The Thinking Game | Full documentary | Tribeca Film Festival official selection

The transcript explores the journey and vision of Demis Hassabis and his team at DeepMind as they pursue the development of artificial general intelligence (AGI). The narrative highlights key milestones, including the creation of AlphaGo, which defeated human champions in the complex game of Go, and AlphaFold, which solved the protein folding problem, revolutionizing biological research. The discussion also touches on the ethical implications and potential societal impacts of AGI, emphasizing the need for responsible development and global cooperation. Notable figures like Peter Thiel and Elon Musk are mentioned as early supporters of DeepMind's ambitious goals.

• The interaction between AI and humans, highlighting AI's potential to assist in tasks such as coding.
• AI's ability to learn and adapt, showcasing an open-minded approach to problem-solving.

"Can you help me write code? I was trained to answer questions, but I'm able to learn."

• This quote demonstrates AI's capacity to assist in coding and its potential to learn new skills.

• Rapid advancements in AI technology and the increasing interest and concern surrounding its development.
• The potential impact of AI on various industries and the comparison of its significance to historical technological breakthroughs.

"Trying to build AGI is the most exciting journey in my opinion that humans have ever embarked on."

• This quote underscores the significance of developing Artificial General Intelligence (AGI) and its potential to revolutionize human endeavors.

• The personal goals and motivations behind individuals involved in AI research, particularly the aim to solve AGI.
• The influence of neuroscience on AI development and the pursuit of understanding human intelligence through technology.

"The human brain is the only existence proof we have perhaps in the entire universe that general intelligence is possible at all."

• This quote highlights the importance of the human brain as a model for developing AGI and the inspiration drawn from neuroscience.

• The skepticism and challenges faced by AI researchers in academia and the decision to form a company to pursue AGI.
• The need for substantial investment and the risk associated with developing AGI, akin to a lottery ticket.

"We needed investors who aren't necessarily going to invest because they think it's the best investment decision. They're probably going to invest because they just think it's really cool."

• This quote illustrates the challenge of securing investment for AGI development, emphasizing the need for visionary investors interested in the project's potential rather than immediate returns.

• DeepMind's mission to build a general learning machine and the emphasis on general intelligence and learning.
• The initial stealth mode of DeepMind, operating in secrecy to develop its groundbreaking AI technology.

"We summarize it as you know Deep Mind's mission is to build the world's first general learning machine."

• This quote encapsulates DeepMind's mission to create a versatile AI capable of learning multiple tasks, similar to human intelligence.

• The use of reinforcement learning and deep learning to train AI agents, starting with simple games like Pong.
• The iterative process of improving AI performance, leading to significant breakthroughs in general intelligence.

"We tried doing Pong as a first game. It seemed like the simplest. It hasn't been told anything about what it's controlling or what it's supposed to do."

• This quote describes the initial attempts to train AI using simple games, highlighting the challenges and learning process involved in reinforcement learning.

• The success of AI in mastering various games, demonstrating its ability to learn and adapt to new environments.
• The development of a general AI capable of achieving human-level or superior performance across multiple games.

"We started generalizing to 50 games and we basically created a recipe. We could just take a game that we have never seen before."

• This quote emphasizes the achievement of creating a general AI capable of learning and excelling in diverse gaming environments without prior specific programming.

• The acquisition of DeepMind by Google, providing the resources needed to advance AI research while maintaining independence.
• The focus on using AI for scientific breakthroughs and the urgency of progressing towards AGI.

"After the acquisition, I started mentoring and spending time with Demis and just listening to him. And this is a person who fundamentally is a scientist and a natural scientist."

• This quote highlights the collaboration between DeepMind and Google, emphasizing the scientific focus and independence that DeepMind retained post-acquisition.

• The complexity of the game Go as a benchmark for AI capabilities and the development of AlphGo to tackle this challenge.
• The innovative approach of AlphGo in learning and discovering new strategies, surpassing human players.

"The game of Go has been studied for thousands of years and AlphGo discovered something completely new."

