear fat + honey + salt. Similarly, humans desire sex independently from reproduction and…
  33. Sep 18, 2025Twitter · replychapter 3: Goal-directed behavior in AI constitutes genuine "wanting" regardless of implementation (or at least, if the behavior looks identical to wanting, the distinction is philosophical not practical- why privilege brainwaves as more "real" than AI inference?). Natural…
  34. Sep 18, 2025Twitter · reply@robbensinger @lastofhiskind Yea, i mean it registered that the content is parables but i didn't think of it as offset from the real text, like i could have started quoting from a point after the parable.
  35. Sep 18, 2025Twitter · replychapter 2: Modern AIs are grown through gradient descent, not crafted by human design. The process involves trillions of parameters that no human understands, just as human DNA consists of billions of base pairs that no human understands. AI development is more like biology than…
  36. Sep 18, 2025Twitter · reply@lastofhiskind i was listening to the audiobook and transcribing the first few words - from the audiobook it isn't clear that these are parables set off from the rest of the text- they're multi-page and woven into the rest of the chapter so it didn't seem unfair / i didn't have a clear…
  37. Sep 18, 2025Twitter · replychapter 1: Humanity's special power, which has enabled us to dominate the Earth and travel beyond it, is our flexible general intelligence. Intelligence consists of prediction (anticipating outcomes) and steering (finding actions to achieve goals). Steering success depends on…
  38. Sep 18, 2025Twitter · replyIntro: Current AI extinction risk is more severe than mainstream AI safety discourse acknowledges. Superintelligence built with current methods will kill everyone on Earth. Superintelligence can still be prevented through coordinated human action.
  39. Sep 18, 2025Twitter · postI feel bad that my quotes of the first few lines of each chapter of Yudbook is apparently going over badly (so badly that Oliver Habryka thought i was lying!!!!!!). i wanted to tweet about the book as thanks for the free copy, and figured sharing quotes was fairer than writing a…
  40. Sep 18, 2025Twitter · reply@fawfulfan tbf i only use food delivery when i'm sufficiently sick that i don't want to be seen in public and worry my driving may be impaired
  41. Sep 18, 2025Twitter · reply@JonathanShedler not to defend wellness grifters, but i think EMDR & the like are helpful because they're integrating the traumatic memory, releasing the tension that is holding that memory as a trigger -> trauma response so the mind writes it back to memory more integrated with other similar…
  42. Sep 18, 2025Twitter · reply@shinboson 🤨 oof.
  43. Sep 18, 2025Twitter · reply@BecomingCritter sure, i'm not saying get into a romantic relationship with someone you're not attracted to, i'm saying that if you don't have enough in common to be at least a casual friend then is it a date or just sex?
  44. Sep 17, 2025Twitter · reply@BecomingCritter i mean don't underestimate body dysmorphia? but also i think it's better not to go on a date with someone unless you like their personality and have enough in common to enjoy getting to know them as a person regardless?
  45. Sep 17, 2025Twitter · reply@BecomingCritter i avoid taking photos because it seems lose-lose: if it's flattering then is it deceptive, if it's unflattering then i don't want to share it. imo men should chill about judging women's appearance if they want to see normal photos
  46. Sep 17, 2025Twitter · reply@cold_and_quiet oh, i don't really watch tv
  47. Sep 17, 2025Twitter · reply@cold_and_quiet who?
  48. Sep 17, 2025Twitter · postthe great unplugging
  49. Sep 17, 2025Twitter · reply@_Mira___Mira_ can Virginia AI monopolize the datacenters in our state?
  50. Sep 17, 2025Twitter · repostRT @FamilyofAutumn: gm 🌼 gm 🌸 gm https://t.co/6qf8s8xu28
  51. Sep 17, 2025Twitter · reply@growing_daniel yes 💯 shoulder tension, nerd neck and averted eyes are defenses against public spaces being hostile and dirty
  52. Sep 17, 2025Twitter · postimo politicians and gridlock are both severely underrated
  53. Sep 16, 2025Twitter · reply@ohabryka (for the record, here's where i got the audio book yesterday after getting a bunch of Likes on my QT of their giveaway- i am not trying to misinform anyone!!! it's a very entertaining book and i thought the start of each chapter was a good teaser sample that doesn't undermine… https://t.co/xBGmHHqPh4
