EP 100. Claude Mythos, Fable 5, 그리고 다음 국면은?
AI Frontier reaches its 100th episode. What began three years ago with GPT-4 and "Sparks of AGI" has now become a conversation about Claude Mythos, Fable 5, RSI, loop engineering, sovereign AI, and independent foundation models. In this episode, we discuss what Anthropic's Fable 5 and Mythos-class model reveal about the next stage of frontier AI. The competition is no longer just about benchmark performance. It increasingly revolves around safety layers, post-training, dataset engineering, test-time compute, pricing, and AI as a strategic asset. We also talk about the current mood in Silicon Valley, how young frontier startups interact with capital, how AI is spreading into biology and drug discovery, and what opportunities Korea may have in the foundation model era. Finally, we reflect on what AI Frontier should become after episode 100, including new sessions around Palantir FDEs and the changing nature of engineering in the agent era. Korean transcript: https://aifrontier.kr/ko/episodes/ep100 Chapters 00:00 100 episodes of AI Frontier and three years of change 03:45 Scaling and the new importance of post-training 06:08 Dataset engineering behind frontier models 07:49 How AI spreads into biology and other domains 12:07 Startup ambition and capital at frontier speed 14:27 Frontier networks and the power of information flow 15:26 The security question behind Mythos and Fable 16:38 What the names Mythos and Fable signal 18:08 Dario Amodei's timeline and the road to RSI 22:03 AI as a strategic asset and market signal 24:19 Loop engineering as a new language for agents 25:55 Fable 5 pricing and test-time compute 30:06 Fable API costs and the commercial logic of loop engineering 34:09 RSI and Anthropic's self-improvement trajectory 35:49 Compute scale and the sovereign AI race 41:35 Korea's foundation model opportunity 43:06 Fable 5 vibe checks and human reflection 48:31 Engineering after Claude Code and agent workflows 51:02 The role of the AI Frontier community 53:10 Plans after episode 100

EP 102. Lessons from San Francisco: "Everyone is crazy"

EP 98. In the Age When AI Executes, What Remains for Humans Is 'Intent' (Hashed CEO Seojoon Kim)

EP 96. LLM Inference Infrastructure and Token Economics

Keynote: After the AI Hype – What’s Real, and What’s Next - Richard Campbell - 2026

EP 91. AI from the Business Perspective — 26.Q1 Update

EP 99. Opus 4.8 Launch, Today's AI Competition, and Human Work

Ilya Sutskever – We're moving from the age of scaling to the age of research

EP 94. Claude Opus 4.7 and the low-hanging fruit

EP 95. Reading the DeepSeek-V4 Paper

Andrej Karpathy: From Vibe Coding to Agentic Engineering w/ Stephanie Zhan

EP 97. People in the Age of AI Psychosis

EP 59. AGI가 오기 전 마지막 사업 기회 - AI창업가 김민석 편

Billionaire's WARNING: I'm SELLING. The Crash Is Already Here!

한 분이라도 더 보셨으면 좋겠습니다 (앤트로픽 풀다큐 번역)

Head of Claude Code: What happens after coding is solved | Boris Cherny

Skill Issue: Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AI

Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit

How GPT, Claude, and Gemini are actually trained and served – Reiner Pope

What rebuilding AlphaGo teaches us about self-play, RL, and future of LLMs - Eric Jang
![[모두콘 2025] 기술창업 6번을 통해서 배운 AI 시대의 기회 - 노정석(비팩토리 대표)](https://i.ytimg.com/vi/PC36opTZ7ow/hqdefault.jpg?sqp=-oaymwEjCNACELwBSFryq4qpAxUIARUAAAAAGAElAADIQj0AgKJDeAE=&rs=AOn4CLABWciQozMXcrDBWl-blXJEn7c92g)
