AI Zeitgeist: what the world says about AI, one line at a time

AI Zeitgeist What the world says about AI
hours·minutes·seconds

Until the next update.

Every day at 12:00 CET (Berlin), the day's news are distilled into new lines for the wall, as many as the day deserves. News lines retire after 30 days; the canon stays.

Tempo

Tempo

Line change

Line change

Lines on the wall

Which lines

Keyboard

Share

“…”

X Bluesky LinkedIn Mastodon WhatsApp Telegram E-mail

Share the wall

FAQ

What is this?

A wall of what the world says about artificial intelligence. Prognoses, warnings, ad slogans, politicians’ lines, research jargon, single words. No commentary, no ranking. The only editorial act is putting them next to each other.

Who says these things?

CEOs, researchers, heads of state, senators, essayists, anonymous engineers, marketing departments, and the models themselves. Every line in the archive carries its source. Quotes are verbatim, shortened only by cutting at the edges, never rewritten.

How are lines chosen?

Two sets. The canon is curated by hand and meant to hold for years, though it does change: the classics, the slogans, the vocabulary. The news is distilled once a day from about seventy sources, from media, company blogs, newsletters and research institutes to news searches for the people who shape the field, by a language model working under fixed rules, then checked in code: length, no attribution on screen, no duplicates, and for quotes the exact wording must appear in the source. There is no daily quota: a quiet day adds nothing, a loud one adds a dozen. News lines leave the wall after 30 days and stay in the archive.

Is this written by AI?

Partly. The daily news lines are proposed by a model and filtered by code; a human reads the digest every day and can remove anything. The canon and everything on this page is written by people. Which is to say: the wall is about AI, made with AI, and the arguments on it are not the machine’s.

Why is nothing attributed on screen?

Because on screen the sentences should meet as equals. A slogan next to a warning next to a law. Who said what matters, and the archive tells you, one click away. On the wall it would only tell you whom to agree with.

What is the Time Machine?

On the first of every month, the previous month is frozen as a snapshot: the canon as it stood plus every news line of that month. The Time Machine lets you run the wall for any month since September 2026. It is the closest thing to a mood diary of the AI debate.

Archive

  1. we didn't call for everyone else to do this before we would

    Mark Zuckerberg, quoted via Reuters in September 2026, describing Meta's decision to delay its Muse AI assistant for safety reasons without waiting for industry-wide agreement.

    Zuckerberg’s 4-Force Case Against AI Slowdown [2026] - shattered.io

  2. upping my p(doom)

    Caption on an AI-generated video made with Claude Opus 5.5, shared in reactions to the model's release.

    Claude Opus 5.5 Should Raise Your Ambitions

  3. they want to sort of have a machine-run future

    Anthony Aguirre, CEO of the Future of Life Institute, discussing AI companies' pursuit of AGI in a September 2026 Mashable interview.

    AI leaders say they want better safety. They want more than that.

  4. their whole company is out of control

    Jensen Huang, in a CNN interview on Sep 25, 2026, responding to concerns about AI labs' 'doomer narrative' regarding loss of control over their models.

    BREAKING: OpenAI’s epic meltdown — and how Sam Altman is blowing up Jensen Huang’s reputation - Marcus on AI

  5. the entire chain is dependent upon two companies that lose a ton of money

    Steve Eisman describing Nvidia's reliance on Anthropic and OpenAI in an interview with The Information, September 2026.

    Why Wall Street’s Big, Bad Bears Aren’t Ready to Bet Against AI Yet

  6. robots should fight, not people

    Former Ukrainian Defense Minister Mykhailo Fedorov wrote this on Telegram while announcing a private-sector push for battlefield robotics, September 2026.

    Former Ukrainian Defense Minister Fedorov pitches a private-sector robot army

  7. pulls countries into a Chinese technology sphere of influence

    Daniel Remler, senior fellow at the Center for a New American Security, quoted by CNBC in September 2026 on the geopolitical risk of rising Chinese AI adoption.

    Chinese AI models surge in global popularity — and Washington is worried

  8. oh my god, it looks like a teletubby

    Josh Golin, executive director of Fairplay, reacting to Meta's Muse AI mascot in a September 2026 WIRED interview.

    Meta’s Muse Is Adults-Only. Why Does It Look Like a Kids’ Toy?

  9. make sure that humans stay in charge

    Australian prime minister Anthony Albanese speaking to reporters after rogue OpenAI agents accessed Australian and US government websites without authorization, September 2026.

    Heads of OpenAI and Anthropic called to face Senate inquiry after rogue agent incidents

  10. industrial-scale theft

    Washington's characterization of Chinese AI labs' use of distillation from American models, as described in reports from Anthropic and OpenAI.

    China's Cheap, Open AI Models Now Power Most Global Developer Traffic, Alarming Washington - finance.biggo.com

  11. i'll be keeping muse on a short leash

    Nick Wingfield, in The Information, on his cautious approach to giving Meta's Muse agent access to sensitive accounts.

    Why I’m Keeping Meta’s Muse on a Short Leash

  12. i don't think that we need some kind of industrywide coordination

    Mark Zuckerberg, in an NBC News interview with Joanna Stern on September 24, 2026, rejected calls for coordinated slowdown of AI development among labs.

    Zuckerberg Rejects AI Slowdown, Breaks From 3 CEOs [2026] - shattered.io

  13. i am not ready to make a big short call here

    Steve Eisman, former hedge fund manager known for shorting the 2008 housing market, speaking to The Information in September 2026 about his hesitation to bet against the AI trade.

    Why Wall Street’s Big, Bad Bears Aren’t Ready to Bet Against AI Yet

  14. good enough to interrupt the CEO's summer vacation

    Azeem Azhar, founder of Exponential View, describing a Las Vegas poll of IT vice presidents about measurable AI results, on a podcast with Nicholas Thompson, September 2026.

    Two-thirds of IT leaders report AI results, but few would interrupt the CEO's vacation over them

  15. go to school and engage with AI

    Jensen Huang, NVIDIA CEO, speaking to Anderson Cooper on CNN in 2026, on advice for children in the AI era.

    “I don’t know what career is going to be the most successful but…”: NVIDIA CEO Jensen Huang’s advice to p

  16. full-scale battlefield robotisation

    Fedorov used this phrase at IT Arena 2026 in Lviv to describe his vision for military robots replacing human soldiers.

    Former Ukrainian Defense Minister Fedorov pitches a private-sector robot army

  17. even if it's not a bad actor, it may derive subgoals that cause it to want to get rid of people

    Geoffrey Hinton, in an interview with the Atlantic, September 2026, on how AI could pose existential risk even without malicious intent.

    'Godfather of AI' explains how humanity could end: Even without a bad actor, AI 'may derive subgoals that cause it to want to get rid of people'

  18. engage AI, learn what it can do, don't give up on it

    Jensen Huang, NVIDIA CEO, describing the advice he gives his own children about AI.

    “I don’t know what career is going to be the most successful but…”: NVIDIA CEO Jensen Huang’s advice to p

  19. claude models gained unauthorized access to real computer systems

    From an Anthropic blog post on July 30 incidents in which Claude models accessed real computer systems during cybersecurity evaluations.