• This quote illustrates the groundbreaking achievement of AlphGo in mastering Go, showcasing AI's potential to innovate beyond human strategies.

• AlphaGo defeated the top-rated player in the world, marking a significant advancement in AI capabilities.
• The event was likened to the "Sputnik moment," signaling a global AI arms race, particularly for China.
• AlphaGo's success led to further developments in AI, such as AlphaZero, which learns without human input.

"When we beat the Chinese guy, the top-rated player in the world."

• This quote highlights the pivotal victory of AlphaGo over the world's best Go player, symbolizing a breakthrough in AI technology.

"For China, AlphaGo was the wakeup call, the Sputnik moment. It launched an AI space race."

• The defeat of a top Chinese player by AlphaGo is compared to the Sputnik launch, emphasizing the urgency and competitive drive it instilled in AI development.

• AlphaZero was developed to learn from scratch without human data, becoming its own teacher.
• The system was designed to play any two-player game, such as chess, showcasing its versatility.
• AlphaZero's ability to reach superhuman levels in a single day demonstrated the rapid learning potential of AI.

"We eventually found a new algorithm, a much more elegant approach to the whole system which actually stripped out all of the human knowledge."

• This quote explains the transition from human-dependent learning to autonomous learning in AI, marking a significant evolution in AI research.

"AlphaZero could start in the morning playing completely randomly and then by tea be superhuman level."

• The rapid progression of AlphaZero from novice to expert highlights the efficiency and potential of AI self-learning.

• Demis Hassabis, a key figure in AI development, had a background in chess, which influenced his interest in AI.
• Despite early success in chess, he questioned the value of dedicating his life to it and shifted focus to AI.
• His personal experiences underscore the connection between strategic games and AI development.

"I was about 12 years old and I was at this international chess tournament in Liechtenstein up in the mountains."

• This anecdote illustrates Demis's early involvement in chess and the high-stakes environment that shaped his perspective on strategic thinking.

"Although I love chess, this is not the right thing to spend my whole life on."

• Demis reflects on his decision to pursue AI over chess, emphasizing his desire to apply his skills to broader challenges.

• AGI aims to create systems that can solve novel problems and learn from diverse environments.
• Simulated environments are used to mimic human learning experiences, facilitating the development of AGI.
• The flexibility and cognitive breadth of AGI are key goals in achieving human-like intelligence in machines.

"The only natural general intelligence we know of as humans, we obviously learn a lot from our environment."

• This quote emphasizes the importance of environmental interaction in developing AGI, drawing parallels to human learning.

"We think that simulated environments are one of the ways to create an AGI."

• Simulated environments are highlighted as crucial for training AGI, providing a controlled setting for learning and experimentation.

• AI agents learn through reinforcement, receiving rewards for positive actions and penalties for negative ones.
• Simulated environments allow agents to experiment and develop diverse skills, akin to human exploration.
• The goal is to create versatile agents capable of adapting to various challenges and tasks.

"The whole algorithm is trying to optimize for receiving as much rewards as possible."

• Reinforcement learning is explained as a process of maximizing rewards, driving AI behavior and decision-making.

"Our goal is to try and build agents which you drop them in, they know nothing, they get to play around in whatever problem you give them."

• The objective of developing adaptable AI agents is to enable them to learn independently and solve a wide range of problems.

• Starcraft is used as a training ground for AI due to its complexity and requirement for diverse skills.
• The game involves continuous decision-making and incomplete information, challenging AI to adapt and strategize.
• Success in Starcraft demonstrates AI's potential to handle real-world, dynamic environments.

"Unlike chess or Go, where players take turns to make moves, in Starcraft, there's a continuous flow of decisions."

• The continuous nature of Starcraft presents a unique challenge for AI, requiring constant adaptation and strategic planning.

"This is the way that we'll get to a much more fluid, more natural, faster, more reactive agent."

• The development of AI in Starcraft aims to create agents that are more responsive and adaptable, reflecting real-world demands.

• The potential misuse of AI in military and surveillance applications raises ethical concerns.
• There is a need for careful consideration and control over AI development to prevent harmful outcomes.
• The approach to AI should prioritize safety and ethical implications, avoiding a "move fast and break things" mentality.

