
Sign up to save your podcasts
Or


The implosion of OpenAI, a new AGI lab, Sutskever regret and a radical new CEO, the last 48 hours have altered the course of Generative AI. Drawing on multiple exclusive accounts, I’ll endeavour to recreate what happened, get more of the why as well as the when, and give you a clearer sense of the main players.
https://www.patreon.com/AIExplained
Altman Twitter: https://twitter.com/sama?ref_src=twsrc%5Egoogle%7Ctwcamp%5Eserp%7Ctwgr%5Eauthor
Brockman: https://twitter.com/gdb/status/1726530200484372688
Bloomberg new AGI Org: https://www.bloomberg.com/news/articles/2023-11-20/microsoft-says-altman-brockman-will-lead-new-in-house-ai-team?srnd=premium
Atlantic Exclusive Extract: https://www.theatlantic.com/technology/archive/2023/11/sam-altman-open-ai-chatgpt-chaos/676050/
Altman New Chipmaker Middle East: https://www.bloomberg.com/news/articles/2023-11-19/altman-sought-billions-for-ai-chip-venture-before-openai-ouster?fromMostRead=true
The Information Exclusives: https://www.theinformation.com/articles/openai-execs-invite-altman-brockman-to-headquarters-in-push-to-reinstate-them?rc=sy0ihq
https://www.theinformation.com/articles/dozens-of-staffers-quit-openai-after-sutskever-says-altman-wont-return?rc=sy0ihq
https://www.theinformation.com/articles/openais-rivals-rush-to-poach-openai-staff-after-altman-ouster?rc=sy0ihq
https://www.theinformation.com/articles/breaking-sam-altman-will-not-return-as-ceo-of-openai?rc=sy0ihq
Twitter Reports: https://twitter.com/ashleevance/status/1726469283734274338
https://twitter.com/emilychangtv/status/1726468006786859101
https://twitter.com/karpathy/status/1726289070345855126
https://twitter.com/satyanadella/status/1726516824597258569
https://twitter.com/elonmusk/status/1726376406785925566
https://twitter.com/ryan_t_lowe/status/1726565898742645104
https://twitter.com/ada_rob/status/1726570956305707265
Altman Board Comments: https://twitter.com/pitdesi/status/1725912296910921900
Emmett Sheer: https://twitter.com/eshear/status/1726526112019382275
https://twitter.com/liron/status/1672986864297578501
https://twitter.com/eshear/status/1703148152138977363
Sutskever: https://www.youtube.com/watch?v=Ft0gTO2K85A&t=2322s
https://twitter.com/ilyasut/status/1726590052392956028 Non-Hype, Free Newsletter: https://signaltonoise.beehiiv.com/
Learn more about your ad choices. Visit megaphone.fm/adchoices
Calmest before the storm? Whatever analogy you want to use, things had gotten quiet toward the end of 2024. But then tonight we got Genie 2, and a series of scheduled announcements from OpenAI...
Assembly AI Speech to Text: https://www.assemblyai.com/?utm_source=youtube&utm_medium=influencer&utm_campaign=ai_explained
Sora is soon here, and o1, but I dive deeper into what it all means and whether reliability is on a path to being solved, ft: two recent papers. Plus Kling Motion Brush, Simple Bench QwQ update and much more.