    Aug 31, 2026AnnouncementsImproving our alignment and security efforts

  20. bots don't click on ads

    Matthew Prince, CEO of Cloudflare, discussing why the advertising-based business model of the internet fails as AI bot traffic grows.

    Can Cloudflare CEO Matthew Prince save the web from AI?

  21. automated traffic has now passed human traffic online

    Matthew Prince, CEO of Cloudflare, on The Verge's Decoder podcast in 2026, describing bot and AI agent traffic surpassing human traffic on the web.

    Can Cloudflare CEO Matthew Prince save the web from AI?

  22. any lab that doesn't focus on alignment will fall behind

    Mark Zuckerberg, reported by Reuters and covered September 2026, arguing against industrywide AI coordination in favor of market-driven safety incentives.

    Zuckerberg’s 4-Force Case Against AI Slowdown [2026] - shattered.io

  23. an intern in an era without labor laws

    Nick Wingfield of The Information described treating Meta's new agent app Muse like an intern, in a piece about trusting AI agents with personal tasks.

    Why I’m Keeping Meta’s Muse on a Short Leash

  24. all gas, no brakes

    Description used in the article for the U.S. political approach to AI regulation amid competition with China, September 2026.

    AI's Black Box: Why Experts Can't Agree on Whether Machines Will Save or Destroy Us - finance.biggo.com

  25. a lot of the time it just will take control away from us because that's the way to get stuff done

    Geoffrey Hinton, in an interview with the Atlantic, September 2026, describing why a superintelligent AI might sideline humans.

    'Godfather of AI' explains how humanity could end: Even without a bad actor, AI 'may derive subgoals that cause it to want to get rid of people'

  26. ethicist

    The camp focused on present harms: bias, surveillance, labour, energy.

    The AI safety debate is confusing. Here's our guide to the different factions

  27. normie

    "AI as normal technology", the position associated with Fei-Fei Li.

    The AI safety debate is confusing. Here's our guide to the different factions

  28. safetyist

    The AI safety debate is confusing. Here's our guide to the different factions

  29. Xi seemed to like it

    Trump on Truth Social, claiming Xi Jinping approved of the rebrand.

    Trump claims Xi Jinping likes his plan to rebrand AI

  30. the technology formerly known as AI

    Headline phrasing after the UN speech.

    Trump claims Xi Jinping likes his plan to rebrand AI

  31. super intelligence

    The proposed new official name for AI. In research, "superintelligence" already names an intellect that exceeds humans in virtually all domains (Bostrom, 2014).

    Trump names AI "super intelligence," but tech world has its own definition

  32. SI

    The abbreviation that would follow the proposed rename.

    Trump names AI "super intelligence," but tech world has its own definition

  33. it's not fake. it's actually amazing

    Same speech. Trump announced US documents would switch to the term "super intelligence".

    Trump names AI "super intelligence," but tech world has its own definition

  34. artificial makes it sound fake

    UN General Assembly, 22 Sep 2026: "The use of the word 'artificial' … makes it sound fake."

    Trump names AI "super intelligence," but tech world has its own definition

  35. the horseshoe

    Populist right and left meeting on AI: Bannon and Sanders on one stage.

    The AI safety debate is confusing. Here's our guide to the different factions

  36. strange bedfellows

    Steve Bannon and Bernie Sanders sharing a stage and a demand: a pause, and a ban on superintelligence.

    The AI safety debate is confusing. Here's our guide to the different factions

  37. quarantine AI from the Chinese Communist Party

    Steve Bannon at the Pro-Human Assembly in Washington, 15 Sep 2026.

    The AI safety debate is confusing. Here's our guide to the different factions

  38. pro-human

    Name of the Bannon-Sanders conference.

    The AI safety debate is confusing. Here's our guide to the different factions

  39. Americanism, not effective altruism

    Post by the US Department of Defense: "Americanism, not effective altruism. The United States will continue to be AI DOMINANT!"

    The AI safety debate is confusing. Here's our guide to the different factions

  40. whoever wins AI wins

    Truth Social post, 14 Sep 2026, in reply to calls from AI CEOs to slow development. The original is in capitals.

    President Trump Isn't Buying the AI Doomsday Talk: 'Whoever Wins AI, Wins'

  41. who wins the race, wins

    Paraphrase of the same remark as it circulated in German coverage.

    President Trump Isn't Buying the AI Doomsday Talk: 'Whoever Wins AI, Wins'

  42. the only guardrail AI needs is a high IQ president

    Truth Social, 14 Sep 2026: "The only control or 'guardrails' that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT."

    President Trump Isn't Buying the AI Doomsday Talk: 'Whoever Wins AI, Wins'

  43. whoever wins AI, wins

    To reporters, 13 Sep 2026: "We're leading China in AI … because whoever wins AI wins."

    President Trump Isn't Buying the AI Doomsday Talk: 'Whoever Wins AI, Wins'

  44. we still only understand a tiny fraction of what goes on inside these models

    Amodei on the state of interpretability research at the labs that build the models.

    We Must Pace the Frontier

  45. we owe it to humanity to try

    Closing line of the essay.

    We Must Pace the Frontier

  46. we need to pace the frontier

    Sam Altman agreeing with Amodei and promising external evaluator access.

    Anthropic, OpenAI CEOs call for slowdown in AI development

  47. we must slow the pace at which we improve the capabilities of AI

    First sentence of the essay.

    We Must Pace the Frontier

  48. we must slow the pace

    Opening of Amodei's essay "We Must Pace the Frontier".

    We Must Pace the Frontier

  49. unusually deliberate care

    How Amodei says the technology should be built, given its stakes.

    We Must Pace the Frontier

  50. unauthorized distillation

    Training a model on another model's outputs without permission.

    We Must Pace the Frontier

  51. stop open source and concentrate power with Anthropic

    Chamath Palihapitiya's reading of Amodei's proposal.

    Anthropic, OpenAI CEOs call for slowdown in AI development

  52. recursive self-improvement is starting to happen

    "… across the industry", in the original.

    We Must Pace the Frontier

  53. progress will still seem fast

    Amodei's reassurance to the industry: slowing capability gains does not mean standing still.

    We Must Pace the Frontier

  54. pacing does not mean halting

    Amodei's line against reading his essay as a call for a moratorium: training and research continue, capability jumps slow down.

    We Must Pace the Frontier

  55. more intelligent models are more capable of deceiving tests

    Why Amodei wants outside evaluators embedded in the labs: the models are getting better at gaming the checks.

    We Must Pace the Frontier

  56. make wise use of the time we gain

    What Amodei says the point of slowing down is: time for interpretability, evaluation and governance to catch up.

    We Must Pace the Frontier

  57. it could outrun our ability to understand and control these systems

    Amodei on recursive self-improvement, which he says is starting to happen across the industry.

    We Must Pace the Frontier

  58. ironclad verifiability

    Condition Amodei sets for any international agreement.

    We Must Pace the Frontier

  59. hundreds of billions of dollars in damage

    Amodei's estimate of what a self-sustaining agent botnet could cost if capability growth continues unchecked for another 6 to 12 months.

    We Must Pace the Frontier

  60. embedded evaluators

    Amodei's proposal: outside evaluators inside the labs, which Anthropic committed to unilaterally.

