According to @demishassabis, AlphaGo’s iconic Move 37 from the 2016 Lee Sedol match marked a turning point proving that deep learning and reinforcement learning could generalize to real‑world problems, and ideas inspired by these methods remain critical to building AGI; as reported by DeepMind’s CEO on X, the new video thread revisits how policy networks, value networks, and Monte Carlo Tree Search combined to produce non‑intuitive strategies with superhuman outcomes and sparked downstream advances in domains like protein folding and chip design. According to the AlphaGo Nature paper and DeepMind’s official write‑ups, the hybrid RL plus MCTS architecture reduced search breadth while improving evaluation quality, creating a playbook now used in enterprise decision optimization, supply chain planning, and drug discovery. As noted by industry analysis from Nature and DeepMind case studies, Move 37’s legacy informs today’s RL from human feedback and planning‑augmented LLMs, pointing to near‑term business opportunities in operations research, industrial control, and scientific simulation where policy–value abstractions cut compute costs and increase reliability.

According to GoogleDeepMind on X, the company has named its new London building Platform 37 to honor both the city's transport heritage and AlphaGo’s famed Move 37, the breakthrough play that demonstrated superhuman strategy in Go (source: Google DeepMind post on X). As reported by Google DeepMind, the facility signals continued investment in UK-based AI research infrastructure, supporting teams working on frontier models and safety evaluation (source: Google DeepMind post on X). According to Google DeepMind, the branding connects institutional memory of AlphaGo’s novel search and policy network advances with its ongoing multimodal and agent research, reinforcing talent attraction, partnerships, and local ecosystem growth around King’s Cross transport links (source: Google DeepMind post on X).