AI Insiders: https://www.patreon.com/AIExplained
Chapters:
00:43 - OpenAI 12 Days, Sora Turbo, o1
03:06 - Genie 2
08:26 - Jensen Huang and Altman Hallucination Predictions
09:45 - Bag of Heuristics Paper
11:40 - Procedural Knowledge Paper
13:02 - AssemblyAI Universal 2
13:45 - SimpleBench QwQ and Chinese Models
14:42 - Kling Motion Brush
Genie 2: https://deepmind.google/discover/blog/genie-2-a-large-scale-foundation-world-model/
Jim Cramer: https://x.com/jimcramer/status/1864068878692675625
Give Us Full o1: https://x.com/tszzl/status/1863882905422106851
Verge Scoop: https://x.com/tomwarren/status/1864326361415925861
O1 Learning to Reason Benchmarks: https://openai.com/index/learning-to-reason-with-llms/
SIMA AI: https://arxiv.org/pdf/2404.10179
Genie Paper: https://arxiv.org/pdf/2402.15391
My Video on Genie: https://www.youtube.com/watch?v=gGKsfXkSXv8
Oasis Minecraft: https://x.com/risphereeditor/status/1852619965511204974
LLMs Procedural Knowledge Paper: https://arxiv.org/pdf/2411.12580
Bag of Heuristics Paper: https://arxiv.org/pdf/2410.21272
Jensen Huang Hallucinations: https://www.tomshardware.com/tech-industry/artificial-intelligence/jensen-says-we-are-several-years-away-from-solving-the-ai-hallucination-problem-in-the-meantime-we-have-to-keep-increasing-our-computation
DeepSeek Interview: https://www.chinatalk.media/p/deepseek-ceo-interview-with-chinas
Kling Motion Brush: https://klingai.com/image-to-video
Tim Rocktaschel Book: https://geni.us/ArtificialIntelligence
The 8 Most Controversial Terms in AI: https://imp.i384100.net/m57g3M
Non-hype Newsletter: https://signaltonoise.beehiiv.com/
Podcast: https://aiexplainedopodcast.buzzsprout.com/
Many people expense AI Insiders for work. Feel free to use the Template in the 'About Section' of my Patreon. https://www.patreon.com/AIExplained
Learn more about your ad choices. Visit megaphone.fm/adchoices
The 9 new AI developments that I genuinely feel will be the most impactful, from insane video dubbing and translation to Apple Ajax and Google Gemini leaks, Roblox AI to an Open Interpreter Demo.
https://www.metaculus.com/ai/?utm_source=ai_explained&utm_campaign=ai_explained
https://www.patreon.com/AIExplained
HeyGen Avatar 2.0 Video Translate: https://labs.heygen.com/video-translate
Altman Original Testimony: https://www.youtube.com/shorts/OXaTpBZGlSo
Open Interpreter: https://twitter.com/hellokillian/status/1699156860073640038
https://openinterpreter.com/
Sam Altman Interview This Week, Norges Bank: https://www.youtube.com/watch?v=O77UyYK51s4
Language Models As Optimizers: https://arxiv.org/pdf/2309.03409.pdf
Bipartisan AI Act: https://twitter.com/SenBlumenthal/status/1700147410880569475/photo/1
Suleyman 80,000 Hours: https://80000hours.org/podcast/episodes/mustafa-suleyman-getting-washington-and-silicon-valley-to-tame-ai/
New York Times IRS AI: https://www.nytimes.com/2023/09/08/us/politics/irs-deploys-artificial-intelligence-to-target-rich-partnerships.html
Google Gemini News: https://www.theinformation.com/articles/google-nears-release-of-gemini-ai-to-rival-openai?rc=sy0ihq
Palm 2 Technical Report: https://ai.google/static/documents/palm2techreport.pdf
Meta Llama 3: https://www.wsj.com/tech/ai/meta-is-developing-a-new-more-powerful-ai-system-as-technology-race-escalates-decf9451
https://twitter.com/agikoala/status/1695125016764157988
Roblox AI: https://twitter.com/rowancheung/status/1701626833970794989
Smell2Text: https://twitter.com/DrJimFan/status/1701611251376497046
NExT-GPT: https://arxiv.org/pdf/2309.05519.pdf
Apple Ajax GPT: https://www.theinformation.com/articles/apple-boosts-spending-to-develop-conversational-ai?rc=sy0ihq
AI Art: https://twitter.com/MrUgleh/status/1702041188482658758/photo/4
https://discord.gg/WxxeHG7HCV
https://www.patreon.com/AIExplained Non-Hype, Free Newsletter: https://signaltonoise.beehiiv.com/
Learn more about your ad choices. Visit megaphone.fm/adchoices
A huge new insider report on GPT Vision is released by Microsoft and just in the last few hours the RT-X series is dropped by Google in Robotics. I will not only break down the 160+ page report on what GPT-4V can do and what it can't, including new use cases, prompting techniques and failure modes, I'll also go through the full RT-2-X and RT-1-X demo, which I am calling the GPT2 moment for robotics. Plus its huge new opensource Open X-Embodiment dataset.
https://www.patreon.com/AIExplained
RT-X Series: Paper: https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/scaling-up-learning-across-many-different-robot-types/Open_X_Embodiment__Robotic_Learning_Datasets_and_RT_X_Models.pdf
Blog: https://www.deepmind.com/blog/scaling-up-learning-across-many-different-robot-types?