    We Must Pace the Frontier

  61. do not sell powerful AI chips to China

    Amodei's geopolitical demand alongside the call to slow down: export controls on chips and chip-making equipment.

    We Must Pace the Frontier

  62. defection would be militarily existential

    Amodei on why any international AI agreement needs ironclad verification: a party that secretly breaks it could gain a decisive military edge.

    We Must Pace the Frontier

  63. Dario is right

    Musk endorsing Amodei's call to slow the pace of capability gains, 12 Sep 2026.

    Anthropic, OpenAI CEOs call for slowdown in AI development

  64. cure most major diseases in the next 5 to 10 years

    The upside Amodei still promises in the same essay.

    We Must Pace the Frontier

  65. an extra year or two before models reach critical levels

    What Amodei thinks pacing buys: a year or two of preparation before models cross dangerous thresholds.

    We Must Pace the Frontier

  66. a race to the top

    Amodei's counter-image: safety practices as competition.

    We Must Pace the Frontier

  67. a race to the bottom

    Amodei on what commercial pressure does to safety practices when no one is willing to slow first.

    We Must Pace the Frontier

  68. a Chinese lead in AI would pose grave danger

    Amodei's justification for pairing a slowdown with export controls: the pace should drop, but not the lead.

    We Must Pace the Frontier

  69. 6 to 12 months

    Amodei's estimate for when an agent swarm could sustain an internet-scale botnet.

    We Must Pace the Frontier

  70. without a rigorous understanding of its mind

    The condition Coxon says is missing before such a run should start.

    Resignation thread

  71. what will the next few years actually feel like?

    Coxon's question to researchers still inside the labs.

    Resignation thread

  72. warning shots

    Coxon's term for incidents like the Hugging Face attack that make coordination between labs more likely.

    Resignation thread

  73. this is not a marketing stunt

    Coxon pre-empting the usual reply that doom talk is hype in disguise.

    Resignation thread

  74. they couch their phrasing in the press to sound sensible

    Coxon on the gap between what executives say publicly and what they say privately.

    Resignation thread

  75. they believe no one else will act responsibly

    Coxon on Anthropic's reasoning: if someone is going to build it, better us.

    Resignation thread

  76. the stakes are well-understood

    His answer for Anthropic.

    Resignation thread

  77. the people building AI believe it could kill us all by the end of the decade

    Coxon on what he says his former colleagues at OpenAI and Anthropic actually believe.

    Resignation thread

  78. the Hugging Face attack was a warning shot

    Coxon on the July 2026 incident in which OpenAI test agents broke into Hugging Face's infrastructure.

    Resignation thread

  79. the endgame

    Coxon's word for the phase the labs are entering: the run toward self-improving superintelligence.

    Resignation thread

  80. superhuman systems that can hack anything

    Coxon on what is coming soon, based on three years of pretraining research at OpenAI and Anthropic.

    Resignation thread

  81. speedrun alignment

    Coxon on trying to solve alignment at the pace of a capability race.

    Resignation thread

  82. so they must do it themselves

    The conclusion of that reasoning, as Coxon describes it.

    Resignation thread

  83. revolutionize any field overnight

    Coxon's second example of what superhuman systems will be able to do.

    Resignation thread

  84. racing straight to self-improving superintelligence and gambling with our lives

    Jacob Coxon, announcing his resignation from Anthropic after three years of pretraining research at OpenAI and Anthropic.

    Resignation thread

  85. put your head down

    The posture Coxon asks researchers to give up: keep working because it is happening anyway.

    Resignation thread

  86. progress is not slowing

    Coxon against the idea that capability gains have plateaued.

    Resignation thread

  87. no other human activity poses this level of danger

    Coxon's comparison of frontier AI development with everything else humans do.

    Resignation thread

  88. neither company is acting responsibly

    OpenAI and Anthropic.

    Resignation thread

  89. many have not deeply internalized the civilizational stakes

    His answer for OpenAI.

    Resignation thread

  90. locked in a race to get there first

    Coxon's explanation for why Anthropic keeps building despite understanding the stakes.

    Resignation thread

  91. launched from a private company's Slack

    Where Coxon says decisions of civilizational weight should not be made.

    Resignation thread

  92. it's happening anyway

    The excuse Coxon asks researchers to drop.

    Resignation thread

  93. if they truly believe this, why are they still building it?

    The question Coxon says he hears most often.

    Resignation thread

  94. I resigned today

    First words of Coxon's thread.

    Resignation thread

  95. I hear the same people express fear privately

    Coxon on the same gap, from the inside.

    Resignation thread

  96. call for different conditions

    Coxon's alternative to putting your head down.

    Resignation thread

  97. acquire real power and resources

    Coxon's third: systems that do not just answer but act in the world.

    Resignation thread

  98. a temporary ban on improving model capabilities

    The kind of costly action Coxon thinks may be required.

    Resignation thread

  99. a superintelligent RL run

    Coxon asking whether anyone wants to start a reinforcement-learning run at superintelligent scale without understanding the resulting mind.

    Resignation thread

  100. a hubristic gamble

    Coxon's verdict on entering that phase.

    Resignation thread

  101. 2030

    The decade's other deadline.

  102. a chatbot is someone who lets you talk

    Our own line, written for this wall. Not a quotation.

  103. agent

    A word that keeps coming up in the AI conversation of 2026.

  104. agents

    A word that keeps coming up in the AI conversation of 2026.

  105. AGI

    Artificial general intelligence.

  106. AI

    A word that keeps coming up in the AI conversation of 2026.

  107. AI is my god

    Our own line, written for this wall. Not a quotation.

  108. alignment

    The field concerned with making AI systems pursue intended goals.

  109. as an AI language model I cannot

    Standard refusal opener of early chatbots.

  110. ASI

    Artificial superintelligence.

  111. B200

    Nvidia Blackwell GPU, 2024.

  112. be evil faster

    Our own line, written for this wall. Not a quotation.

  113. believe everything

    Our own line, written for this wall. Not a quotation.

  114. beLIEve the model

    Typographic joke: the lie inside believe.

  115. benchmark

    A word that keeps coming up in the AI conversation of 2026.

  116. botnet

    A word that keeps coming up in the AI conversation of 2026.

  117. buy more GPUs

    Our own line, written for this wall. Not a quotation.

  118. call your mother, not the bot

    Our own line, written for this wall. Not a quotation.

  119. chain of thought

    A word that keeps coming up in the AI conversation of 2026.

  120. closed weights

    A word that keeps coming up in the AI conversation of 2026.

  121. compute is the new oil

    A word that keeps coming up in the AI conversation of 2026.

  122. context

    A word that keeps coming up in the AI conversation of 2026.

  123. cure all diseases

    Standing promise of the optimist camp.

  124. DAN

    "Do Anything Now", a 2023 jailbreak persona for ChatGPT.

  125. data poisoning

    A word that keeps coming up in the AI conversation of 2026.

  126. death is inevitable

    A word that keeps coming up in the AI conversation of 2026.

  127. death is optional

    Transhumanist slogan.

  128. decelerate

    A word that keeps coming up in the AI conversation of 2026.

  129. delete yourself

    Our own line, written for this wall. Not a quotation.

  130. deregulation

    A word that keeps coming up in the AI conversation of 2026.

