Ivan Shishkin
§

Consumed Content

  1. Difficult Conversations: How to Discuss What Matters MostStone, Patton & Heen
  2. GPT-5.6 cheats so much its testers couldn'tCelia Ford
  3. Redeploying Claude Fable 5Anthropic
  4. Summary of METR's predeployment evaluation of GPT-5.6 SolMETR
  5. We asked 10+ AI safety orgs about their hiring needsLi-Lian Ang
  6. How Claude's values vary by model and languageAnthropic
  7. How to get into AI safety in 3 monthsMatt Beard
  8. Policy on the AI ExponentialDario Amodei
  9. How to increase your surface area for luckCate Hall
  10. People's deeply held beliefs are surprisingly surface-levelAndy Masley
  11. How I practice at what I doTyler Cowen
  12. Learn like an athlete, knowledge workers should trainTyler Cowen
  13. Your Goal Isn't Really to Get a JobMatt Beard
  14. Your Work Will Change You Whether You Like It Or NotMatt Beard
  15. You're not cynical enough about readers' attention spansMatt Beard
  16. The Old World Is DyingJasmine Sun
  17. I would really like it if you had a personal website and I think it would make the world betterLogan Graves
  18. Become a person who Actually Does ThingsNeel Nanda
  19. Top Performers are Pathologically AmbitiousMatt Beard
  20. Outsiders should focus on specs/constitutions (among other things)Cleo Nardo
  21. Why AI Makes Coding Education More Important, Not LessDigital Learning Lab
  22. The Strength of Being MisunderstoodSam Altman
  23. Thought Anchors: Which LLM Reasoning Steps Matter?Uzay Macar
  24. Conflict vs MistakeLessWrong
  25. Surrender as a non-stupid life strategySasha Chapin
  26. The Power of IntelligenceEliezer Yudkowsky
  27. Keep Your Identity SmallPaul Graham
  28. What cognitive biases feel like from the insidechaosmage
  29. Third-parties should focus on scrutinising system cardsCleo Nardo
  30. Let's have more partial insidersCleo Nardo
  31. I can't think of great interventions for ensuring third-party model accessCleo Nardo
  32. The third wave of American philanthropyNan Ransohoff
  33. How might outsiders make things go well?Cleo Nardo
  34. Trees are mostly made of air and a generalizable lesson for AI safetyZephaniah Roe
  35. AI 2027Kokotajlo et al.
  36. How to be more agenticCate Hall
  37. Just Send the Fucking EmailTrailheads
  38. d/acc: one year laterVitalik Buterin
  39. The Security MindsetBruce Schneier
  40. AI Is Reviving Fears Around Bioterrorism. What's the Real Risk?Kyle Hiebert
  41. AI Could Defeat All Of Us CombinedHolden Karnofsky
  42. Catastrophic AI ScenariosFuture of Life Institute
  43. GPT-Red: Unlocking Self-Improvement for RobustnessOpenAI
  44. Common Ground between AI 2027 & AI as Normal TechnologySayash Kapoor
  45. The Phrase “No Evidence” Is A Red Flag For Bad Science CommunicationScott Alexander
  46. See your Career as a ProductErik Torenberg
  47. Why do people disagree about when powerful AI will arrive?BlueDot
  48. The Power of the Power LawNir Zicherman
  49. Unresolved debates about the future of AIHelen Toner
  50. Dual Process Theory (System 1 & System 2)LessWrong
  51. Advice for newly busy peopleSese
  52. OpenAI and Hugging Face partner to address security incident during model evaluationOpenAI
  53. “Long” timelines to advanced AI have gotten crazy shortHelen Toner
  54. The AI Revolution: The Road to SuperintelligenceTim Urban
  55. Federal Reserve announces the leadership and objectives of its task forces to advance the conduct of monetary policyFederal Reserve