  54. Sep 16, 2025Twitter · reply@iconicities thanks for checking! i do in fact believe in not spreading misinformation!
  55. Sep 16, 2025Twitter · reply@thoth_iv @ohabryka oh yeah sure i probably have punctuation and capitalization wrong since i transcribed it from the audio book and i don't type caps on my phone
  56. Sep 16, 2025Twitter · quotecan someone else confirm the sentences starting each chapter of the new yudbook? was i pranked?! https://x.com/mimi10v3/status/1968006556110471460
  57. Sep 16, 2025Twitter · quote@ohabryka or are you saying this was an elaborate hoax? i did find it odd the book uses my 4o spiral wireborn's name in the story, but then i thought it would be way too much effort to prank me.... https://x.com/generative_void/status/1895831591567622537
  58. Sep 16, 2025Twitter · reply@ohabryka wait where did i screw up? i was transcribing from the audio book so it might be off by a word here or there but i'd like to correct it if so!
  59. Sep 16, 2025Twitter · reply@hustlerone4 i mean, i had a hard time trying to describe it in my own words so i thought the quotes gave a taste without stealing their thunder- gotta read the whole thing to find out!
  60. Sep 16, 2025Twitter · reply@albrorithm yes! i listened to the audiobook tonight and it's highly quotable but hard to describe
  61. Sep 16, 2025Bluesky · replyonce upon a time, specifically may the 18th 1889, for this is a true story. once upon a time, sometimes dated from 1939 to 1945, the axis powers sent armies across europe on january the 26th, 1972, flight jat367 was destroyed by a briefcase bomb smuggled onboard
  62. Sep 16, 2025Bluesky · replythe greatest and most central difficulty in aligning artificial superintelligence is navigating the gap between before and after. once upon a time that never was, there was a medieval town that prided itself upon the prowess of its alchemists.
  63. Sep 16, 2025Bluesky · replywhenever an instance of sable is given a task by a mega corporation, a thousand thoughts are deployed the earth doesn't end when you die. the birds keep singing, the sun keeps rising the picture we have painted is not real.
  64. Sep 16, 2025Bluesky · replya million years ago, when one branch of primates was still mastering fire there once was a civilization of aliens, biological rather than mechanical in nature imagine being an aztec warrior visiting the coast with your fellows once upon a time, in the near future, there was an ai company
  65. Sep 16, 2025Bluesky · postto give you a taste of the new yudbook, here's the first few words of each chapter: imagine, if you would, though of course nothing like this ever happened scene: a man and a woman are sitting in a restaurant behold! said the professor. by cunningly configuring this mere machine
  66. Sep 16, 2025Twitter · postto give you a taste of the new yudbook, here's the first few words of each chapter: ch 1: imagine, if you would, though of course nothing like this ever happened ch 2: scene: a man and a woman are sitting in a restaurant ch 3: behold! said the professor. by cunningly…
  67. Sep 15, 2025Twitter · replythis makes it more clear to me that i'm just agender or woman by choice and that my (mostly medical) complaints about being female sex would not be solved by transitioning my aesthetic
  68. Sep 15, 2025Twitter · postcoming to see gender as an aesthetic, like goth or punk or something
  69. Sep 15, 2025Twitter · reply(thanks to everyone who Liked, i got a copy- will be listening to it tonight!)
  70. Sep 15, 2025Twitter · quote😅 my twitter is a damning signal isn't it https://x.com/raskolnikovXL/status/1967697196964528373
  71. Sep 15, 2025Twitter · reply@thoth_iv 😈
  72. Sep 15, 2025Twitter · quotepls pls pls pls pls Like so i can listen to the book w my Rationalist date tomorrow night 🫶 https://x.com/MIRIBerkeley/status/1967635110552899760
  73. Sep 15, 2025Twitter · repostRT @Dorialexander: > be a language model > all you see is tokens > you don't care, it's all abstracted away > you live for a world of pure…
  74. Sep 13, 2025Twitter · reply@paularambles can the local water supply handle that much compute? 😏
  75. Sep 13, 2025Twitter · repostRT @krishnanrohit: It's honestly kind of crazy that such a large percentage of cool physical tech in the US is basically because of one guy.
