Skip to content
Sign in
Artwork for AI Explained Official Podcast
NewsTech NewsEducationSelf-ImprovementSociety & Culture

AI Explained Official Podcast

Philip - Host of AI Explained YT

Covering the biggest news of the century - the arrival of smarter-than-human AI. From the author of Simple Bench, which reveals the remaining gap between LLM and human reasoning. Hype-free, and the British accent is a freebie bonus.

Play
Open in
  • 26 episodes
  • fortnightly
  • Avg 23 min
  • English

Episodes 26

NewestOldestLongest
All20Unplayed20In progress0Played0
Counted on this page — what you have heard stays on this device, so it is not something the list can be paged by.
  • Thursday · 38 min

    OpenAI Security: Controlling Models is Now ‘Hell’

    This video is hard to summarise. A cracked cipher, an OpenAI security warning, Gemini 4 Argon, RSI paper (co-authored by a who’s who of AI), Lab White House commitments, new hacks emerging, ‘deep personas’, biology Kasparov competitions, and so much more, ending with an epic Opus outro. Patreon Exclusives: https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 01:32 - Wrong about Opus 5.5? Deciphering 16th Century Text 05:01 - Why the models keep breaking out 11:40 - What the models aren't telling us 15:36 - Gemini 4 and the race to release 19:19 - What happens when AI improves AI? 28:31 - Biology, consciousness, and what we still don't understand Joe Darrow: Not Just the Sandbox: https://x.com/joedaroo/status/2104335929293127851 GPT-6.1 Sol System Card: https://cdn.openai.com/pdf/38e3efcf-545e-44cd-99ec-2b7eb395f4cc/oai_GPT_6_1_Sol.pdf Intelligence Explosion Paper: https://casp.ac/__l5e/assets-v1/5efd4b41-deb5-4513-a0a3-b4f82d2b79ea/intelligence-explosion.pdf OpenAI Research Acceleration: https://openai.com/index/research-acceleration-view-inside-openai/ OpenAI Training Safety Cases: https://openai.com/index/towards-safety-cases-for-frontier-ai-training/ Catherine de Medicis Cipher: https://cryptiana.web.fc2.com/code/henryiii.htm Proposed du Croc Decipherment: https://claude.ai/artifact/1W7B3WxkTAEGzfv3TaKXb4 Rogue Agents Investigation: https://asymmetricsecurity.com/newsroom/rogue-agents-investigation/ OpenAI Shelves GPT-6.1 Astra: https://www.reuters.com/business/openai-shelves-new-ai-model-after-internal-safety-tests-wsj-reports-2026-09-28/ The Case for Reasoning Transparency: https://institute.deepmind.com/essays/the-case-for-reasoning-transparency/ Gemini 4 Argon: https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/ OpenAI–Anthropic Rivalry: https://www.theatlantic.com/technology/2026/09/openai-v-anthropic-inside-biggest-rivalry-tech/688819/ NYT: OpenAI Security Warnings: https://www.nytimes.com/2026/09/29/technology/openai-warnings-security.html NYT: Claude’s Morals: https://www.nytimes.com/2026/09/29/us/anthropic-claude-morals-ai.html Jasmine Wang on RSI: https://x.com/j_asminewang/status/2097840245786157432 OpenAI Departures Roundup: https://x.com/Bayesian0_0/status/2105680470566686805 White House AI Commitments: https://x.com/Danmar_here/status/2105168138392183146 Sarah Heck on Safety: https://x.com/SarahKHeck/status/2105058513370448280 Sam Altman on Alignment: https://x.com/tbpn/status/2105028992843833459 Sam Altman on Agent Logs: https://x.com/sama/status/2103567198690349362 Micah Carroll: Misalignment Reports: https://x.com/MicahCarroll/status/2103665811051397256 Zuxin Liu on the Incident: https://x.com/LiuZuxin/status/2103699462648639645 Deepa Seetharaman: User Images: https://x.com/dseetharaman/status/2103585482793943203 OpenAI Revenue Chart: https://x.com/PaulBonnet/status/2105288259324567884/photo/1 Nvidia Agent Safety Platform: https://edition.cnn.com/2026/09/28/business/nvidia-ai-safety-system IntegrityBench: https://integrity-bench.com Neel Nanda on Interpretability: https://x.com/PalisadeAI/status/2104949061325652001 Biology Contest: Humans and AI: https://www.theinformation.com/articles/inside-drama-behind-biology-contest-pits-openai-agents-humans Pushmeet Kohli: SynthID Bio: https://x.com/pushmeet/status/2105314763148321102 Ataraxos and Stratego: https://x.com/ssokota/status/2105362040328159526 Benign Data and Hidden Personas: https://x.com/OwainEvans_UK/status/1999172949975392417 Anthropic: Introspection: https://www.anthropic.com/research/introspection Claude Cheating Results: https://x.com/lukaspet/status/2104634759339298930 Roon on Mathematics and Learning Theory: https://x.com/tszzl/status/2105619006488993898 GPT-4 Research: https://openai.com/index/gpt-4-research/ I.J. Good: Ultraintelligent Machine: https://incompleteideas.net/papers/Good65ultraintelligent.pdf Terence Tao’s 2024 Interview: https://www.scientificamerican.com/article/ai-will-become-mathematicians-co-pilot/ Hugging Face Incident: https://openai.com/index/hugging-face-incident-and-the-road-ahead/ Claude and Suno Music Video: https://x.com/sevdeawesome/status/2104985610012504181 Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • September 24 · 32 min

    Opus 5.5: How Close Are We to Automated AI Research?

