In this episode of Alephic Research, we dive into DeepSeek AI’s meteoric rise as hedge-fund veteran Liang Wen Feng and a team of fresh PhD graduates turned U.S. GPU export curbs into a software-centric breakthrough—R1, an open-source reasoning model that cost just $5.6 M to train yet wiped nearly $600 billion off Nvidia’s market cap in a single day. We unpack the clever mixture-of-experts architecture, reinforcement-learning training, and viral open-source launch that slashed usage costs by up to 96%. Along the way, we draw big-picture lessons on how constraints, lean teams, and shared innovation can upend tech giants and democratize AI.