Github: https://robotics-transformer-x.github.io/
utm_source=twitter&utm_medium=social&utm_campaign=RT-X
The Dawn of LLMS: https://arxiv.org/abs/2309.17421
GPT 4 vs Pali 17B: https://openai.com/research/gpt-4
GPT V Inception: https://twitter.com/markchen90/status/1708867434491380057/photo/1
PaLI-X: https://arxiv.org/pdf/2305.18565.pdf
PromptBreeder: https://arxiv.org/pdf/2309.16797.pdf
Wozniak Definition: https://en.wikipedia.org/wiki/Artificial_general_intelligence
Gobi from The Information: https://www.theinformation.com/articles/openai-hustles-to-beat-google-to-launch-multimodal-llm?rc=sy0ihq
https://www.patreon.com/AIExplained Non-Hype, Free Newsletter: https://signaltonoise.beehiiv.com/
Learn more about your ad choices. Visit megaphone.fm/adchoices
In this video, I will not only show you how to get smarter results from GPT 4 yourself, I will also showcase SmartGPT, a system which I believe, with evidence, might help beat MMLU state of the art benchmarks.
This should serve as your ultimate guide for boosting the automatic technical performance of GPT 4, without even needing few shot exemplars.
The video will cover papers published in the last 72 hours, like Automatically Discovered Chain of Thought, which beats even 'Let's think Step by Step' and the approach that combines it all.
Yes, the video also touches on the OpenAI DeepLearning Prompt Engineering Course but the highlights come more from my own experiments using the MMLU benchmark, and drawing upon insights from the recent Boosting Theory of Mind, and Let’s Work This Out Step By Step, and combining it with Reflexion and Dialogue Enabled Resolving Agents.
Prompts Frameworks:
Answer: Let's work this out in a step by step way to be sure we have the right answer
You are a researcher tasked with investigating the X response options provided. List the flaws and faulty logic of each answer option. Let's work this out in a step by step way to be sure we have all the errors:
You are a resolver tasked with 1) finding which of the X answer options the researcher thought was best 2) improving that answer, and 3) Printing the improved answer in full. Let's work this out in a step by step way to be sure we have the right answer:
Automatically Discovered Chain of Thought: https://arxiv.org/pdf/2305.02897.pdf
Karpathy Tweet: https://twitter.com/karpathy/status/1529288843207184384
Best prompt: Theory of Mind: https://arxiv.org/ftp/arxiv/papers/2304/2304.11490.pdf
Few Shot Improvements: https://sh-tsang.medium.com/review-gpt-3-language-models-are-few-shot-learners-ff3e63da944d
Dera Dialogue Paper: https://arxiv.org/pdf/2303.17071.pdf
MMLU: https://arxiv.org/pdf/2009.03300v3.pdf
GPT 4 Technical report: https://arxiv.org/pdf/2303.08774.pdf
Reflexion paper: https://arxiv.org/abs/2303.11366
Why AI is Smart and Stupid: https://www.youtube.com/watch?v=SvBR0OGT5VI&t=1s
Lennart Heim Video: https://www.youtube.com/watch?v=7EwAdTqGgWM&t=67s
https://www.patreon.com/AIExplained Non-Hype, Free Newsletter: https://signaltonoise.beehiiv.com/
Learn more about your ad choices. Visit megaphone.fm/adchoices
Google just announced at least 12 things that are each worthy of a video, but here are the top I/O highlights. From Veo 3 to Deep Research now being useable, Deep Think breaking records to Gemini Diffusion, Gemini 2.5 Flash changing how AI is priced and GemmaVerse, SynthID Detector and Imagen 4. And even this intro is missing other announcements covered in the vid! And yes, they’ll be plenty of Veo 3 clips to enjoy…
https://80000hours.org/aiexplained
AI Insiders ($9!): https://www.patreon.com/AIExplained
Chapters:
00:00 - Introduction
00:48 - Veo 3
02:10 - Gemini 2.5 Flash
03:13 - Universal Assistant
03:47 - Usage Skyrockets + OpenAI dig
04:51 - Gemini Pro Deep Think
06:21 - Overviews and AI Mode
07:26 - Deep Research Updates (new) + Jules
08:53 - Make and Deploy Apps with Gemini
09:12 - Imagen 4
10:00 - Gemini Diffusion
11:46 - Try It On
12:17 - SynthID Detector
13:30 - GemmaVerse, SignGemma, Gemma3n, medGemma
14:24 - Outro + Clips
Event: https://www.youtube.com/watch?v=o8NiE3XMPrM
Ntaive Audio: https://aistudio.google.com/generate-speech
Gemini Diffusion: https://deepmind.google/models/gemini-diffusion/#capabilities
New Gemini 2.5 Flash: https://deepmind.google/models/gemini/flash/
SignGemma (See end of this vid): https://www.youtube.com/watch?v=GjvgtwSOCao
Deep Think: https://blog.google/technology/google-deepmind/google-gemini-updates-io-2025/#flash-improvements
Google Parallel Sampling: https://www.patreon.com/posts/next-level-good-127441188
Price Plans: https://blog.google/products/google-one/google-ai-ultra/
Imagen 4 Benchmarks: https://deepmind.google/models/imagen/
Jules: https://jules.google/
SynthID Detector: https://blog.google/technology/ai/google-synthid-ai-content-detector/
Veo 3 Benchmarks: https://deepmind.google/models/veo/evals/
MedGemma: https://deepmind.google/models/gemma/medgemma/
Build Apps: https://aistudio.google.com/apps
Non-hype Newsletter: https://signaltonoise.beehiiv.com/
Podcast: https://aiexplainedopodcast.buzzsprout.com/
Learn more about your ad choices. Visit megaphone.fm/adchoices
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/
Learn more about your ad choices. Visit megaphone.fm/adchoices
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/
Podcast: https://aiexplainedopodcast.buzzsprout.com/
Learn more about your ad choices. Visit megaphone.fm/adchoices
Time for something different - a tour of ChatGPT getting things wrong, including a whole new category of errors that you might find illuminating, concerning or just entertaining.