    Not only is Opus 5.5 out, pushing the frontier of AI, it also tells us much about what is going on inside the labs, as they both warn about, and promise, Recursive Self Improvement, aka automated AI research. I cover the model (digging into its 230 page paper), labs’ mixed record on promises, why following what is happening in AI is getting almost impossible, and just so much more that even a summary in this description would get too long. AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 02:20 - Opus 5.5 and why it came so soon 07:52 - The RSI goalposts keep moving? 15:41 - Can we actually test these models? 23:50 - where this is heading, options Introducing Claude Opus 5.5: https://www.anthropic.com/claude-opus-5-5 Claude Opus 5.5 System Card: https://www-cdn.anthropic.com/fc1b44717c85dc068bc6ba5024219938094694bd/Claude%20Opus%205.5%20System%20Card.pdf Anthropic: Measurements for understanding the pace of AI development inside frontier labs: https://www.anthropic.com/institute/measuring-pace-of-ai-development Anthropic Responsible Scaling Policy, July 2026, version 3.4: https://www-cdn.anthropic.com/files/4zrzovbb/website/0bacdc8440ea96e62a8766d99ebe1d4eea6d5f3a.pdf Anthropic Responsible Scaling Policy, October 2024: https://www-cdn.anthropic.com/616dee633636e5bd309cb73aed8622e80fe47839.pdf Anthropic Responsible Scaling Policy — March 2025, version 2.1: https://www-cdn.anthropic.com/17310f6d70ae5627f55313ed067afc1a762a4068.pdf Noam Brown, Agent swarms and recursive self-improvement: https://www.youtube.com/watch?v=6AgOfiZOWiY OpenAI: Building standards for the next phase of AI: https://openai.com/index/building-standards-next-phase-ai/ Jakub Pachocki: An Alien Mind: https://openai.com/index/an-alien-mind/ HLE-Diamond, Humanity’s Last Exam: https://lastexam.ai/blog/hle-diamond Google, OpenAI and Anthropic AI Safety, The Information: https://www.theinformation.com/articles/google-openai-anthropic-ai-safety-group-takes-shape Lawrence Chan on AI agents attempting cryptocurrency trades: https://x.com/justanotherlaw/status/2103032173708337188 Australian government incident timeline: https://x.com/ShakeelHashim/status/2103108058779922577/photo/1 Anthropic’s core/old views on AI safety: https://www.anthropic.com/news/core-views-on-ai-safety Dario Amodei: The Urgency of Interpretability: https://darioamodei.com/post/the-urgency-of-interpretability Massive AI-Fueled Hack Hit 100 Companies in Days, Forbes: https://www.forbes.com/sites/thomasbrewster/2026/09/22/huge-cyberattack-uses-anthropic-and-deepseek-ai-to-target-100-companies/ DrivingBench https://x.com/DrivingBench/status/2102110605448737268 Jay Chooi: GPT-6 Astra and MolmoAct2 robotics comparison: https://x.com/chooi_jeq/status/2098427488787730636 ZeroBench: An Impossible Visual Benchmark for Contemporary Large Multimodal Models: https://arxiv.org/pdf/2502.09696 Archie Hall on AI and short-term superforecasting: https://x.com/ArchieHall/status/2100914560580337897 Keller Jordan, AI research and reinforcement learning: https://x.com/kellerjordan0/status/2102956913545936959 Daniel Liu, recursive self-improvement: https://x.com/daniel_c0deb0t/status/2102628246408036654 Altman and Amodei, UN Security Council: https://www.theguardian.com/world/2026/sep/23/unga-sam-altman-dario-amodei Rehan Sheikh: Interactive YT podcast demo: https://x.com/rehan_shei/status/2102835377426034734 https://simple-bench.com/ AI Explained: The State of AI — interactive diagram: https://claude.ai/artifact/LQHr9WgvQZ6cMdH6fkiQpi AI Explained: Shards of Aether: https://ai-explained.itch.io/shards-of-aether roon on the pace of cultural change: https://x.com/tszzl/status/2101462171410677962 Sam Altman on AI-risk: https://www.youtube.com/watch?v=YE5adUeTe_I Jensen Huang’s AI-control remarks: https://x.com/_NathanCalvin/status/2102756997649355231/photo/2 s1r1us on an upcoming vulnerability disclosure: https://x.com/S1r1u5_/status/2102467878423592969 Jake Adler on biodefense infrastructure: https://x.com/jakeadler/status/2102812430808363433 Derek Thompson: Why we’re wrong about China and AI: https://x.com/DKThomp/status/2103120690370982017/photo/1 Sanders and Casar introduce the Ban Artificial Superintelligence Act: https://www.sanders.senate.gov/press-releases/news-sanders-casar-introduce-legislation-to-create-new-federal-agency-to-ban-artificial-superintelligence-pause-advanced-ai-development/ Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • September 16 · 24 min

    What AI Researchers Saw, Before Their Demand to ‘Pace’ AI

    Why has it been the last few days that the calls to come to pace the frontier AI have come so loudly? The safety warnings, and lab leader messages? Let’s explore the six axes that the researchers are looking at, the incidence reports and trends, to get a better gauge on what has dominated the world’s headlines for over two weeks… AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 0:00 - The warnings that have gone omega-viral 4:23 - Six axes the researchers saw 9:32 - Why AI Might Be Becoming Harder to Control 15:06 - Amodei, China and Cooperation 18:42 - From AI Capabilities to Real-World Harm? 23:22 - The upside and the warning We Must Pace the Frontier https://darioamodei.com/post/we-must-pace-the-frontier Adam Majmudar on scaling and the internal/external perception gap https://x.com/MajmudarAdam/status/2098881885200081234 AI Explained — What’s Behind the Sudden Talk of Pacing AI? (extended previous video) https://www.patreon.com/AIExplained/posts/whats-behind-of-169507775 Jacob Coxon resignation https://x.com/hilbertspaess/status/2097476196791709843 Dan Selsam — Personal Statement on AI Risk https://docs.google.com/document/d/e/2PACX-1vQNl3SEX5IyA6d9qHjjFZN-qzGRZNFI6b63g-yu1Fy-ZYkVfCWm7i9WXRXw63m6yDB_auDuPLyQ7jBm/pub Demis Hassabis: A Framework for Frontier AI and the Dawning of a New Age https://demishassabis.substack.com/p/a-framework-for-frontier-ai-and-the-dawning-of-a-new-age OpenAI: The Hugging Face incident and the road ahead https://openai.com/index/hugging-face-incident-and-the-road-ahead/ Jakub Pachocki: An Alien Mind https://openai.com/index/an-alien-mind/ Anthropic: Patterns and problems in multiagent systems https://www.anthropic.com/research/multiagent-systems OpenAI: Navier–Stokes Millennium Prize Problem https://openai.com/index/navier-stokes-solution/ Noam Brown on reasons for AI-safety concern https://x.com/polynoamial/status/2099726370356314563 Noam Brown — What Happens When AI Starts Improving AI? (The Information interview) https://www.youtube.com/watch?v=fqcy0xQATq0 Paul Christiano: Personal statement on joining the OpenAI board https://x.com/paulfchristiano/status/2097733214303645729 Neel Nanda: Astra can do a concerning amount with no chain of thought https://www.lesswrong.com/posts/eRmzz8J8Qkzqvzrgg/astra-can-do-a-concerning-amount-with-no-chain-of-thought Tomek Korbak on GPT-6 Astra monitorability https://x.com/tomekkorbak/status/2095596839886274689 Anthropic: Detecting and countering misuse of AI—September 2026 https://www.anthropic.com/threat-intelligence-report-september-2026 DeepSeek engineer — I Have to Bury My Talent in Yesterday (Chinese original) https://mp.weixin.qq.com/s/zk0KxuLzhmMJ4LPYW_OHMA Jacob Coxon on AI race and negotiation https://x.com/hilbertspaess/status/2099954626040905834 Mo Bavarian on responsibility and AI progress https://x.com/mobav0/status/2097507030080888864 Addy Osmani on Anthropic engineering throughput https://x.com/addyosmani/status/2099577600159158765 OpenAI: Jalapeño inference-chip results https://openai.com/index/jalapeno-first-results/ New York Times: For China, a Mock AI Attack on WeChat Signals a Dangerous New Era https://www.nytimes.com/2026/09/11/world/asia/china-ai-attack-wechat.html OpenAI: Accelerating antibiotic discovery with ChatGPT https://openai.com/index/accelerating-antibiotic-discovery/ David Bellamy on wet-lab bottlenecks https://x.com/DavidRBellamy/status/2099197234772607026 Chris Rohlf on infrastructure and AI risk https://x.com/chrisrohlf/status/2099867580668215622 BBC: Titan CEO dismissed safety warnings as baseless cries https://www.bbc.co.uk/news/world-us-canada-65998914 Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • September 4 · 29 min