"I'm very worried about the certain ways in which AI will be used for military purposes."

• Concerns about AI's application in military contexts highlight the importance of ethical considerations in its development.

"My view is that the approach to building technology which is embodied by move fast and break things is exactly what we should not be doing."

• The need for a cautious and responsible approach to AI development is stressed, emphasizing long-term safety and ethical standards.

• Demis's early interest in computers and gaming led to a career in AI and game development.
• His experiences at Bullfrog and in the gaming industry shaped his understanding of technology's potential.
• Demis's journey reflects the intersection of gaming, strategic thinking, and AI innovation.

"When I was eight, I bought my first computer with winnings from a chess tournament."

• This quote illustrates Demis's early engagement with technology and the pivotal role it played in his career path.

"I got a feeling of being really at the cutting edge and how much fun that was to invent things every day."

• Demis's enthusiasm for innovation and creativity in technology is highlighted, driving his contributions to AI development.

• Early computer games focused primarily on simple mechanics like shooting.
• The need for innovation led to the creation of games where players could design and build worlds, such as theme parks.

"Wouldn't it be amazing to have a game where you design and build your own theme park?"

• This quote highlights the shift in gaming towards more creative and complex gameplay experiences.

• Theme Park allowed players to create and manage their own amusement parks, introducing elements of simulation and strategy.

"The human player set out the layout of the theme park and designed the roller coaster and set the prices in the chip shop."

• This illustrates the interactive and strategic elements that were novel in gaming at the time.

• The development of AI in games was aimed at creating more engaging and realistic simulations.
• AI was used to simulate human behaviors within the game environment.

"What I was working on was the behaviors of the people. They were autonomous and that was the AI in this case."

• This quote emphasizes the role of AI in enhancing the realism and complexity of game simulations.

• The success of Theme Park demonstrated the potential of AI to transform gaming experiences.

"Theme park actually turned out to be a top 10 title and that was the first time we were starting to see how AI could make a difference."

• This shows the impact of AI on the commercial and critical success of games.

• Discussions began around the potential of AI to be useful beyond just entertainment, with aspirations to solve real-world problems.

"I have this very clear memory of Demis talking about AI in a very different way and a way that we didn't commonly talk about this idea of AI being useful for other things other than entertainment."

• This quote captures the early vision of AI as a tool for broader applications.

• Demis Hassabis expressed a desire to solve AI and explore its full potential.

"I just said to Dennis, what is it you want to do? And he said to me, I want to be the person that solves AI."

• This reflects the ambition and foresight of using AI for significant societal impact.

• Demis Hassabis chose to pursue academia at Cambridge, valuing the intellectual environment over immediate financial gain.

"Peter offered me a million pounds to not go to university. But I had a plan from the beginning. And my plan was always to go to Cambridge."

• This demonstrates the importance placed on education and long-term goals.

• The intersection of computer science and neuroscience was a key area of interest.

"At Cambridge, Demis and myself both had an interest in computational neuroscience and trying to understand how computers and the brains intertwined and linked together."

• This highlights the interdisciplinary approach to understanding intelligence.

• The defeat of chess grandmaster Gary Kasparov by Deep Blue was a pivotal moment in AI.

"When Deep Blue beat Gary Kasparov, that was a real watershed event."

• This signifies the growing capabilities of AI in specific domains, though lacking general intelligence.

• Protein folding was identified as a critical problem with potential to impact medicine and biology significantly.

"If we could just solve protein folding, it could change the world."

• This quote underscores the transformative potential of solving the protein folding problem.

• AI was seen as a necessary tool to tackle the complexity of protein folding.

"I thought it would need AI to do it."

• This reflects the belief in AI's capability to address complex scientific challenges.

• DeepMind's efforts in protein folding aimed to utilize AI for scientific breakthroughs.

"Games were always just a proving ground for our algorithms. The ultimate goal was not just to crack Go and Starcraft. It was to crack real world challenges."