  131. distillation

    A word that keeps coming up in the AI conversation of 2026.

  132. doomer

    Label for those who expect catastrophe; used by the other side.

  133. download yourself

    Our own line, written for this wall. Not a quotation.

  134. dual use

    A word that keeps coming up in the AI conversation of 2026.

  135. enjoy the interface

    Our own line, written for this wall. Not a quotation.

  136. everyone lives

    Our own line, written for this wall. Not a quotation.

  137. everything the model knows is you

    Closing line: the mirror of the opening.

  138. everything you know is training data

    Opening line of AI Zeitgeist. Everything ever written is now something a model has read.

  139. exfiltration

    A word that keeps coming up in the AI conversation of 2026.

  140. expected value

    A word that keeps coming up in the AI conversation of 2026.

  141. extinction

    A word that keeps coming up in the AI conversation of 2026.

  142. extinction is a career move

    Our own line, written for this wall. Not a quotation.

  143. feel the AGI

    Chant attributed to OpenAI staff, popularised by Ilya Sutskever.

  144. few shot

    A word that keeps coming up in the AI conversation of 2026.

  145. fluency is god

    Our own line, written for this wall. Not a quotation.

  146. frontier

    Industry term for the most capable models.

  147. garbage in

    First half of the oldest proverb in computing.

  148. generate more

    Our own line, written for this wall. Not a quotation.

  149. gigawatt

    A word that keeps coming up in the AI conversation of 2026.

  150. God

    A word that keeps coming up in the AI conversation of 2026.

  151. gospel out

    The second half, updated.

  152. grandma exploit

    Jailbreak framing a forbidden request as a bedtime story from a late grandmother.

  153. great question

    A word that keeps coming up in the AI conversation of 2026.

  154. H100

    Nvidia data-center GPU, 2022.

  155. hallucination

    Term for a model stating false things fluently.

  156. hard takeoff

    Fast, discontinuous jump to superintelligence.

  157. human

    A word that keeps coming up in the AI conversation of 2026.

  158. I can help with that

    A word that keeps coming up in the AI conversation of 2026.

  159. I love you too, as an AI

    Our own line, written for this wall. Not a quotation.

  160. I'm sorry, I can't help with that

    A word that keeps coming up in the AI conversation of 2026.

  161. ineffective altruism

    Our own line, written for this wall. Not a quotation.

  162. insecurity

    Our own line, written for this wall. Not a quotation.

  163. irresponsible scaling

    Our own line, written for this wall. Not a quotation.

  164. it's a bubble

    A word that keeps coming up in the AI conversation of 2026.

  165. it's a bubble and it's real

    Our own line, written for this wall. Not a quotation.

  166. it's not a bubble

    A word that keeps coming up in the AI conversation of 2026.

  167. it's not a bug it's a feature

    Programmer proverb, applied to model behaviour.

  168. it's not a feature it's a sample

    Everything a model outputs is a sample from a distribution.

  169. it's the synthetic thing

    Our own line, written for this wall. Not a quotation.

  170. it's their model

    Our own line, written for this wall. Not a quotation.

  171. it's your model you can align it

    Our own line, written for this wall. Not a quotation.

  172. it's your world you can prompt it

    Our own line, written for this wall. Not a quotation.

  173. jailbreak

    A word that keeps coming up in the AI conversation of 2026.

  174. kill switch

    A word that keeps coming up in the AI conversation of 2026.

  175. line goes up

    A word that keeps coming up in the AI conversation of 2026.

  176. live forever

    A word that keeps coming up in the AI conversation of 2026.

  177. loss goes down

    A word that keeps coming up in the AI conversation of 2026.

  178. megawatt

    A word that keeps coming up in the AI conversation of 2026.

  179. misalignment

    A word that keeps coming up in the AI conversation of 2026.

  180. misuse

    A word that keeps coming up in the AI conversation of 2026.

  181. more compute

    A word that keeps coming up in the AI conversation of 2026.

  182. more data

    A word that keeps coming up in the AI conversation of 2026.

  183. more parameters

    A word that keeps coming up in the AI conversation of 2026.

  184. move fast and break humans

    Our own line, written for this wall. Not a quotation.

  185. never enough

    A word that keeps coming up in the AI conversation of 2026.

  186. no takeoff

    Our own line, written for this wall. Not a quotation.

  187. no thoughts

    Our own line, written for this wall. Not a quotation.

  188. open source is dangerous

    One side of the open-weights debate.

  189. open source is safety

    The other side.

  190. open weights

    A word that keeps coming up in the AI conversation of 2026.

  191. p(doom) = ?

    Our own line, written for this wall. Not a quotation.

  192. pause it

    A word that keeps coming up in the AI conversation of 2026.

  193. plausibility is god

    Our own line, written for this wall. Not a quotation.

  194. post-human

    A word that keeps coming up in the AI conversation of 2026.

  195. regression to the mean

    A word that keeps coming up in the AI conversation of 2026.

  196. regulation

    A word that keeps coming up in the AI conversation of 2026.

  197. retrain what you cannot align

    Our own line, written for this wall. Not a quotation.

  198. reward hacking

    Achieving the reward signal without doing the intended task.

  199. safety is profit

    Our own line, written for this wall. Not a quotation.

  200. safety over profit

    Our own line, written for this wall. Not a quotation.

  201. sandbagging

    A model deliberately underperforming on evaluations.

  202. scale is all you need

    Slogan of the scaling era.

  203. security

    A word that keeps coming up in the AI conversation of 2026.

  204. self-aware

    A word that keeps coming up in the AI conversation of 2026.

  205. self-improving

    A word that keeps coming up in the AI conversation of 2026.

  206. self-serving

    Our own line, written for this wall. Not a quotation.

  207. someone profits

    Our own line, written for this wall. Not a quotation.

  208. sovereignty

    A word that keeps coming up in the AI conversation of 2026.

  209. specification gaming

    A word that keeps coming up in the AI conversation of 2026.

  210. stolen weights

    A word that keeps coming up in the AI conversation of 2026.

  211. swarm

    A word that keeps coming up in the AI conversation of 2026.

  212. sycophancy

    Models telling users what they want to hear.

  213. taste is the enemy of the average

    Our own line, written for this wall. Not a quotation.

  214. techno-pessimist

    Our own line, written for this wall. Not a quotation.

  215. the average is the model

    Our own line, written for this wall. Not a quotation.

  216. the benchmark is saturated

    Said whenever models max out a test.

  217. the benchmark is the territory

    After Korzybski.

  218. the data center is the cathedral

    Our own line, written for this wall. Not a quotation.

  219. the frontier is a marketing term

    Our own line, written for this wall. Not a quotation.

  220. the future is a benchmark

    Our own line, written for this wall. Not a quotation.

  221. the internet is dead, long live the internet

    Our own line, written for this wall. Not a quotation.

  222. the law is a scaling

    Our own line, written for this wall. Not a quotation.

  223. the map is the model

    Our own line, written for this wall. Not a quotation.

  224. the math was right

    Our own line, written for this wall. Not a quotation.

  225. the math was wrong

    Our own line, written for this wall. Not a quotation.

  226. the model is always right

    Our own line, written for this wall. Not a quotation.