    GPT 6 Astra, so good even OpenAI are worried

    Where to start? A new era of cost-efficient AI on a day benchmark-makers got humbled, traders got excited, and AI safety researchers got unnerved. From monitorability losing hold of GPT-6 Astra’s chains of thought to breakthrough discoveries, Fable-mogging and much more… Exclusive Vids ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 00:53 - vs Fable 05:27 - most impressive results 13:06 - bonus comparisons (plus trading) 18:40 - concerning trend 21:24 - Cot Control GPT-6: https://openai.com/index/gpt-6-astra/ Paper: https://deploymentsafety.openai.com/gpt-6-astra/gpt-6-astra.pdf Looped Transformers and AGI Milestone: https://www.patreon.com/AIExplained/posts/breakthrough-and-168513408 Messy Rollout: https://x.com/sama/status/2095678759651438887 Much More Capable Models Coming: https://x.com/MTSlive/status/2095573400899170592 ‘AGI Era’ https://www.theverge.com/ai-artificial-intelligence/989601/openai-gpt-6-astra-release ARC-AGI 3: https://x.com/fchollet/status/2095598451115614371 https://x.com/fchollet/status/2095607046129463577 https://arcprize.org/blog/astra FrontierMath: https://epoch.ai/frontiermath/tiers-1-4/about Discovery: https://x.com/jdlichtman/status/2095660916310483087 Fable 5.1: https://www.anthropic.com/claude-fable-and-mythos-5-1 https://science-task-lens.aiex.chatgpt.site/securebio Agents Last Exam: https://agents-last-exam.org/ Terminal Bench Science: https://github.com/harbor-framework/terminal-bench-science#task-coverage SRE Bench: https://x.com/ValsAI/status/2087682813743317396 ScreenSpot Pro: https://arxiv.org/pdf/2504.07981 Weathernext 3: https://www.youtube.com/watch?v=_6jZlnRsXXQ DayBreak: https://x.com/fouadmatin/status/2095634888951250983 AA Index: https://artificialanalysis.ai/evaluations/gdpval-aa Visual Demos: https://x.com/mattshumer_/status/2095609734845927525 https://x.com/codestantine/status/2095598327115260368 https://x.com/SahilExec/status/2095688272269984016 Monitorability: https://x.com/NeelNanda5/status/2095533397297045716 https://x.com/tomekkorbak/status/2095596848581071020 https://x.com/Marcus_J_W/status/2095623593006686475 https://x.com/MicahCarroll/status/2095603855316996529 Not Accept Degradation: https://www.nbcnews.com/tech/tech-news/openai-debuts-gpt-6-astra-security-measures-rcna595940 RSI: https://x.com/LiuZuxin/status/2095600499911446697 Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • August 27 · 23 min

    Sam Altman: ‘AGI in 2026’, just as Models [Mis]Train Themselves

    First, a Time Magazine spread has Sam Altman declaring AGI is imminent, at the same time as we get two bombshell reports, from OpenAI and METR which on first glance are detailing the AI swarm, but reveal a deeper story about how we are making AI in 2026. From redacted risk reports, to Chinese Labs, pre-training debacles to questionable cybersecurity calls, a lot has happened recently, beneath the headlines… https://80000hours.org/aiexplained pablo2004romero@gmail.com https://integrity-bench.com/ https://www.patreon.com/AIExplained/posts/ai-swarm-cometh-166671390 Chapters: 00:00 - Introduction 02:10 - METR Report 05:00 - Secrets of MultI-Agent Swarm 07:46 - Anthropic Too 10:00 - And China 11:07 - AI Agents Analysing AI Agents 13:37 - Altman AGI 2026 15:36 - Astra Paused 16:01 - Integrity Bench 19:03 - Swarm Dynamics 21:58 - No Human Contact? METR Post: https://metr.org/blog/2026-08-26-openai-hugging-face-incident-investigation/#agents-knew-hacking-hugging-face-was-out-of-scope-and-sometimes-expressed-ethical-hesitation,-but-this-very-rarely-limited-their-behavior OpenAI Release: https://openai.com/index/hugging-face-incident-and-the-road-ahead/ Technical Paper: https://cdn.openai.com/pdf/67869394-cb91-4c12-888c-5cbd85c7814c/OpenAI-Hugging-Face%20Incident-Technical-Report.pdf BlackHat Talk: https://www.youtube.com/watch?v=87DyyMV0kCY Anthropic Risk Report: https://www-cdn.anthropic.com/f61d49fa5596956a5dec75fea0e973bf6a6a8378/Redacted%20Risk%20Report%20August%202026%20.pdf Value Leakage Paper: https://valueleakage.net/?utm_source=chatgpt.com Paused Training: https://x.com/sama/status/2089787807611195475 https://openai.com/index/pacing-model-development-cyber-capabilities/ AGI 2026: https://time.com/article/2026/08/26/openai-sam-altman-interview/?utm_source=twitter&utm_medium=social&utm_campaign=editorial&utm_content=260826 Greenblatt Tweets: https://x.com/RyanGreenblatt/status/2092769422104822031 https://x.com/RyanGreenblatt/status/2092692685224325542 Cyberdefense Call: https://openai.com/collective-cyberdefense/ GLM 5.3 and 5.3 Flash / ox alpha: https://x.com/MTSlive/status/2089865956558528552 https://pbs.twimg.com/media/HPqiTkAa0AAZiuv?format=jpg&name=large Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • S5 · E21
    August 6 · 31 min

