Sam Altman Finally Broke His Silence on GPT-5 Backlash!

Video Description

Let me know, what do you think of GPT-5 so far? I think GPT-5 was what I expected but not what I hoped for, useful, fast and cheap but not a monumental leap. The International math Olympiad winners are the next generation and we'll probably see them in a few months, vastly more capable, smarter and also more expensive. GPT-5's public showing disappointed many, shifting market sentiment. Google's AI model, Gemma, defied expectations, impacting the AI landscape and causing a significant shift in market odds. This shift highlights the rapid changes and surprises within the **ai news** and **technology news** as the **future of ai** unfolds, showing that even giants like **openai** can face setbacks. The **ai model** landscape is constantly evolving. *************** 0:00 Overview 2:39 The Reality of GPT-5 Launch 8:05 OpenAI's Smarter Internal Model 12:12 New HLE World Record 13:55 DeepMind and MIT: Is AI Hitting A Wall? 15:43 Google's Insane Image Model (Nano Banana) 17:01 Artificial Gems *************** Related Links Nano Banana in lmarena https://lmarena.ai/ Market & leaderboards Polymarket event — Which company has the best AI model by end of August: https://polymarket.com/event/which-company-has-best-ai-model-end-of-august Chatbot Arena leaderboard: https://lmarena.ai/leaderboard OpenAI / GPT-5 OpenAI post-launch update thread: https://x.com/OpenAI/status/1956461718097494196 Sam Altman on rollout issues & capacity constraints (Yahoo Finance): https://finance.yahoo.com/news/sam-altman-admits-openai-totally-172345973.html “Issues after release” note: https://x.com/sama/status/1956483306951938134 GPT-5 announcement page (pricing/features): https://openai.com/index/introducing-gpt-5/ CNBC on enterprise uptake: https://www.cnbc.com/2025/08/14/gpt-5-openai-ai-enterprise.html System prompts (example leak – Claude): https://github.com/asgeirtj/system_prompts_leaks/blob/main/claude.txt Google / Gemini / Gemma / Image models Gemini 2.5 DeepThink blog: https://blog.google/products/gemini/gemini-2-5-deep-think/ Gemma 3 (270M on-device): https://deepmind.google/models/gemma/gemma-3/ Image model examples thread: https://x.com/madpencil_/status/1958156444471308727 Olympiad winners & new techniques Reuters on IMO gold wins context: https://www.reuters.com/technology/artificial-intelligencer-why-ais-math-gold-wins-matter-2025-07-24/ Sam Altman on experimental IMO-level models: https://x.com/sama/status/1946569252296929727 Reflective test-time scaling paper (Metastone/Xbai): https://arxiv.org/pdf/2507.01951 Caesar & HLE Caesar announcement: https://x.com/caesar_data/status/1956385408834470088 Caesar site: https://www.caesar.xyz/ HLE evaluations overview: https://artificialanalysis.ai/evaluations/humanitys-last-exam “Peak AI” & scaling laws “AI has peaked?” article: https://futurism.com/scientists-worried-ai-pleateau DeepMind + MIT scaling paper (sparse vs dense): https://people.csail.mit.edu/suvinay/pubs/2025.journey.iclr.pdf Medical results Emory/Winship oncology paper on GPT-5: https://arxiv.org/pdf/2508.08224 Gems & extras Fun demos thread: https://x.com/Rainmaker1973/status/1951649185436180854 Fusion reactor footage (context on energy → AI): https://www.youtube.com/watch?v=0RwBwx6ELko *************** All materials in these videos are used for educational purposes and fall within the guidelines of fair use. No copyright infringement is intended. If you are or represent the copyright owner of materials used in this video and have a problem with the use of said material, please contact me via my email in the "about" page on my channel. ************** #GPT5 #Gemini #LLMs #ChatbotArena #Polymarket #TestTimeScaling #CaesarAI #HLE #Gemma3 #OnDeviceAI #AIResearch #AIImageEditing #DeepMind #MIT #Emory #AIBenchmarks #Agents #CodingAI