У нас вы можете посмотреть бесплатно The State of AI in 2025: Reflections with Tengyu Ma (Voyage AI) | People Who Ship Episode 6 или скачать в максимальном доступном качестве, видео которое было загружено на ютуб. Для загрузки выберите вариант из формы ниже:
Если кнопки скачивания не
загрузились
НАЖМИТЕ ЗДЕСЬ или обновите страницу
Если возникают проблемы со скачиванием видео, пожалуйста напишите в поддержку по адресу внизу
страницы.
Спасибо за использование сервиса ClipSaver.ru
Subscribe to MongoDB YouTube→ https://mdb.link/subscribe In the final 2025 episode of People Who Ship, host Apoorva sits down with Tengyu Ma, Chief AI Scientist and CEO/Co-founder of Voyage AI, for a discussion about the most transformative year in artificial intelligence. Following the acquisition of Voyage AI by MongoDB, Tengyu reflects on the evolution of embedding models and re-rankers, explaining why high-quality retrieval remains the backbone of successful AI applications and why it is here to stay for 2026. The conversation explores the explosion of AI agents, the rise of reasoning models (like Open AI's 01 and DeepSeek), and the technical debate between long context windows and Retrieval-Augmented Generation (RAG). Tengyu also shares his personal workflow for 10x productivity, his "hot take" on why human talent—not just compute—is the ultimate moat in AI, and his bold predictions for robotics and faster thinking models in 2026. 00:00:00 - Introduction to the 2025 AI Year in Review 00:00:46 - The Vision Behind Voyage AI & MongoDB Acquisition 00:02:40 - Why RAG is More Effective Than Fine-Tuning 00:04:19 - Core Components: Understanding Embeddings and Re-rankers 00:05:46 - Mission Update: Building a One-Stop AI Developer Solution 00:07:31 - 2025 Highlights: The Rise of Agents & Reasoning Models 00:09:27 - Double Acceleration: Using AI to Build Better AI 00:11:15 - The Real AI Moat: Why Talent Outweighs Compute 00:13:30 - Tengyu’s Favorite Tools & 10x Productivity Hacks 00:15:31 - Trust Levels: Human Verification in AI Workflows 00:17:15 - Debunking the Myth: Why Long Context Won't Kill RAG 00:20:13 - Future Outlook: The Next 3-5 Years of AI Tooling 00:22:20 - Missing Links: What We Need to Reach AGI 00:24:20 - Rapid Fire: DeepSeek, Coding Agents, and Vibe Coding 00:25:29 - 2026 Predictions: Robotics, Video, and Advice for Founders Visit Mongodb.com → https://mdb.link/MongoDB Read the MongoDB Blog → https://mdb.link/Blog Read the Developer Blog → https://mdb.link/developerblog