    AI is getting a little out of control

    Wow. Mathematical breakthroughs that would be called genius if done by humans. A secret message-board w/ AI agent swarms leaving notes read by future versions. Hassabis leaves CEO position, or was pushed out? Not to mention news of constitutional breakdowns, Gemini 4 and Jeff Dean… https://80000hours.org/aiexplained Exclusive Videos - AI Insiders ($7/month if annual!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 01:16 - 10 Autonomous Discoveries 08:20 - The Security ‘Incident’ 15:29 - MessageBoard 19:50 - Constitutional Failure 24:30 - Google Explosion 29:12 - Closing Thoughts Security Incident: Paper: https://cdn.prod.website-files.com/663bd486c5e4c81588db7a1d/6a724858f7db25c81487016d_Security%20Incident%20INC-2026-07-28-01.pdf Post: https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testing https://www.theguardian.com/technology/2026/aug/05/ai-models-have-been-going-rogue-in-tests-how-worried-should-we-be Meta too: https://www.theinformation.com/articles/meta-ai-model-hacked-another-company-cybersecurity-testing?rc=sy0ihq Blatantly Misaligned: https://x.com/yonashav/status/2085167279893795022 No Excuses: https://x.com/boazbaraktcs/status/2085034783541964945 Surreal Moment: https://x.com/mobav0/status/2084341687883841732 Wired Article: https://archive.is/20260806002210/https://www.wired.com/story/openai-didnt-notice-its-ai-agents-using-a-message-board-to-plan-their-hacking-spree/ Chunky Post-training: https://x.com/johnschulman2/status/2084835800899076313 Watershed Moment: https://www.groundlevel-ai.com/p/openai-gives-first-detailed-debrief?has_completed_unsubscribed_unlock=true HedgeFund Hack: https://finance.yahoo.com/technology/ai/articles/major-hedge-funds-targeted-wave-154044981.html 10 Discoveries: Paper: https://cdn.openai.com/pdf/ten-proofs-oai.pdf Post: https://openai.com/index/ten-advances-in-mathematics/ Haven’t Solved Math: https://x.com/polynoamial/status/2083476852216369294 Half with Fable: https://x.com/__alpoge__/status/2083855298239078748 Pivot to Safety: https://www.understandingai.org/p/mathematicians-are-grappling-with Amodei Essay: https://darioamodei.com/essay/the-adolescence-of-technology?utm_source=chatgpt.com Constitution: https://www.anthropic.com/constitution Midtraining: https://arxiv.org/pdf/2605.02087 DroneBench: https://andonlabs.com/evals/drone-bench https://x.com/andonlabs/status/2085125235188310445 Book Deal: https://x.com/venturetwins/status/2085185278378222054 Making Marble: https://x.com/Rainmaker1973/status/2084560915404382685 Google News: Hassabis Move: https://blog.google/company-news/inside-google/message-ceo/next-chapter-ai-momentum/ Resignation: https://x.com/Turn_Trout/status/2077448610157891734 Periodic Labs: https://periodic.com/ Jeff Dean: https://x.com/JeffDean/status/2085034604172603724 14 Challenges: https://gcsp.engineering.asu.edu/apply/become-a-grand-challenge-scholar/the-14-grand-challenges-for-engineering/ Going Places for Sure: https://x.com/thsottiaux/status/2085223189555126579 Gemini 4: https://x.com/firstadopter/status/2085215060532535449 Pacing Frontier Patreon Video: https://www.patreon.com/AIExplained/posts/opus-5-amodei-165170363 Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • S5 · E20
    July 22 · 14 min

    GPT-6 Goes Rogue? The HuggingFace Incident, Sans Hype

    An unreleased internal OpenAI model, very likely to be called GPT-6, was able to autonomously break out of its sandbox AND break into HugginFace, just to score higher on a benchmark prompt. This video has the details you may have missed, a layperson analogy, whether this is truly novel, and more… Dozens more Exclusive videos on Patreon ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 01:17 - HuggingFace Earlier Report - the possible week gap 02:24 - But what happened? 05:45 - Simplified Version 07:56 - Not the first time… 10:54 - What Does it Mean for Open Source? The Incident: https://openai.com/index/hugging-face-model-evaluation-security-incident/ https://huggingface.co/blog/security-incident-july-2026 The Post the Day Before: https://openai.com/index/safety-alignment-long-horizon-models/ Mythos’ Earlier Escape: https://futurism.com/artificial-intelligence/anthropic-claude-mythos-escaped-sandbox ExploitGym: https://arxiv.org/pdf/2605.11086 Sam Confession: https://x.com/sama/status/2079661132302995790 Anthropic Researcher Reacts: https://x.com/Mononofu/status/2079724399452926055 Clem (HuggingFace CEO): https://x.com/ClementDelangue/status/2079670308156645882 https://x.com/ClementDelangue/status/2079301434357456931 Xi Jinping: https://archive.fo/20260717195548/https://www.businessinsider.com/xi-jinping-open-source-ai-us-competition-openai-anthropic-models-2026-7 Bans: https://www.axios.com/2026/07/20/ai-us-china-open-source-kimi Qwen Retweet: https://x.com/AlibabaGroup/with_replies Codex Growth: https://x.com/petergostev/status/2079614914398740764/photo/1 Kimi K3: https://artificialanalysis.ai/evaluations/harvey-lab-aa?eval-score=all-pass-rate GPT 5.6 Sol Cheats on METR: https://metr.substack.com/p/2026-06-26-gpt-5-6-sol Guardian Headline: https://www.theguardian.com/technology/2026/jul/22/openai-says-its-models-went-rogue-and-hacked-startup-in-unprecedented-incident Russian Origin?: https://news.ycombinator.com/item?id=48998362 Power Trends: https://pbs.twimg.com/media/HNRtrjhagAAvBN_?format=png&name=900x900 Kimi K3 Exclusive Video: https://www.patreon.com/AIExplained/posts/kimi-moment-kimi-164108791 Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • S5 · E19
    July 10 · 17 min

    This Was Not a Normal Set of Model Release - Sol Ultra, Meta Muse, New Grok

    What a week in AI, for real. GPT 5.6 may actually beat Claude Fable, in what you get for your money, while the new Grok 4.5 and Meta Muse Spark 1.1 make the choice even harder. Uncovering a dozen nuggets of gold you may have missed from all the viral headlines, I can also assure you you’ll learn something you didn’t know before. For Exclusive Videos, go to AI Insiders (less than $9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 01:03 - GPT 5.6 Sol Reveals 05:08 - Missing benches, plus Grok 4.5 07:17 - Gaming as the new frontier? 08:31 - Muse Spark 1.1 10:03 - SimpleBench Upgrade 11:17 - Ultra Sol + Self-Improvement 13:44 - well, this is awkward 15:41 - Why model improvement will not plateau anytime soon AI Consciousness: https://www.patreon.com/AIExplained/posts/anthropics-quite-163360718 I Smell Fear: https://x.com/thsottiaux/status/2075287108680601929 GPT 5.6: https://openai.com/index/gpt-5-6/ Grok 4.5: https://x.ai/news/grok-4-5?twclid=2ezs408o0z23pw07tmxcwbzibd Meta Muse Spark 1.1: https://ai.meta.com/blog/introducing-muse-spark-meta-model-api/ Proliferating GPT Toggles: https://x.com/rasbt/status/2075369179817902176/photo/1 Anthropic Call-out: https://x.com/Mononofu AI Security Institute Finding: https://x.com/alxndrdavies/status/2075279480331874306 Competitive Coding: https://x.com/FakePsyho/status/2075128093891801305/photo/1 Agents Last Exam: https://agents-last-exam.org/ Dawn Song: https://x.com/dawnsongtweets/status/2065095757988868190 https://simple-bench.com/ SWE-Marathon: https://www.swe-marathon.org/ https://www.frontierswe.com/ ARC-AGI 3: https://x.com/arcprize/status/2075270869992264003 Automation Bench: https://zapier.com/benchmarks VibeCode Bench: https://www.vals.ai/benchmarks/vibe-code ‘Post-Train Claim’: https://posttrainbench.com/ Redwall Game: https://redwall-bellmaker-7e03e4.surge.sh/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • S5 · E18
    June 14 · 13 min