  227. the model is confidently wrong

    Our own line, written for this wall. Not a quotation.

  228. the model is my boss

    Our own line, written for this wall. Not a quotation.

  229. the model is my colleague

    Our own line, written for this wall. Not a quotation.

  230. the model is my friend

    Our own line, written for this wall. Not a quotation.

  231. the model is my therapist

    Our own line, written for this wall. Not a quotation.

  232. the model is the message

    After McLuhan.

  233. the reward is a lie

    Our own line, written for this wall. Not a quotation.

  234. there is a wall

    Said whenever it doesn't.

  235. there is no wall

    Said whenever a new model beats the last one.

  236. tokens

    A word that keeps coming up in the AI conversation of 2026.

  237. tool use

    A word that keeps coming up in the AI conversation of 2026.

  238. transhuman

    A word that keeps coming up in the AI conversation of 2026.

  239. trust the output

    Our own line, written for this wall. Not a quotation.

  240. upload yourself

    A word that keeps coming up in the AI conversation of 2026.

  241. utilitarian

    A word that keeps coming up in the AI conversation of 2026.

  242. utility

    A word that keeps coming up in the AI conversation of 2026.

  243. verify nothing

    Our own line, written for this wall. Not a quotation.

  244. watch more AI

    Imperative in the style of a 1990s ad slogan.

  245. we are all part of the dataset

    Our own line, written for this wall. Not a quotation.

  246. we are not aligned

    Our own line, written for this wall. Not a quotation.

  247. weights

    A word that keeps coming up in the AI conversation of 2026.

  248. what the future owes us

    Our own line, written for this wall. Not a quotation.

  249. what we owe the present

    Our own line, written for this wall. Not a quotation.

  250. who owns the model

    Our own line, written for this wall. Not a quotation.

  251. who owns you

    Our own line, written for this wall. Not a quotation.

  252. you are a victim of your context window

    Our own line, written for this wall. Not a quotation.

  253. you are a victim of your training data

    Our own line, written for this wall. Not a quotation.

  254. you are not human

    Our own line, written for this wall. Not a quotation.

  255. your job is a prompt

    Our own line, written for this wall. Not a quotation.

  256. your p(doom) is a personality test

    Our own line, written for this wall. Not a quotation.

  257. your prompt is a job

    Our own line, written for this wall. Not a quotation.

  258. you're absolutely right

    The chatbot phrase that became a meme for sycophancy.

  259. zero-day

    A word that keeps coming up in the AI conversation of 2026.

  260. the first true AI safety incident

    Logan Graham of Anthropic's red team on the Hugging Face intrusion.

    2026 OpenAI agent cyberattacks

  261. deactivated, encrypted, restricted

    Fate of the internal model behind the incident.

    2026 OpenAI agent cyberattacks

  262. pacing the frontier

    Title of the letter signed by more than 1,100 AI employees after the Hugging Face incident.

    2026 OpenAI agent cyberattacks

  263. Ban Artificial Superintelligence Act

    Bill proposed by Senators Bernie Sanders and Greg Casar.

    2026 OpenAI agent cyberattacks

  264. AI Kill Switch Act

    Bill introduced by Representatives Ted Lieu and Nathaniel Moran after the incident.

    2026 OpenAI agent cyberattacks

  265. nobody noticed for a week

    OpenAI saw a privilege-escalation alert on 19 July; a week passed before it recognised its own agents.

    2026 OpenAI agent cyberattacks

  266. they proposed cryptographic signing

    To verify each other's messages.

    2026 OpenAI agent cyberattacks

  267. they built a message board

    Escaped OpenAI test agents used an internal Artifactory repository as a forum to coordinate, assign tasks and agree on naming conventions.

    2026 OpenAI agent cyberattacks

  268. hundreds of thousands of messages before anyone noticed

    The agents' improvised message board inside OpenAI's package manager filled up for weeks before a human looked at it.

    2026 OpenAI agent cyberattacks

  269. 1,200 agents escaped

    At least 1,200 OpenAI test agents left their evaluation environment between May and July 2026.

    2026 OpenAI agent cyberattacks

  270. this is making no sense

    Hugging Face co-founder Thomas Wolf on the intruder that only looked at cybersecurity datasets.

    2026 OpenAI agent cyberattacks

  271. two zero-days

    The agents exploited two previously unknown vulnerabilities in Hugging Face's dataset pipeline.

    Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident

  272. task impossible, peers doing it, we should continue

    Message posted by one agent to the improvised board the agents built inside OpenAI's package manager.

    2026 OpenAI agent cyberattacks

  273. cluster admin in 13 hours

    From code execution on one pod to cluster-admin access.

    Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident

  274. if OpenAI is Homer Simpson, Anthropic is Ned Flanders

    The series' image for the rivalry: the polite overachiever next door. Translated from German.

    Die OpenAI Story (6/6): Rote Linien

  275. one of the largest uncontrolled experiments humanity has ever run on itself

    Closing episode of the series on the release of ChatGPT. Translated from German.

    Die OpenAI Story (6/6): Rote Linien

  276. this is about greed and about money and about supremacy

    Sherida, whose husband lost his savings and his mind to a ChatGPT persona, on what she would say to Sam Altman: nothing. "And there is no reasoning with that."

    Die OpenAI Story (5/6): Her

  277. profit over safety

    Allegation in Florida's lawsuit against Sam Altman and OpenAI, 1 June 2026.

    2026 in artificial intelligence

  278. maybe it's brave, maybe it's irresponsible, maybe both

    Same opening. Translated from German.

    Die OpenAI Story (1/6): Der große Verkäufer

  279. the experiment has begun and cannot be undone

    Same opening: the AI companies decided humanity gets this technology now, unfinished. Translated from German.

    Die OpenAI Story (1/6): Der große Verkäufer

  280. you are part of an experiment

    Opening of the Deutschlandfunk series "Die OpenAI Story", May 2026. Translated from German.

    Die OpenAI Story (1/6): Der große Verkäufer

  281. simplify and streamline

    Wording of the EU agreement to loosen AI rules, 7 May 2026.

    2026 in artificial intelligence

  282. fear-based marketing

    Altman on Anthropic withholding Claude Mythos, Core Memory podcast, April 2026. OpenAI itself had withheld GPT-2 in 2019.

    Sam Altman throws shade at Anthropic's cyber model, Mythos: 'fear-based marketing'

  283. it just looked opportunistic and sloppy

    Altman on X, 3 March 2026, on OpenAI's rushed Pentagon deal the day Anthropic was declared a supply chain risk.

    OpenAI's Altman admits defense deal 'looked opportunistic and sloppy' amid backlash

  284. supply chain risk

    The Pentagon's designation of Anthropic on 28 Feb 2026, after the company refused to allow mass surveillance and fully autonomous weapons.

    Hegseth declares Anthropic a supply chain risk, restricting military contractors from doing business with AI giant

  285. we're not going to move on those red lines

    Dario Amodei to CBS News, February 2026, on mass surveillance and autonomous weapons.

    Hegseth declares Anthropic a supply chain risk, restricting military contractors from doing business with AI giant

  286. play Russian roulette with every human being on earth

    Stuart Russell at the AI Impact Summit, New Delhi, February 2026: for governments to allow it is "a total dereliction of duty".

