AlphaGo
DeepMind's system that achieved superhuman play in the game of Go, famous for the novel 'move 37.'
As defined in Demis Hassabis #2
Where it comes up · 24
- The future of intelligence | Demis Hassabis (Co-founder and CEO of DeepMind)6 claims
- Sir Demis Hassabis on The Future of Knowledge | Institute for Advanced Study5 claims
- Demis Hassabis — Scaling, superhuman AIs, AlphaZero atop LLMs, AlphaFold4 claims
- Google DeepMind C.E.O. Demis Hassabis on Living in an A.I. Future | EP 1373 claims
- Demis Hassabis — Scaling, superhuman AIs, AlphaZero atop LLMs, AlphaFold3 claims
- Human Dignity in the Age of AI | Yuval Noah Harari2 claims
- Ex-Google Exec Reveals The Shocking Truth About AI with Mo Gawdat | Bad Decisions Podcast #642 claims
- Rodney Brooks on Artificial Intelligence2 claims
and 16 more
Most important claims · 10 of 47Search all claims for “AlphaGo” →
- Hassabis's team believed that pattern-matching human play would not reach the level of the top human player, so the system needed to discover novel, non-human moves; the early version learned human strategies, then discarded some that could be counteracted, and became alien in style as it strengthened.#416: The Relentless Missionary Creating AGI: Demis Hassabis
- There is a trade-off between model quality and search: AlphaZero plays beyond world-champion level using only tens of thousands of evaluated positions versus millions for brute-force systems like Deep Blue, while human grandmasters use only a few hundred—because richer, more accurate world models let you pick which nodes to expand, so improving models makes search more efficient and lets you go further.Demis Hassabis — Scaling, superhuman AIs, AlphaZero atop LLMs, AlphaFold
- 2016 was the true breakthrough year for AI, not 2023, because two things happened: (1) reinforcement learning became practical, moving from punishing failed AI (culling) to correcting mistakes like a teacher, vastly reducing the need for labeled data and enabling exploration, and (2) AlphaGo defeated world champion Lee Sedol, and more importantly AlphaGo Zero beat AlphaGo Master without ever seeing human play by training only against itself.Ex-Google Exec Reveals The Shocking Truth About AI with Mo Gawdat | Bad Decisions Podcast #64
- AlphaGo's victory over the top Go player in Shanghai spurred China into action and catalyzed their AI investment because Go is central to Chinese culture (comparable to chess in the West), and the defeat was significant enough that the Chinese government restricted televised coverage to prevent citizens from witnessing what they saw as a national loss.Eric Schmidt and Craig Mundie with David Rubenstein: Genesis
- AlphaGo was preceded by TD-Gammon where Gerry Tesauro used reinforcement learning and temporal difference learning to beat the world's best backgammon players; in some sense, AlphaGo was merely a scaling up of that process.Richard Sutton – Father of RL thinks LLMs are a dead end
- In 2016, AlphaGo invented a move in the game of Go that had never been seen in 2,500 years of human play because the system was organized to maintain a greater than 50 percent chance of winning and calculated this move as optimal strategy.The AI Revolution Is Underhyped | Eric Schmidt | TED
- Any pattern generated or found in nature can be efficiently discovered and modeled by a classical learning algorithm, because natural systems have structure imposed by evolutionary and selection processes, which creates a low-dimensional manifold a neural network can learn rather than requiring intractable brute-force enumeration.Demis Hassabis #2
- Public interest in chatbots (like ChatGPT) arrived earlier and more intensely than DeepMind anticipated, creating a different research environment with more venture capital and urgency than if systems had developed more quietly.Demis Hassabis — Scaling, superhuman AIs, AlphaZero atop LLMs, AlphaFold
- RL agents like AlphaZero and AlphaGo are active learners whose next actions determine which next data or experience they receive, creating a feedback loop; this active learning, including in realistic simulation environments, is essential for understanding how actions affect the world and for robotics.Demis Hassabis — Scaling, superhuman AIs, AlphaZero atop LLMs, AlphaFold
- AlphaGo discovered moves that no human had ever played, and humans then learned from the AI and updated their understanding of the game, suggesting that AI can not just solve problems but generate novel insights humans can learn from.Sam Altman and Greg Brockman on AI and the Future (Full Audio)