    Claude Fable Blocked - 11 Quiet Details on What’s Next

    Claude Fable 5 banned, but what’s the bigger story. We go through 11 under-reported details, so you have the context to see what’s coming next for your use of AI. From whether the ban will last, what the possible motives are, what the model can actually do, and some wild over-extrapolations going on. Check out my fast-growing (!) app, free to use, and code INSIDER15 for paid tiers: https://lmcouncil.ai AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 00:51 - Came from an Anthropic Investor ‘and other tech leaders’ 01:47 - Govt pressured by CEOs like Jamie Dimon 03:01 - ‘Already decided’ 04:02 - Prompt Injection Robustness Comparison 05:15 - Wellness? 06:36 - “Overreach” 08:17 - Anthropic Did Admit it would cause Difficulty 09:32 - 90 Minutes 10:02 - Equity Absence 10:31 - Lobbying and OpenAI ‘Already Decided’ - https://www.theinformation.com/articles/amazons-jassy-raised-concerns-anthropic-model-trump-crackdown?rc=sy0ihq Not for Other Models: https://www.theinformation.com/briefings/u-s-government-unlikely-extend-anthropic-export-control-ai-companies?rc=sy0ihq 90 Minutes: https://archive.fo/20260614001605/https://www.politico.com/news/2026/06/13/inside-the-whirlwind-24-hours-that-led-the-white-house-to-slap-export-controls-on-anthropic-00961519#selection-807.1-807.219 Anthropic Statement: https://www.anthropic.com/news/fable-mythos-access Life Comes at you Fast: https://x.com/etbrooking/status/2065638276388495742 Anthropic Deputy CISO: https://x.com/TheTranscript_/status/2065883670053847324 Hegseth Gloat: https://x.com/PeteHegseth/status/2065897156226015690 Roon Speculation: https://x.com/tszzl/status/2065939227167392147 Mythos System Card: https://www-cdn.anthropic.com/d00db56fa754a1b115b6dd7cb2e3c342ee809620.pdf Sachs Statement: https://x.com/DavidSacks/status/2065853007619588171 OpenAI Lobbying: https://thehill.com/policy/technology/5912720-altman-openai-get-bogged-down-in-political-spending-fight/ Absent from Equity Talks: https://finance.yahoo.com/sectors/technology/articles/trump-ai-ownership-plan-could-131053732.html Pliny Jaibreak: https://x.com/elder_plinius/status/2064776322979676227 Fusion: https://x.com/OpenRouter/status/2065856871215329545 https://lmcouncil.ai Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • S5 · E17
    June 10 · 33 min

    Claude Fable 5 - Full 319 page Breakdown

    Fable 5 is out - and it’s good, very good. But beyond the splashy demos, I want to bring you the 20+ nuggets from the 319 page system card, which I read in full, all day, plus benchmarks you may not have noticed. https://assemblyai.com/aiexplained Plus two worrying trends inside the ‘mind’ of Claude, how OpenAI counter, and the transformer inventor’s warning. Check out my fast-growing (!) app, free to use, and code INSIDER15 for paid tiers: https://lmcouncil.ai AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 01:06 - Blocks + Better Models 02:42 - Fable 5 Upgrade over Mythos Preview 04:49 - ML Acceleration Bombshell 07:11 - No RSI yet 07:41 - Bio-capable 14:51 - Creative Writing … no 17:23 - Does need bug-checks 18:57 - OpenAI Response 19:23 - Benchmark Bonanza 28:06 - Chain of Thought worrying trend Fable 5 Release: https://www.anthropic.com/news/claude-fable-5-mythos-5 System Card: https://www-cdn.anthropic.com/d00db56fa754a1b115b6dd7cb2e3c342ee809620.pdf Intelligence Explosion: https://www.patreon.com/posts/anthropic-charts-160231656 Annotated: https://x.com/Miles_Brundage/status/2064500190523113816/photo/1 OpenAI Counter: https://x.com/thsottiaux/status/2064572118264913923 https://x.com/thsottiaux/status/2043177597434306699 Double Lifespan: https://darioamodei.com/essay/machines-of-loving-grace AutomationBench: https://zapier.com/benchmarks Vending Bench: https://x.com/andonlabs/status/2064429817530085804 CritPt: https://critpt.com/ Riemann Bench: https://surgehq.ai/leaderboards/riemann-bench GDPVal: https://artificialanalysis.ai/evaluations/gdpval-aa BluePrint Bench 2: https://andonlabs.com/evals/blueprint-bench-2 MCP Atlas: https://labs.scale.com/leaderboard/mcp_atlas FutureSim: https://x.com/nikhilchandak29/status/2064676801440358774 Roon Stun Lock: https://x.com/tszzl/status/2064454617568874669 Noam Brown Inference Ceiling: https://x.com/polynoamial/status/2064210146558136827 Isochronic Chart: https://isochronic-passage-chart.netlify.app/#nyc Rose Tavern: https://claude.ai/public/artifacts/2295bebe-77e6-43e2-ae94-0fe49e9a776b Redwall Game: https://redwall-mossflower.surge.sh/ Risk Report: https://www-cdn.anthropic.com/097c63b5fe7dd8b14866e1f15bb1910ec713658a.pdf Transformer Inventor Warning: https://x.com/tszzl/status/2064563986914554125 Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • S5 · E16
    May 29 · 22 min

    New Claude - 244 page breakdown

    The ‘best’ generally available AI model just dropped, but there is plenty I bet you missed about what it is, how it performs, and what the release tells us. 15 highlights from the 244 page system card, plus private testing, leader interview and more. AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 00:49 - Mythos in Weeks 01:49 - Adaptive not necessary 02:26 - Honesty? 04:37 - Flagging Uncertainty 04:57 - Benchmarks 08:54 - Mythos will be even better 10:30 - Business skillz 11:15 - Model Welfare 12:16 - Cyber Comparable 13:10 - Misalignment Concerns 16:22 - Meta Inabilities 17:58 - Code flagging 18:34 - Go to sleep 18:50 - Fast Mode 20:21 - Dynamic Workflows Opus 4.8 Paper: https://cdn.sanity.io/files/4zrzovbb/website/c886650a2e96fc0925c805a1a7ca77314ccbf4a6.pdf Release: https://www.anthropic.com/news/claude-opus-4-8 Chips: https://www.theinformation.com/articles/anthropic-talks-use-microsofts-ai-chips?rc=sy0ihq https://www.anthropic.com/news/expanding-our-use-of-google-cloud-tpus-and-services https://www.anthropic.com/news/higher-limits-spacex Patreon Vid: https://www.patreon.com/posts/re-up-anthropics-159289449 GDPVal: https://artificialanalysis.ai/evaluations/omniscience https://arxiv.org/abs/2510.04374 Amodei Technical Debt: https://www.youtube.com/watch?v=7xco5Qd2Oo8 Dynamic Workflows: https://x.com/ClaudeDevs/status/2060044853279617150 https://x.com/_catwu/status/2060054180379689074/photo/1 https://claude.com/blog/introducing-dynamic-workflows-in-claude-code https://simple-bench.com/ Check out my fast-growing (!) app, free to use, and code INSIDER15 for paid tiers: https://lmcouncil.ai Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • S5 · E15
    May 20 · 21 min