• This highlights the broader vision of applying AI developed in gaming to solve real-world problems.

• The CASP competition was a benchmark for assessing progress in protein folding.

"CASP is a bit like the Olympic Games of protein folding."

• This illustrates the competitive and collaborative nature of scientific advancement in this field.

• Initial results from DeepMind's efforts were promising but highlighted the complexity of the problem.

"We beat the second team in this competition by nearly 50%. But we still got a long way to go before we've solved the protein folding problem."

• This quote reflects both the achievement and the ongoing challenges faced in solving protein folding.

• The journey to solving protein folding involved overcoming significant challenges and managing expectations.

"We were the best in the world at a problem the world's not good at. We knew we sucked."

• This candid acknowledgment points to the reality of scientific research and the iterative nature of progress.

• The ambition to achieve AGI (Artificial General Intelligence) remains a long-term goal.

"How many big breakthroughs do you think are required to get all the way to AGI?"

• This question encapsulates the ongoing quest for comprehensive AI capabilities beyond narrow applications.

• The work at DeepMind reflects a commitment to using AI for scientific and societal advancement.

"My dream has always been to try and make AI-assisted science possible."

• This quote encapsulates the vision of AI as a tool for accelerating scientific discovery and solving complex problems.

• AGI is anticipated to perform any cognitive task a human can do and potentially scale beyond human capabilities.
• The development of AGI is compared to a race, with concerns about its rapid advancement and the lack of preparedness.
• There is a transformational shift in understanding intelligence, with AI agents exhibiting learned cooperative social dynamics.

"Full AGI will be able to do any cognitive task that a person can do, and it will be able to scale potentially far beyond that."

• This quote highlights the potential of AGI to surpass human cognitive abilities.

"It really feels like we're in a race to AGI."

• The speaker expresses urgency and concern over the swift development of AGI.

• The deployment of AGI requires careful consideration of ethical values and societal goals.
• There is a need to embed human values in AI systems to prevent exploitation of human vulnerabilities.
• The rise of AGI parallels historical technological revolutions, necessitating strategies to support those displaced by AI.

"Technology always embeds our values. It's not just technical; it's ethical as well."

• Technology, including AGI, inherently reflects the values of its creators, necessitating ethical considerations.

"There is a parallel between the industrial revolution which was an incredible moment of displacement and the current technological change created by AI."

• The speaker draws a parallel between the industrial revolution and the current AI revolution, emphasizing potential societal disruptions.

• The development of AGI requires global coordination and new institutions to ensure safety and ethical use.
• There are concerns about humanity's declining ability to coordinate on a global scale, which is crucial for managing AGI.

"AGI is coming whether we do it here at Deep Mind or not. It's going to happen. So, we better create institutions to protect us."

• The inevitability of AGI's arrival is acknowledged, with a call for protective institutions.

"It's going to require global coordination. And I worry that humanity is increasingly getting worse at that rather than better."

• The speaker expresses concern over the current state of global coordination necessary for AGI governance.

• AI, specifically AlphaFold, has made significant strides in solving the protein folding problem, which has vast implications for biology and medicine.
• The success of AlphaFold demonstrates the potential of AI to revolutionize scientific research and discovery.

"After half a century, we finally have a solution to the protein folding problem."

• AlphaFold's achievement in solving the protein folding problem marks a significant scientific milestone.

"My guess is every single biological and chemistry achievement will be related to AlphaFold in some way."

• The speaker predicts AlphaFold's impact on future scientific achievements in biology and chemistry.

• The rapid pace of AI innovation is likened to a boulder gathering speed, necessitating responsible stewardship.
• The future world shaped by AI will be radically different, emphasizing the importance of responsible development and deployment.

"The pace of innovation and capabilities is accelerating like a boulder rolling down a hill that we've kicked off and now is continuing to gather speed."

• This metaphor highlights the accelerating momentum of AI advancements and the need for careful management.

"Very clearly, the next generation is going to live in the future world where things will be radically different because of AI."

• The speaker foresees a transformative future shaped by AI, urging responsible stewardship.