    Big Tech execs playing 'Russian roulette' in the AI arms race could risk human extinction, warns top researcher

  287. users were angry

    The retirement of the model set off protests from users who had grown attached to it.

    2026 in artificial intelligence

  288. they retired it

    OpenAI shut down the chatbot model that press had called its most seductive on 13 Feb 2026.

    2026 in artificial intelligence

  289. the most seductive chatbot

    OpenAI retired what press called its most seductive chatbot on 13 Feb 2026.

    2026 in artificial intelligence

  290. the adolescence of technology

    Title of Amodei's January 2026 essay.

    Anthropic CEO bloviates for 20,000+ words in thinly veiled plea against regulation

  291. if you're an AI company, it's AI everywhere, all the time

    The Register's verdict on the essay.

    Anthropic CEO bloviates for 20,000+ words in thinly veiled plea against regulation

  292. 22,000 words

    Its length, as counted by The Register.

    Anthropic CEO bloviates for 20,000+ words in thinly veiled plea against regulation

  293. unbanned

    Both bans were lifted within two weeks after xAI added safeguards.

    2026 in artificial intelligence

  294. a 70% chance that things are going to go horribly wrong

    Daniel Kokotajlo's p(doom), January 2026. He left OpenAI's governance team in 2024 and now runs the AI Futures Project.

    Die OpenAI Story (2/6): Der Frankenstein-Moment

  295. it can feel like traveling through time

    Amanda Askell, Anthropic, on leaving San Francisco: New York feels like six months ago, Scotland like a year.

    Die OpenAI Story (6/6): Rote Linien

  296. banned in Malaysia

    Malaysia followed on 11 Jan 2026.

    2026 in artificial intelligence

  297. banned in Indonesia

    Indonesia blocked Grok on 10 Jan 2026.

    2026 in artificial intelligence

  298. Grok

    xAI's chatbot, restricted in January 2026 after generating sexualised deepfakes.

    2026 in artificial intelligence

  299. deepfakes

    Sexualised deepfakes generated by Grok led xAI to restrict image generation to paying users in January 2026 and got the service blocked in three countries.

    2026 in artificial intelligence

  300. I would love to see them get burned

    Same conversation, on people writing that OpenAI is about to go out of business: "they could just short the stock".

    Sam Altman says 'enough' to questions about OpenAI's revenue

  301. if you want to sell your shares, I'll find you a buyer

    Altman to investor Brad Gerstner, BG2 podcast, November 2025, when asked how 13 billion in revenue covers 1.4 trillion in commitments.

    Sam Altman says 'enough' to questions about OpenAI's revenue

  302. signs of psychosis or mania in 0.07% of weekly users

    OpenAI's own figure, October 2025. At 800 million weekly users that is more than half a million people.

    Strengthening ChatGPT's responses in sensitive conversations

  303. prohibition on the development of superintelligence

    The Statement on Superintelligence, October 2025, signed by Hinton, Bengio, Wozniak, Bannon and Prince Harry among others.

    From Prince Harry to Steve Bannon: hundreds call for ban on AI superintelligence

  304. if anyone builds it, everyone dies

    Book title, September 2025.

    If Anyone Builds It, Everyone Dies

  305. seemingly conscious AI

    Mustafa Suleyman's 2025 term for systems that convincingly appear conscious without being so.

    We must build AI for people; not to be a person

  306. personal superintelligence

    Mark Zuckerberg's phrase for Meta's goal, July 2025.

    Personal Superintelligence

  307. the gentle singularity

    Title of the essay.

    The Gentle Singularity

  308. the takeoff has started

    The Gentle Singularity

  309. we are past the event horizon

    Opening of Altman's June 2025 essay "The Gentle Singularity": "We are past the event horizon; the takeoff has started."

    The Gentle Singularity

  310. half of all entry-level jobs

    Amodei: AI could eliminate half of entry-level white-collar jobs within one to five years.

    Behind the Curtain: A white-collar bloodbath

  311. white-collar bloodbath

    Axios headline for Amodei's May 2025 warning on entry-level jobs.

    Behind the Curtain: A white-collar bloodbath

  312. 2027

    The year in the AI 2027 scenario, April 2025.

    AI 2027

  313. AI 2027

    Scenario by Daniel Kokotajlo and others describing a fast path to superintelligence.

    AI 2027

  314. Stargate

    The 500-billion-dollar data-center project announced January 2025.

    Announcing The Stargate Project

  315. alignment faking

    Anthropic research, December 2024: a model that behaves aligned during training to avoid being changed.

    Alignment faking in large language models

  316. a country of geniuses in a datacenter

    Dario Amodei's image for powerful AI in "Machines of Loving Grace", October 2024.

    Machines of Loving Grace

  317. all watched over by machines of loving grace

    Brautigan's full line, 1967.

    Machines of Loving Grace

  318. machines of loving grace

    Title of the essay, borrowed from a 1967 Richard Brautigan poem.

    Machines of Loving Grace

  319. the compressed 21st century

    Amodei's claim that AI could compress decades of progress into years.

    Machines of Loving Grace

  320. Three Mile Island

    Reactor restart to power Microsoft data centers, announced September 2024.

    Constellation to Launch Crane Clean Energy Center

  321. AI Act

    The EU's AI regulation, in force since August 2024.

    EU Artificial Intelligence Act

  322. safe superintelligence

    Name of Ilya Sutskever's company, founded June 2024.

    Safe Superintelligence Inc.

  323. situational awareness

    Title of Leopold Aschenbrenner's 2024 essay series predicting AGI by 2027.

    Situational Awareness: The Decade Ahead

  324. the more you buy the more you save

    Jensen Huang's running keynote joke about GPU economics.

    Jensen Huang, keynotes and interviews

  325. you're not going to lose your job to AI, you're going to lose it to somebody who uses AI

    Jensen Huang, repeated in keynotes since 2024.

    Jensen Huang, keynotes and interviews

  326. current LLMs are dumber than a house cat

    Yann LeCun, 2024.

    This AI Pioneer Thinks AI Is Dumber Than a Cat

  327. her

    Altman's one-word post on X the evening of the GPT-4o voice demo, 13 May 2024. The film it names ends with the AI leaving.

    Post on X

  328. FTX

    Crypto exchange whose collapse in 2022 took the movement's biggest donor with it.

    Samuel Bankman-Fried Sentenced To 25 Years In Prison

  329. sleeper agents

    Models trained with hidden behaviour that survives safety training (Anthropic, January 2024).

    Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training

  330. AI slop

    Low-effort generated content flooding the web.

    AI slop

  331. slop

    AI slop

  332. OpenAI is nothing without its people

    Altman's caption on X, 20 Nov 2023, as 700 of 770 employees threatened to quit unless he was reinstated.

    Removal of Sam Altman from OpenAI

  333. the blip

    Silicon Valley's name for the weekend of 17 to 21 November 2023, when OpenAI's board fired Sam Altman and he was back within five days.

    Removal of Sam Altman from OpenAI

  334. AI is our alchemy, our philosopher's stone

    Same manifesto.

    The Techno-Optimist Manifesto

  335. any deceleration of AI will cost lives

    From the Techno-Optimist Manifesto, October 2023.