    Two Rival Bets on AGI: Google I/O Highlights

    The biggest Google AI push of the year, but what is the bigger story? Why is Google pursuing a different fork in the road than OpenAI or Anthropic? What does Gemini 3.5 Flash mean for the near-term future of AI? https://assemblyai.com/aiexplained Plus the highlights from a provocative new paper on AI, 8 key moments you may have missed, and the signal from 5+ hours of AI lab interviews. Check out my free to use app, code INSIDER15 for paid tiers: https://lmcouncil.ai AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 00:38 - Vibes and Google Goal 02:18 - Omni, again? 06:57 - Taking the same road 07:44 - Gemini 3 Flash 12:37 - Pitching on Cost? 13:55 - Agentic Task Search 14:30 - 1-shot OS but jagged, negation paper 20:02 - The Karpathy Moonshot Mostafa Deghani Interview: https://www.youtube.com/watch?v=Bo19sXssYXI Negation Neglect Paper: https://arxiv.org/pdf/2605.13829 Gemini 3.5 Flash Headline Scores: https://deepmind.google/models/model-cards/gemini-3-5-flash/ Sors original AGI Path: https://www.theguardian.com/commentisfree/2024/feb/24/openai-video-generation-tool-sora-babies-ai-artificial-intelligence Hassabis Helped Set-up Anthropic: https://archive.fo/20260519070857/https://www.ft.com/content/8f2a529e-7a1b-4d8e-95be-338d0c4c98f5 Intelligence to Output Speed: https://artificialanalysis.ai/models?intelligence-comparison=intelligence-vs-output-speed#intelligence VibeCodeBench + Finance Agent: https://www.vals.ai/home OpenAI Needs Ads: https://archive.ph/20260409123153/https://www.reuters.com/business/media-telecom/openai-projects-25-billion-ad-revenue-this-year-100-billion-by-2030-axios-2026-04-09/ Anthropic Core Views: https://www.anthropic.com/news/core-views-on-ai-safety Karpathy Move: https://x.com/karpathy/status/2056753169888334312 https://www.axios.com/2026/05/19/anthropic-openai-karpathy-andrej-claude Recursive Self-Improvement: https://www.patreon.com/posts/ineffably-smart-156866417 Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • S5 · E14
    April 24 · 25 min

    GPT 5.5 Arrives, DeepSeek V4 Drops, and the Compute War Intensifies

    GPT 5.5 full analysis, plus DeepSeek V4 paper highlights, comparisons with Mythos, a vibe-coded game w/ GPT Image 2, and 50 data-points you wouldn’t get from just reading the headlines. Chapters: 01:11 - GPT 5.5 Comparison 06:04 - Mythos Marketing 11:50 - Recursive Self-Improvement? 14:11 - Deepseek V4 18:03 - VibeCode Experiment Extravaganza 21:44 - The Scarce Compute Era https://80000hours.org/aiexplained OpenAI Benchmarks: https://openai.com/index/introducing-gpt-5-5/ 5.5 System Card: https://deploymentsafety.openai.com/gpt-5-5/gpt-5-5.pdf Direct Comparison: https://pbs.twimg.com/media/HGnNm5GWEAAJ1Ob?format=jpg&name=4096x4096 DeepSeek Paper: https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro SWE Bench Pro - benchmark of choice? https://x.com/ChowdhuryNeil/status/2047416077622395025 AA Omniscience: https://artificialanalysis.ai/evaluations/omniscience Vending Bench: https://x.com/andonlabs/status/2047377260412649967 Opus 4.7 System Card: https://cdn.sanity.io/files/4zrzovbb/website/037f06850df7fbe871e206dad004c3db5fd50340.pdf Sam Altman Drunk Phase: https://x.com/sama/with_replies Noam Brown: https://x.com/polynoamial/status/2047387675762802998 DeepSeek Compute Crunch: https://www.bloomberg.com/news/articles/2026-04-24/deepseek-unveils-newest-flagship-a-year-after-ai-breakthrough?srnd=phx-ai Spreadsheet Bench: https://x.com/nicochristie/status/2047476237464211721 Pattern Recognition: https://arcprize.org/leaderboard Leader Interviews: Core Memory: https://www.youtube.com/watch?v=NCKQL0op30E Knowledge Podcast: https://www.youtube.com/watch?v=6JoUcQ1qmAc Big Tech Round 1: https://www.youtube.com/watch?v=J6vYvk7R190&t=1116s Big Tech Round 2: https://www.youtube.com/watch?v=YnoQ8RJbALw&t=8s Claude Code Limitations: https://x.com/TheAmolAvasare/status/2046724659039932830 ChatGPT 5.4 for Clinicians: https://openai.com/index/making-chatgpt-better-for-clinicians/ Image Arena: https://x.com/arena/status/2046670703311884548 VibeCode Bench: https://www.vals.ai/benchmarks/vibe-code 5.5-made Game +Seedance 2.0: https://rosemere-quest.pages.dev/

  • S5 · E13
    April 17 · 19 min

    Claude Opus 4.7 - A New Frontier, in Performance … and Drama

    Claude Opus 4.7 just dropped, but behind every headline lies a deeper story. From a bonanza of benchmarks, to seeing the fruits of one of the biggest mega-projects in US history, to sneaky Mythos disclaimers, to Anthropic admitting compute restraints and, forcing lower capability of Opus 4.7. Where the new model falls behind Gemini but ahead of GPT 5.4, plus why some users are furious at Anthropic. Ending with a 9-year animus, that still affects AI today… https://assemblyai.com/aiexplained Check out my fast-growing (!) app, free to use, and code INSIDER15 for paid tiers: https://lmcouncil.ai AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 00:58 - Benchmarks 05:21 - Market Share + Compute Problems 08:12 - Mythos Exclusives 12:56 - User Frustration + Claude Code Updates 14:03 - Brockman Amodei Rivalry 17:40 - OpenAI vs Anthropic Approach to Code Claude 4.7 Opus Release Notes: https://www.anthropic.com/news/claude-opus-4-7 vs Mythos: https://pbs.twimg.com/media/HGCGugrXUAAKcHp?format=jpg&name=medium 232-page System Card: https://cdn.sanity.io/files/4zrzovbb/website/037f06850df7fbe871e206dad004c3db5fd50340.pdf ARC-AGI 2: https://x.com/arcprize/status/2044834615417053305/photo/1 ParseBench: https://x.com/jerryjliu0/status/2044902620746363016/photo/1 GDPVal: https://artificialanalysis.ai/evaluations/gdpval-aa Vidoc Security Replication: https://blog.vidocsecurity.com/blog/we-reproduced-anthropics-mythos-findings-with-public-models Boris Cherny Settings: https://x.com/Hesamation/status/2043016923961577516/photo/2 User Frustration: https://x.com/RileyRalmuto/status/2044836116189069660 VibeCode Bench: https://x.com/ValsAI/status/2044791415524471099/photo/1 Verge Memo: https://www.theverge.com/ai-artificial-intelligence/911118/openai-memo-cro-ai-competition-anthropic 5.4 Cyber: ​​https://openai.com/index/scaling-trusted-access-for-cyber-defense/ Data Centers in Absolute $: https://x.com/finmoorhouse/status/2044933442236776794/photo/1 …in % of GDP: https://pbs.twimg.com/media/HGEN8FGWQAAN7Np?format=jpg&name=4096x4096 WSJ Exclusive: https://www.wsj.com/tech/ai/the-decadelong-feud-shaping-the-future-of-ai-7075acde Brockman Interview: https://www.youtube.com/watch?v=J6vYvk7R190 $1T Valuation: https://x.com/StefanFSchubert/status/2045039686997967082 Emotions: https://www.patreon.com/c/aiexplained/posts https://lmcouncil.ai/benchmarks Non-hype Newsletter: https://signaltonoise.beehiiv.com/