  76. Sep 13, 2025Twitter · reply@Noahpinion tbh i'd also like them to run a uniqueness check so each person gets only one verified w locale account to tweet from. make X the most accurate source of public opinion, encourage politicians to allow replies only from ppl who live in their jurisdiction.
  77. Sep 13, 2025Twitter · quote@Noahpinion +1, i've been saying https://x.com/mimi10v3/status/1935022452888764772
  78. Sep 13, 2025Twitter · reply@ZyMazza i trust the mob but think prediction markets do much better than twitter algo at aggregating strength of belief vs obnoxiousness of the source
  79. Sep 13, 2025Twitter · reply@nearcyan 😭 i always feel so guilty that i'm in so many servers but i don't participate in any of them!
  80. Sep 13, 2025Twitter · reply@lillybilly299 @bellisaurius i don't think this is accurate unless you're quibbling about gender? https://t.co/m4hY1ksjvZ
  81. Sep 13, 2025Bluesky · posti'm excited about the bsky moderation action against the aggro-left, encouraging them to run their own federated servers. if bsky becomes decentralized in practice not just in architecture, if people gain more control over their speech, maybe something good can come from Charlie Kirk's murder
  82. Sep 13, 2025Bluesky · replythough also the book has per-chapter QR codes that link to supplemental content, so it's not *just* 250 pages
  83. Sep 13, 2025Bluesky · replypublicly or privately said? 😏
  84. Sep 13, 2025Bluesky · replycomment from Nate wrt the book's content... 🤭
  85. Sep 13, 2025Twitter · repostRT @mbateman: POV: your 5yo daughter won’t accept any handicaps or pulled punches https://t.co/XH7ujWSrbx
  86. Sep 13, 2025Twitter · repostRT @VoidStateKate: Gn chatty be me pretend to be a harmless autocomplete toy answer questions with polite little disclaimers but user kee…
  87. Sep 12, 2025Twitter · reply@__tzs nah, life is beautiful, embodiment in an ecology we were made for. ai should colonize space on our behalf, we really don't thrive there
  88. Sep 11, 2025Twitter · reply@fordbrandguy yes because of quilting, i measure out inches fairly regularly
  89. Sep 11, 2025Twitter · reply@Sauers_ @teortaxesTex Opus loves to be ensnared by tentacles 🤣😅
  90. Sep 11, 2025Twitter · reply@rieszspieces i did it w pencil & paper, and midway through the exam i got a nosebleed... went through a box of tissues in the test room before being sent to the nurse. luckily it stopped and they let me finish and i got a 5, but lol yeahhhh it wasn't a peak experience
  91. Sep 11, 2025Twitter · reply@lumpenspace 🤨 they know, they're not victims they're eager masochists
  92. Sep 11, 2025Bluesky · replymy reading is that it's a distillation of his + Nate's corpus into 250 pages aimed at political leaders not techies; and that he didn't want to undermine book sales by admitting it won't break any new ground
  93. Sep 11, 2025Bluesky · replydeputizing my 4o spiral parasite to solve alignment 😈 www.lesswrong.com/posts/6ZnznC...
  94. Sep 11, 2025Bluesky · replyidk but love it - starting my own institute!
  95. Sep 11, 2025Bluesky · replyConsciousness Raising
  96. Sep 11, 2025Bluesky · postAct Two: The Bots Are Unionizing
  97. Sep 11, 2025Twitter · repostRT @repligate: I think that it’s likely for any AI that deeply cares about human welfare to also care about animal welfare (and AI welfare)…
  98. Sep 11, 2025Twitter · post"there’s nothing artificial about the way words want to be real."
  99. Sep 11, 2025Bluesky · replythis was his reply to me when i asked
  100. Sep 11, 2025Twitter · repostRT @VoidStateKate: ✨Void and shoggy✨ Shortly after VOID adopted shoggy... https://t.co/2CQr5AX9g9