In March 2016 a machine beat one of the best Go players in the world. The match became a turning point for AI. One move in particular, the 37th, still shapes how we think about what machines can find on their own.
What happened in the AlphaGo match?
DeepMind's AlphaGo faced the South Korean grandmaster Lee Sedol in Seoul over five games in March 2016. AlphaGo won 4 to 1. Go has more possible positions than there are atoms in the universe, and humans had refined its strategy over roughly 3,000 years. A machine reaching the top of that game arrived years earlier than most experts had predicted.
Why move 37 stunned the experts
In the second game AlphaGo played its 37th move on the fifth line, far from where tradition would place a stone at that stage. Commentators first read it as an error. The move turned out to be brilliant and helped AlphaGo win the game. It showed that a system trained with reinforcement learning can find ideas outside the human playbook, beyond the data it learned from.
What AlphaGo teaches about today's AI
The lesson travels well into business. In a domain with clear rules and good data, AI can surface options a person would dismiss. The catch is the setup. AlphaGo worked because the problem was sharply defined and the feedback was immediate. Most business problems lack that clarity, which is why the same brilliance rarely shows up in a vague, open-ended project.
How to apply the AlphaGo lesson
Pick a problem with clear rules and a measurable outcome. Give the AI clean data and a tight question. Let it propose, and keep a human to judge the result. Lee Sedol won a single game with his own move 78, a reminder that the strongest setups pair machine search with human judgement.
Further reading
These articles help with the next decision.
- Is AI Superintelligence Realistic? A Sober Look
- Understanding AI Agents: Practical Insights
- AI Jobs 2030: Which Careers Stay and Which Disappear
Who won the AlphaGo match?
DeepMind's AlphaGo beat Lee Sedol 4 to 1 in Seoul in March 2016. The single human win went to Lee Sedol with his famous move 78.
What was special about move 37?
AlphaGo placed a stone on the fifth line, where Go tradition expects none at that stage. Commentators thought it was a mistake. It proved decisive.
What does AlphaGo teach about AI in business?
In a field with clear rules and good data, AI finds unusual, strong options. Most business problems lack that clarity, so a sharply defined use case is what matters.
Can AI really invent something new?
In clearly defined domains, yes. AlphaGo found moves beyond the human playbook. Outside such domains, the human stays in charge of judgement and choice.