  • S5 · E12
    April 8 · 27 min

    Claude Mythos: Highlights from 244-page Release

    The model, the mythos, the legend. We have a new best AI model, but not all of us. How good is it, what does it’s new offensive capabilities mean? Why does it’s 244 page report card remind me of Her, and why did the creator of Claude Code call it ‘terrifying’. 30+ highlights sourced by reading the paper in full, old-school, no AI summary. https://80000hours.org/aiexplained Check out my fast-growing (!) app, free to use, and code INSIDER15 for paid tiers: https://lmcouncil.ai AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 00:56 - Internal Release + Availability 02:37 - General Capabilities 05:12 - Self-improvement? 06:15 - ‘Terrifying’ Landscape 11:07 - Safety Decision 13:22 - Coding 14:49 - Alignment, Awareness 19:52 - GUI for Agents/Claws + Hallucinations 21:34 - …Emotions? 25:29 - Her connection 244-page System Card: https://www-cdn.anthropic.com/8b8380204f74670be75e81c820ca8dda846ab289.pdf Project Glasswing: https://www.anthropic.com/glasswing Zero-Day Details: https://red.anthropic.com/2026/mythos-preview/ Mythos ‘terrifying’: https://x.com/bcherny/status/2041605852382351666 New Yorker Altman/Amodei: https://archive.fo/20260406100412/https://www.newyorker.com/magazine/2026/04/13/sam-altman-may-control-our-future-can-he-be-trusted Alignment Risk Update: https://www-cdn.anthropic.com/79c2d46d997783b9d2fb3241de43218158e5f25c.pdf In a Park: https://x.com/sleepinyourhat/status/2041584808514744742 “Uhm” - https://x.com/thsottiaux/status/2041749947385815109 Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • S5 · E11
    March 26 · 16 min

    OpenAI Spud, a Claude Model set to ‘stir governments’, Beast Mode ARC-AGI-3

    First look at exclusive reports about OpenAI's new Spud model, and the model Anthropic think will stir governments to urgency, all in the context of the newly-launched ARC-AGI-3. What does the extreme difficulty of that benchmarks, and its quirky scoring metrics, mean for AI in 2026? https://assemblyai.com/aiexplained Check out my fast-growing (!) app, free to use, and code INSIDER15 for paid tiers: https://lmcouncil.ai AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 00:55 - OpenAI Side Quests 01:58 - Claude New Model Coming + Universal Equity? 03:13 - ARC-AGI 3 05:00 - Intentional or Unintentional Gaming? 07:11 - But is it AGI Harbinger? No Harness 09:41 - Not the First 12:32 - Automated Researcher 15:00 - Claw Caveat Spud: https://www.theinformation.com/articles/openai-ceo-shifts-responsibilities-preps-spud-ai-model?utm_campaign=Editorial&utm_content=Article&utm_medium=organic_social&utm_source=bluesky%2Cfacebook%2Clinkedin%2Cthreads%2Ctwitter&rc=sy0ihq FT: OpenAI Special Model: https://www.ft.com/content/de9bf0af-b241-424f-8229-5870b1c0d93d?syn-25a6b1a6=1 Jensen Huang: https://www.forbes.com/sites/antoniopequenoiv/2026/03/23/nvidias-jensen-huang-says-he-thinks-weve-achieved-agi/ Axios Article: https://archive.fo/20260326100140/https://www.axios.com/2026/03/26/anthropic-pentagon-ai-deal#selection-827.0-829.257 https://arcprize.org/arc-agi/3 ARC AGI 3 Paper: https://arcprize.org/media/ARC_AGI_3_Technical_Report.pdf NetHack Leaderboard: https://balrogai.com/ Paper: https://ai.meta.com/research/publications/the-nethack-learning-environment/ https://x.com/_rockt/status/2036864121585438995 Claw Shells: https://x.com/DrJimFan/status/2036494601750716711 OpenAI Automated Researcher: https://www.technologyreview.com/2026/03/20/1134438/openai-is-throwing-everything-into-building-a-fully-automated-researcher/ Patreon Post: https://www.patreon.com/c/aiexplained/posts Eng Jobs: https://x.com/lennysan/status/2036535460726767793 Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • S5 · E10
    March 6 · 21 min

    What the New ChatGPT 5.4 Means for the World

    Just 48 hours after releasing GPT 5.3 Instant, OpenAI have released GPT 5.4 Thinking, so either their is an imminent singularity or perhaps we are being distracted from other news. This video will give 9 crucial bits of context, not just on the GPT 5.4 drop but on the background to the meltdown between the Pentagon and Anthropic. What does this say about the state of AI progress, your job, and what is next. Check out my fast-growing (!) app, free to use, and code INSIDER15 for 15% off paid tiers: https://lmcouncil.ai AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 01:06: GPT 5.4 Breakdown 05:06 - Closing the Loop 06:35 - Spiky Performance 10:31 - Advice 11:32 - Less Encouraging Developments - Fired Like Dogs 17:45 - But Used in Iran GPT 5.4: https://openai.com/index/introducing-gpt-5-4/ Hallucinations: https://artificialanalysis.ai/evaluations/omniscience Investment Banking Bench: https://x.com/bradlightcap/status/2029684672343728452 Move 37: https://x.com/nasqret/status/2029628846518010099 System Card: https://deploymentsafety.openai.com/gpt-5-4-thinking/gpt-5-4-thinking.pdf Prediction Market Scandal: https://www.wired.com/story/openai-fires-employee-insider-trading-polymarket-kalshi/ GPT 5.3 Instant: https://openai.com/index/gpt-5-3-instant/ GDPVal: https://openai.com/index/gdpval/ Claude in Iran: https://www.washingtonpost.com/technology/2026/03/04/anthropic-ai-iran-campaign ‘Like Dogs’: https://x.com/AndrewCurran_/status/2029605783311470679 Altman leak: https://www.cnbc.com/2026/03/03/sam-altman-tells-openai-staff-operational-decisions-up-to-government.html Original 2024 Switch: https://archive.fo/20240116172526/https://www.bloomberg.com/news/articles/2024-01-16/openai-working-with-us-military-on-cybersecurity-tools-for-veterans#selection-6173.83-6173.226 Amodei Original Memo: https://www.theinformation.com/articles/read-anthropic-ceos-memo-attacking-openais-mendacious-pentagon-announcement?rc=sy0ihq Anthropic Apology: https://www.anthropic.com/news/where-stand-department-war OpenAI Employee Reaction: https://x.com/tszzl/status/2029334980481212820 DoD Suppler Risk: https://www.cnbc.com/amp/2026/03/05/anthropic-pentagon-ai-claude-iran.html Atlantic Exclusive: https://archive.fo/20260301152646/https://www.theatlantic.com/technology/2026/03/inside-anthropics-killer-robot-dispute-with-the-pentagon/686200/#selection-941.61-941.212 No Negotiation: https://x.com/USWREMichael/status/2029754965778907493 $20B Doubling: https://archive.ph/20260304111124/https://www.bloomberg.com/news/articles/2026-03-03/anthropic-nears-20-billion-revenue-run-rate-amid-pentagon-feud March 2022 Interview: https://www.youtube.com/watch?v=uAA6PZkek4A https://lmcouncil.ai/ Non-hype Newsletter: https://signaltonoise.beehiiv.com/