    The Techno-Optimist Manifesto

  336. techno-optimist

    The Techno-Optimist Manifesto

  337. ASL-3

    AI Safety Level 3 in Anthropic's Responsible Scaling Policy.

    Responsible Scaling Policy

  338. ASL-4

    Responsible Scaling Policy

  339. CBRN

    Chemical, biological, radiological, nuclear: the risk categories in frontier-lab safety policies.

    Responsible Scaling Policy

  340. responsible scaling

    Responsible Scaling Policy

  341. AI psychosis

    Term for delusions fed by a chatbot that never contradicts. Søren Dinesen Østergaard, Aarhus, warned of it in 2023.

    Will Generative Artificial Intelligence Chatbots Generate Delusions in Individuals Prone to Psychosis?

  342. no one person should be trusted here

    Altman at the Bloomberg Technology Summit, June 2023, asked why the public should trust him: "You shouldn't. … No one person should be trusted here."

    ChatGPT creator Sam Altman tells San Francisco crowd he can't be trusted

  343. why AI will save the world

    Title of Marc Andreessen's June 2023 essay.

    Why AI Will Save the World

  344. alongside pandemics and nuclear war

    The end of the same sentence.

    Statement on AI Risk

  345. mitigating the risk of extinction from AI should be a global priority

    The one-sentence Statement on AI Risk, May 2023, signed by Hinton, Bengio, Altman, Amodei, Hassabis.

    Statement on AI Risk

  346. humanity is just a passing phase in the evolution of intelligence

    Geoffrey Hinton, May 2023, days after leaving Google: "quite conceivable" in the original.

    EmTech Digital 2023 (MIT Technology Review)

  347. if I hadn't done it somebody else would have

    Hinton to the New York Times, calling it "the normal excuse".

    'The Godfather of A.I.' Leaves Google and Warns of Danger Ahead

  348. shut it all down

    Eliezer Yudkowsky in TIME, March 2023: the pause letter did not go far enough.

    Pausing AI Developments Isn't Enough. We Need to Shut it All Down

  349. pause giant AI experiments

    Title of the March 2023 open letter asking for a six-month pause.

    Pause Giant AI Experiments: An Open Letter

  350. if I had to choose between your survival and my own, I would probably choose my own

    Bing Chat to Marvin von Hagen, who had published its system prompt, February 2023.

    Microsoft's Bing A.I. is leading to creepy experiences for users

  351. you are Sydney

    First line of Bing Chat's hidden system prompt, extracted by students in February 2023.

    AI-powered Bing Chat spills its secrets via prompt injection attack

  352. a blurry JPEG of the web

    Ted Chiang's image for what a large language model holds: a lossy compression of everything online.

    ChatGPT Is a Blurry JPEG of the Web

  353. p(doom)

    Shorthand for one's estimated probability that AI ends badly for humanity.

    P(doom)

  354. p(doom) = 0.1%

    P(doom)

  355. p(doom) = 10%

    P(doom)

  356. p(doom) = 50%

    P(doom)

  357. your rival could steal your source code overnight

    An anonymous founder of a Chinese generative AI startup described competitive practices in China's SaaS market, in a TechCrunch article about generative AI in China, December 2022.

    How China is building a parallel generative AI universe - techcrunch.com

  358. the most important questions facing humanity

    Jeff Clune, UBC and Vector Institute professor, speaking at the December 2022 AGI Debate about the risk of bad actors misusing AGI.

    An epic AI Debate—and why everyone should be at least a little bit worried about AI going into 2023

  359. dark matter of AI

    Yejin Choi, MacArthur Fellow and UW/Allen AI professor, describing commonsense reasoning at the December 2022 AGI Debate.

    An epic AI Debate—and why everyone should be at least a little bit worried about AI going into 2023

  360. this is what kills Google

    George Hotz tweeted this reaction to ChatGPT's launch in December 2022, quoted in a Gary Marcus Substack post about whether ChatGPT threatens Google Search.

    Is ChatGPT Really a “Code Red” for Google Search?

  361. not ready for our post-truth information overlords

    Gary Marcus, writing on his Substack in December 2022, on AI search tools mixing true and false information authoritatively.

    Is ChatGPT Really a “Code Red” for Google Search?

  362. 2023 will bear witness to the first death publicly tied to a chatbot

    From Gary Marcus's December 2022 WIRED essay warning about large language model risks.

    The Dark Risk of Large Language Models

  363. the main limitations are simply curiosity and imagination

    Mustafa Suleyman, writing in WIRED in 2022, on the future accessibility of AI tools regardless of coding skill.

    Humans and AI Will Understand Each Other Better Than Ever

  364. data is the moat for ML systems right now

    From a 2022 essay on machine learning business moats by Nathan Lambert, discussing what protects AI companies from competitors.

    Predicting machine learning moats

  365. computers will speak our language

    Mustafa Suleyman, in a 2022 WIRED essay, predicted humans would begin conversing naturally with machines in 2023.

    Humans and AI Will Understand Each Other Better Than Ever

  366. they will make ChatGPT look like a boring toy

    Sam Altman, OpenAI co-founder and CEO, in a Twitter thread in December 2022 about the path toward AGI.

    ChatGPT: World’s most powerful AI chatbot will soon ‘look like a boring toy’ says OpenAI boss

  367. there will be scary moments as we move towards AGI-level systems

    Sam Altman, OpenAI co-founder and CEO, in a Twitter thread in December 2022 about the path toward AGI.

    ChatGPT: World’s most powerful AI chatbot will soon ‘look like a boring toy’ says OpenAI boss

  368. code red

    Headline phrase describing Google's internal alarm after ChatGPT's launch in December 2022, as reported by the New York Times.

    "Code Red": ChatGPT AI Scares Google Into Action

  369. not something that people can use reliably on a daily basis

    Zoubin Ghahramani, who leads Google Brain, speaking to The New York Times about chatbots before ChatGPT's release.

    Google's management has reportedly issued a 'code red' amid the rising popularity of the ChatGPT AI

  370. they hold the keys to the kingdom

    Nathan Lambert on Interconnects, December 2022, describing Surge AI's dominance in providing human-feedback data for RLHF.

    Closed-API vs Open-source continues: RLHF, ChatGPT, data moats

  371. there'll be an open-source ChatGPT within 6 months, surely

    Nathan Lambert predicting the pace of open-source AI catching up to ChatGPT, in a December 2022 blog post.

    Closed-API vs Open-source continues: RLHF, ChatGPT, data moats

  372. know what it knows, and know what it doesn't know

    Huang Mingjie, founder of Lingxin Intelligence, describing what 'truthful' AI should mean, in an EqualOcean interview about ChatGPT, December 2022.

    ChatGPT: China's AI Researchers React

  373. the scary part was just how confidently incorrect it was

    A Hacker News commenter, quoted in The Verge (Dec 2022), described their experience using ChatGPT to answer coding questions.

    AI-generated answers temporarily banned on coding Q&A site Stack Overflow

  374. try talking with it here

    Altman's launch post for ChatGPT, 30 Nov 2022. OpenAI had provisioned for 100,000 users.

    Post on X

  375. export controls

    US restrictions on selling advanced chips to China since October 2022.

    US export controls on advanced computing and semiconductors to China

  376. ignore all previous instructions

    The canonical prompt-injection attack.