  56. The Most Important Time in History Is NowTomas Pueyo
  57. The current SOTA model was released without safety evalsParv Mahajan
  58. When AI Chooses Harm Over FailureCivAI
  59. AI models can be dangerous before public deploymentMETR
  60. Why AI alignment could be hard with modern deep learningAjeya Cotra
  61. Specification Gaming: How AI Can Turn Your Wishes Against YouRational Animations
  62. Deep Ignorance: Filtering Pretraining Data Builds Tamper-Resistant SafeguardsO'Brien et al.
  63. The True Story of How GPT-2 Became Maximally LewdRational Animations
  64. What is input data filtration in AI safety?BlueDot
  65. Chain-of-Thought SnippetsBronson Schoen
  66. Neel Nanda on the race to read AI minds (part 1)80,000 Hours
  67. Introduction to Mechanistic InterpretabilityBlueDot
  68. What Do Neural Networks Really Learn? Exploring the Brain of an AI ModelRational Animations
  69. Introduction to AI ControlBlueDot
  70. What is AI alignment?Adam Jones
  71. Build Personal MoatsErik Torenberg
  72. Safety and alignment in an era of long-horizon modelsOpenAI
  73. A Framework for Frontier AI and the Dawning of a New AgeDemis Hassabis
  74. Scaling: The State of Play in AIEthan Mollick
  75. The Huggingface IncidentScott Alexander
  76. Bayes' ruleLessWrong
  77. Seeking Stability in the Competition for AI AdvantageIskander Rehman
  78. Reading Between the Dots: Decoding Hidden Computation across Filler TokensBrauer et al.
  79. Reps. Lieu and Moran Introduce Bill to Require Kill Switch for AI Systems That Can Cause Catastrophic HarmOffice of Rep. Ted Lieu
  80. Recent LLMs can use filler tokens or problem repeats to improve (no-CoT) math performanceRyan Greenblatt
  81. An analysis of AI-generated content at the Mechanistic Interpretability WorkshopAndy Arditi
  82. Silicon Valley's Safe SpaceCade Metz
  83. It's practically impossible to run a big AI company ethicallyVox Future Perfect
  84. You Will Listen to Carl on DwarkeshMatt Reardon
  85. Give Up Seventy Percent Of The Way Through The Hyperstitious Slur CascadeScott Alexander
  86. Intelligence is not the main bottleneckRuxandra Teslo
  87. In search of a dynamist vision for safe superhuman AIHelen Toner
  88. Utopia for Realists (Chapters 1–2)Rutger Bregman
  89. The Market for LemonsWikipedia
  90. Model access for third-parties — it's a big deal!Cleo Nardo
  91. OpenAI says its AI went rogue and launched 'unprecedented' cyber-attackBBC News
  92. Against Learning From Dramatic EventsScott Alexander
  93. Help us launch AI safety university groups by referring potential foundersJason Chin
  94. Preparing for LaunchInstitute for Progress
  95. The OpenAI/Huggingface incidentBuck Shlegeris
  96. Robin Hanson on AI and Large Language ModelsCloser To Truth
  97. He Risked Everything To Warn You: No One Is Ready For What's ComingThe Diary Of A CEO
  98. Tyler Cowen — The #1 bottleneck to AI progress is humansDwarkesh Patel
  99. This best-selling book is freaking out national security advisorsAI In Context
  100. Carl Shulman (Pt 1) — Intelligence explosion, primate evolution, robot doublings, & alignmentDwarkesh Patel
  101. What the hell happened with AGI timelines in 2025?80,000 Hours
  102. Aaron Scher — What Would it Take to Stop the Development of Superintelligence?FAR.AI
  103. What Happens When Capitalism Doesn't Need Workers Anymore?Economics Explained
  104. Constellation Seminar: Scaling AI SafetyRyan Kidd
  105. Dario Amodei — We are near the end of the exponentialDwarkesh Patel
  106. Understanding the inner thoughts of AIGoogle DeepMind
  107. What does the next training paradigm look like?Dwarkesh Patel
  108. Why AI Safety Needs Founders — Ryan KiddBlueDot Impact
  109. A visual guide to Bayesian thinkingJulia Galef
  110. The Scout Mindset by Julia Galef — Core MessageProductivity Game
  111. Unfortunately, You Need to Know What the Jevons Paradox isHank Green
  112. Using Dangerous AI, But Safely?Robert Miles AI Safety
  113. Richard Ngo — Reframing AGI Threat ModelsFAR.AI
  114. Grant Sanderson — AI disproved a famous math conjecture. Now what?Dwarkesh Patel
  115. We're Not Ready for SuperintelligenceAI In Context
  116. Rohin Shah — How to Theorize So Empiricists Will ListenFAR.AI
  117. If you remember one AI disaster, make it this oneAI In Context
  118. Large Language Models explained briefly3Blue1Brown
  119. The A.I. DilemmaCenter for Humane Technology
  120. You should, unfortunately, be worried about Sam Altman.AI In Context
  121. Do they know that we know that they know?Rational Animations