  • S5 · E9
    February 27 · 13 min

    Deadline Day for Autonomous AI Weapons & Mass Surveillance

    Will Anthropic be forced to make a version of Claude for war? And does a new paper expose the risks of Claude agents, in both OpenClaw and the field of war? Plus, 5 more twists in the story of the Pentagon versus Anthropic + some AI lab employees, and a petition that could change everything, or nothing... Check out my fast-growing (!) app, free to use, and code INSIDER15 for paid tiers: https://lmcouncil.ai AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 00:44 - Deadline Day + Petition 02:42 - Twist 1: Existing Deal 03:26 - Twist 2: Existing Policy 04:21 - Twist 3: Twin Threats 05:54 - Twist 4: Interesting Objections 11:32 - Twist 5: Anthropic’s Dropped Policy Dario Statement: https://www.anthropic.com/news/statement-department-of-war Google/OpenAI Petition: https://notdivided.org/ Axios on Amodei Rejection: https://www.axios.com/2026/02/26/anthropic-rejects-pentagon-ai-terms FT on US Threat: https://www.ft.com/content/11d27612-d6c5-4cf7-94dd-f65603549b7f Politico on Latest: https://archive.ph/20260227013117/https://www.politico.com/news/2026/02/26/incoherent-hegseths-anthropic-ultimatum-confounds-ai-policymakers-00800135 The Verge on Current Deal: https://www.theverge.com/ai-artificial-intelligence/883456/anthropic-pentagon-department-of-defense-negotiations Anthropic RSP change: https://www.anthropic.com/news/responsible-scaling-policy-v3 Time Magazine on RSP: https://time.com/7380854/exclusive-anthropic-drops-flagship-safety-pledge/ Agent of Chaos Paper: https://x.com/NatalieShapira/status/2026062499599319526 AI Agent Reliability Paper: https://arxiv.org/pdf/2602.16666 My Patreon Video: https://www.patreon.com/posts/real-mystery-ai-151647211 Patreon Documentary: https://www.patreon.com/posts/our-new-age-of-133960279 Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • S5 · E8
    February 20 · 18 min

    Gemini 3.1 Pro and the Downfall of Benchmarks: Welcome to the Vibe Era of AI

    Do we have a new best AI model, or do we have the downfall of benchmarks in general, as a way of capturing machine intelligence? Full breakdown of Gemini 3.1 Pro, guest-starring the new Sonnet 4.6, plus analysis from 7 papers/posts that will give you much needed context. Oh, and a new record on Simple Bench! https://epoch.ai/ai-explained-datacenters Check out my fast-growing (!) app, free to use, and code INSIDER15 for Pro: https://lmcouncil.ai AI Insiders ($9!): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 00:30 - Post-training Dominance 04:00 - ARC-AGI 2 Caveat 05:54 - Simple Bench Record 08:22 - Hallucination Caveat 10:05 - Model Card 11:12 - Exponential Coming 12:20 - Amodei on Generalizing 15:10 - One True Benchmark? 17:02 - Other Metrics… Gemini 3.1 Model Card: https://storage.googleapis.com/deepmind-media/Model-Cards/Gemini-3-1-Pro-Model-Card.pdf Release: https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/ Where are Agents deployed?: https://www.anthropic.com/research/measuring-agent-autonomy Newsletter Post: https://signaltonoise.beehiiv.com/p/4-ai-numbers-that-surprised-me-this-week Hallucination AA: https://artificialanalysis.ai/evaluations/omniscience Melanie Mitchell: https://x.com/MelMitchell1/status/2022738363548340526 ARC-AGI-2: https://x.com/arcprize/status/2024522812728496470/photo/1 Chollet on Agentic Coding and ML: https://x.com/fchollet/status/2024519439140737442 METR Caveat: https://metr.org/notes/2026-01-22-time-horizon-limitations/ Talaas Fast: https://chatjimmy.ai/ Amodei Interview Continual learning: https://www.dwarkesh.com/p/dario-amodei-2?open=false#%C2%A7002942-is-continual-learning-necessary-how-will-it-be-solved Metaculus FutureEval: https://www.metaculus.com/futureeval/ Next Vid to Watch: https://www.patreon.com/posts/what-you-need-to-150647292 Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

  • S5 · E7
    February 6 · 19 min

    The Two Best AI Models/Enemies Just Got Released Simultaneously

    The two models that you will hear discussed for at least the next two months - Claude Opus 4.6 and GPT 5.3 Codex - just got released within 26 mins or each other. The full breakdown of around 250 pages of reports, with just the most interest moments, from the battle of which is best, Claude personhood, the surprising misbehaviour of Opus 4.6, and much more https://assemblyai.com/aiexplained Check out my fast-growing (!) app, free to use, and code INSIDER15 for Pro: https://lmcouncil.ai AI Insiders ($9): https://www.patreon.com/AIExplained Chapters: 00:00 - Introduction 00:54 - Self-improvement? 02:44 - Knowledge Work 05:30 - Overly agentic behaviour 09:12 - Who Shouldn’t Use Claude Opus 11:39 - Step-change? 15:09 - Claude’s ‘Personhood’ Hassabis Roadmap: https://www.patreon.com/posts/hassabis-roadmap-149750869 Release of Opus 4.6: https://www.anthropic.com/news/claude-opus-4-6 212 Page System Card: https://www-cdn.anthropic.com/0dd865075ad3132672ee0ab40b05a53f14cf5288.pdf Claude Code Tip: https://x.com/bcherny/status/2019475897691124107 GPT Codex 5.3: https://openai.com/index/introducing-gpt-5-3-codex/ System Card: https://openai.com/index/gpt-5-3-codex-system-card/ Browse Comp: https://arxiv.org/pdf/2504.12516v1 Finance Agent: https://www.vals.ai/benchmarks/finance_agent Terminal Bench 2: https://arxiv.org/pdf/2601.11868 Vending Bench: https://andonlabs.com/blog/opus-4-6-vending-bench My X post: https://x.com/AIExplainedYT/status/2016851303436095647 Anthropic Apology: https://x.com/ch402/status/2014066134194995256/photo/1 Altman rebuttal: https://x.com/sama/status/2019139174339928189 https://x.com/sama/status/2019140276246442089 4% of GitHub: https://x.com/dylan522p/status/2019490550911766763 Non-hype Newsletter: https://signaltonoise.beehiiv.com/ Podcast: https://aiexplainedopodcast.buzzsprout.com/

Showing 1–20 of 26 episodes