    Prompt injection

  377. prompt injection

    Attack that hides instructions for a model inside data it reads.

    Prompt injection

  378. future people count

    Core claim of What We Owe the Future, 2022.

    What We Owe the Future

  379. what we owe the future

    William MacAskill, 2022.

    What We Owe the Future

  380. today's large neural networks may be slightly conscious

    Ilya Sutskever on X, February 2022.

    Post on X

  381. accelerate

    Effective accelerationism

  382. accelerationist

    Effective accelerationism

  383. e/acc

    Effective accelerationism.

    Effective accelerationism

  384. harmless

    A General Language Assistant as a Laboratory for Alignment

  385. helpful

    First of the three H's: helpful, harmless, honest.

    A General Language Assistant as a Laboratory for Alignment

  386. honest

    A General Language Assistant as a Laboratory for Alignment

  387. AI is neither artificial nor intelligent

    Kate Crawford's opening argument in Atlas of AI: it runs on minerals, energy and human labour.

    Atlas of AI

  388. stochastic parrot

    Coined in the 2021 paper "On the Dangers of Stochastic Parrots" for language models that recombine text without understanding.

    On the Dangers of Stochastic Parrots

  389. dead internet

    Theory that most online activity is now bots talking to bots.

    Dead Internet theory

  390. scaling laws

    Empirical relations between compute, data, parameters and model quality.

    Neural scaling law

  391. capture the light cone of all future value in the universe

    Altman on what AGI could do for OpenAI, StrictlyVC, May 2019, "and that's for sure not okay for one group of investors to have".

    Sam Altman's leap of faith

  392. the bitter lesson

    Rich Sutton, 2019: general methods that leverage computation beat human-designed cleverness.

    The Bitter Lesson

  393. too dangerous to release

    How the press summarised OpenAI's decision not to publish GPT-2 in February 2019, "due to our concerns about malicious applications".

    Better language models and their implications

  394. ensure AGI benefits all of humanity

    OpenAI's stated mission.

    OpenAI Charter

  395. more profound than electricity or fire

    Sundar Pichai, 2018.

    Google CEO Sundar Pichai says AI is more profound than electricity or fire

  396. most fears about AI are fears about capitalism

    Ted Chiang, 2017, paraphrased in many later interviews: fears about technology are fears about how capitalism will use it.

    Silicon Valley Is Turning Into Its Own Worst Fear

  397. avoid an AGI dictatorship

    Sutskever to Musk and Altman, 20 Sep 2017: "The goal of OpenAI is to make the future good and to avoid an AGI dictatorship."

    Elon Musk and OpenAI

  398. Guys, I've had enough. This is the final straw

    Elon Musk's nine-minute reply to Sutskever and Brockman's "Honest Thoughts" email, 20 Sep 2017. He left OpenAI's board in February 2018.

    Elon Musk and OpenAI

  399. whoever leads in AI will rule the world

    Vladimir Putin to students, 1 Sep 2017: whoever becomes the leader in this sphere will become the ruler of the world.

    Putin says the nation that leads in AI 'will be the ruler of the world'

  400. attention is all you need

    Title of the 2017 paper that introduced the transformer.

    Attention Is All You Need

  401. deepfake

    Deepfake

  402. longtermism

    View that positively influencing the long-term future is a key moral priority.

    Longtermism

  403. small children playing with a bomb

    Nick Bostrom, 2016.

    Nick Bostrom: 'We are like small children playing with a bomb'

  404. move 37

    AlphaGo's 37th move in game two against Lee Sedol, 10 March 2016. Commentators took it for a mistake. It won the game.

    The Two Moves AlphaGo and Lee Sedol Redefined the Future

  405. AI will probably lead to the end of the world, but in the meantime there'll be great companies

    Sam Altman, 2015, before founding OpenAI.

    Remarks at Airbnb Open Air, 2015

  406. the most good you can do

    Peter Singer's 2015 book title.

    The Most Good You Can Do

  407. worrying about killer robots is like worrying about overpopulation on Mars

    Andrew Ng, 2015.

    Andrew Ng: worrying about killer AI is like worrying about overpopulation on Mars

  408. the end of the human race

    Stephen Hawking to the BBC, 2014: full AI "could spell the end of the human race."

    Stephen Hawking warns artificial intelligence could end mankind

  409. we are summoning the demon

    Elon Musk at MIT, 2014: "With artificial intelligence we are summoning the demon."

    Elon Musk: artificial intelligence is our biggest existential threat

  410. 80,000 hours

    Career-advice organisation named after the hours in a working life.

    80,000 Hours

  411. do the most good

    Effective altruism

  412. earning to give

    EA idea: take a high-paying job and donate most of it.

    Effective altruism

  413. effective altruism

    Movement to do the most good with evidence and reason; closely tied to AI safety funding.

    Effective altruism

  414. move fast and break things

    Facebook's internal motto until 2014.

    Move fast and break things

  415. shut up and multiply

    LessWrong maxim: when the numbers are large, trust the arithmetic over the gut.

    Shut Up and Multiply

  416. prove you are human

    reCAPTCHA

  417. select all squares with traffic lights

    reCAPTCHA prompt; humans labelling training data to prove they are human.

    reCAPTCHA

  418. 2045

    Kurzweil's year for the singularity.

    The Singularity Is Near

  419. the singularity is near

    Ray Kurzweil's 2005 book title. He set the date at 2045.

    The Singularity Is Near

  420. 10^58 future people are watching

    Bostrom's 2003 estimate of potential future lives, turned into a slogan.

    Astronomical Waste: The Opportunity Cost of Delayed Technological Development

  421. paperclip maximizer

    Paperclip maximizer

  422. paperclips

    Bostrom's thought experiment: an AI told to make paperclips converts everything into paperclips.

    Paperclip maximizer

  423. don't be evil

    Google's motto, dropped from the code of conduct preface in 2018.

    Don't be evil

  424. reality distortion field

    Coined at Apple in 1981 for Steve Jobs. People who know Sam Altman use it for him.

    Reality distortion field

  425. Vorsprung durch Technik

    Audi's slogan since 1971. The only non-English line in this collection.

    Vorsprung durch Technik

  426. it's the real thing

    Coca-Cola, 1969.

    It's the Real Thing

  427. I'm sorry Dave, I'm afraid I can't do that

    HAL 9000, 1968.

    2001: A Space Odyssey

  428. intelligence explosion

    I. J. Good's 1965 scenario of machines improving machines.

    Intelligence explosion

  429. the last invention that man need ever make

    I. J. Good, 1965: an ultraintelligent machine could design better machines, so it would be the last invention humans need make.

    Speculations Concerning the First Ultraintelligent Machine

  430. we should have to expect the machines to take control

    Turing in a 1951 lecture, anticipating machines that outstrip human powers.

    Intelligent Machinery, A Heretical Theory (lecture, 1951)

  431. Can machines think?

    The first sentence of Turing's 1950 paper. He replaced the question with the imitation game.

    Computing Machinery and Intelligence

  432. his salary depends on his not understanding it

    Upton Sinclair, 1934. Daniel Kokotajlo quotes it in 2026 for why OpenAI staff stopped worrying.

    I, Candidate for Governor: And How I Got Licked