From investigating whether GPT 4 does indeed have theory of mind, to how easily it is jailbroken, to testing Inflection 1, Bard and Claude on the same puzzle that flummoxes ChatGPT to arguing that GPT 4 will just double down on bad logic, this video showcases GPT getting irrational.
Inverse Scaling Paper: https://arxiv.org/pdf/2306.09479.pdf
Theory of Mind Perturbations: https://arxiv.org/pdf/2302.08399.pdf
Unfaithful Explanations: https://arxiv.org/pdf/2305.04388.pdf
Decoding Trust: https://arxiv.org/pdf/2306.11698.pdf
Inflection Memo: https://inflection.ai/assets/Inflection-1.pdf
Heypi from Inflection: https://heypi.com/talk
Jailbreaks Will Always be Possible: https://arxiv.org/pdf/2304.11082.pdf
https://www.patreon.com/AIExplained Non-Hype, Free Newsletter: https://signaltonoise.beehiiv.com/
Learn more about your ad choices. Visit megaphone.fm/adchoices
Phi-2 is a tiny model that could fit on a phone, but it outperforms huge language models like Llama 2. I explain more about how it was made and what it means. Then we see Imagen-2, the most stunning text-to-image yet, at least according to Google' images. We then glimpse Optimus 2, smaller in the sense that it's 10kg lighter! With more degrees of freedom, it's movements look a lot more humanoid. And then the full launch of AI Insiders, plus a recap of why we shouldn't use the MMLU to 2 decimal places!
https://www.patreon.com/AIExplained
phi2 now on HuggingFace: https://huggingface.co/microsoft/phi-2
Bubeck Video, min 19: https://www.youtube.com/watch?v=24O1KcIO3FM
Phi 2: https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/
Shital Shah: https://twitter.com/sytelus/status/1734882570603753814
Shoggoth: https://twitter.com/norabelrose/status/1702488701782352097
Mamba 3B: https://www.together.ai/blog/mamba-3b-slimpj
Phi 1.5B: https://arxiv.org/abs/2309.05463
Phi 1: https://arxiv.org/abs/2306.11644
Microsoft Prompting: https://www.microsoft.com/en-us/research/blog/steering-at-the-frontier-extending-the-power-of-prompting/
SmartGPT Video: https://www.youtube.com/watch?v=hVade_8H8mE&t=554s
The Information: https://www.theinformation.com/articles/how-microsoft-is-trying-to-lessen-its-addiction-to-openai-as-ai-costs-soar?rc=sy0ihq
Imagen 2: https://twitter.com/googledeepmind/status/1734954295655534780
https://deepmind.google/technologies/imagen-2/?utm_source=twitter&utm_medium=social
https://twitter.com/elonmusk/status/1734763060244386074
Greg Technology: https://twitter.com/TheRundownAI/status/1734544659953623509
Swyx: https://www.latent.space/
AI Engineer: https://youtube.com/@aiDotEngineer?si=HhMdIwClGS3re9UK
Shawn Wang: https://x.com/swyx?s=09
https://www.patreon.com/AIExplained Non-Hype, Free Newsletter: https://signaltonoise.beehiiv.com/
Learn more about your ad choices. Visit megaphone.fm/adchoices