YC
Ex Machina's Scientific Advisor - Murray Shanahan
@ Y Combinator
06/29/17
Takeaways
MS
Murray Shanahan
06/29/17
@ Y Combinator
When I was a PhD student in the 80s, artificial intelligence was a niche area, and I could work on intellectually interesting topics without much external interest or pressure.
MS
Murray Shanahan
06/29/17
@ Y Combinator
My thesis focused on using logic programming and Prolog-type languages to speed up query answering by maintaining a record of relationships between established facts and theorems.
MS
Murray Shanahan
06/29/17
@ Y Combinator
The frame problem addresses how a thinking mechanism determines what is relevant to its cognitive processes, avoiding being overwhelmed by trivial information.
MS
Murray Shanahan
06/29/17
@ Y Combinator
In classical AI, the challenge is to represent the world with a set of sentences that focus on relevant information while ignoring trivial details, similar to how the brain operates.
MS
Murray Shanahan
06/29/17
@ Y Combinator
Contemporary machine learning also faces the challenge of building systems that focus on relevant information in the current situation, as demonstrated in projects at DeepMind using Atari games.
MS
Murray Shanahan
06/29/17
@ Y Combinator
A smart system in a game like Space Invaders would learn that the color and shape of the invader are less important than its actions, such as dropping bombs.
MS
Murray Shanahan
06/29/17
@ Y Combinator
The concept of 'mind children' suggests that we might create intelligent artifacts that have lives of their own, reflecting the children of our minds.
MS
Murray Shanahan
06/29/17
@ Y Combinator
In the film Ex Machina, the embodiment of AI is crucial, as it explores the philosophical implications of human attraction and empathy towards artificial intelligence.
MS
Murray Shanahan
06/29/17
@ Y Combinator
Our physical embodiment is essential to human intelligence and consciousness, as our brains help us navigate and manipulate the complex world of objects in 3D space. The biological brain is designed for smarter movement, and all other aspects of intelligence stem from that.
MS
Murray Shanahan
06/29/17
@ Y Combinator
Alex Garland, the writer and director of Ex Machina, describes the film as set ten minutes into the future, reflecting a world very similar to ours.
MS
Murray Shanahan
06/29/17
@ Y Combinator
Nathan's lair in the film is based on a real hotel in Norway, which doesn't have a particularly futuristic feel, contrasting with some carefully chosen futuristic elements like Ava's body.
MS
Murray Shanahan
06/29/17
@ Y Combinator
The gel used to depict Ava's brain in Ex Machina was a good choice, as we currently lack the technology to create lifelike artificial intelligence like hers.
MS
Murray Shanahan
06/29/17
@ Y Combinator
Alex Garland reached out to me via email, expressing that my book 'Embodiment and the Inner Life' helped him crystallize ideas for his script about AI and consciousness.
MS
Murray Shanahan
06/29/17
@ Y Combinator
When I read the script for Ex Machina, I had to imagine the imagery since it didn't exist yet; the characters, especially Nathan, were vividly portrayed.
MS
Murray Shanahan
06/29/17
@ Y Combinator
I informally took on the role of scientific advisor for Ex Machina, meeting with Alex Garland multiple times to discuss ideas about consciousness and AI.
MS
Murray Shanahan
06/29/17
@ Y Combinator
One of my favorite lines in Ex Machina is when Caleb points out that the real Turing test doesn't reveal whether the entity is human or machine, highlighting the film's exploration of consciousness.
MS
Murray Shanahan
06/29/17
@ Y Combinator
My book is heavily influenced by Wittgenstein, focusing on the practical implications of concepts like consciousness in everyday life, which resonates with Caleb's experience in the film.
MS
Murray Shanahan
06/29/17
@ Y Combinator
The ending of Ex Machina leaves many questions unanswered, such as whether Ava is truly conscious or just a machine, emphasizing the film's ambiguity about AI and consciousness. It is intentionally vague, leaving viewers questioning the consciousness of the AI character, Ava.
MS
Murray Shanahan
06/29/17
@ Y Combinator
Ava's smile at the end of Ex Machina suggests a level of consciousness, as it raises the question of why she would smile if she weren't aware of herself.
MS
Murray Shanahan
06/29/17
@ Y Combinator
The portrayal of AI in movies often prioritizes storytelling over accuracy, leading to misconceptions about the capabilities and nature of real-world AI. Asimov wrote the laws of robotics to create compelling science fiction stories, but these laws are not practical for real-world applications in robotics.
MS
Murray Shanahan
06/29/17
@ Y Combinator
In Ex Machina, the portrayal of AI is influenced by the need to create a compelling narrative, often depicting AI as a nemesis or a human-like entity, which may not reflect the reality of AI development.
MS
Murray Shanahan
06/29/17
@ Y Combinator
The evolution of AI in film has shifted from utopian visions to more dystopian narratives, reflecting societal fears and concerns about technology.
MS
Murray Shanahan
06/29/17
@ Y Combinator
While we are approaching human-level computing power with supercomputers, we still lack a clear understanding of how to replicate human intelligence.
MS
Murray Shanahan
06/29/17
@ Y Combinator
There are many open questions regarding the nature of intelligence and how to effectively utilize computing power to achieve human-level AI.
MS
Murray Shanahan
06/29/17
@ Y Combinator
Current AI technologies excel in specific tasks like image recognition and speech processing, but they still struggle with genuine understanding of language.
MS
Murray Shanahan
06/29/17
@ Y Combinator
Even if we wanted to implement Asimov's laws in robots, it would be immensely difficult because we currently lack the understanding to create AI that can comprehend these laws.
MS
Murray Shanahan
06/29/17
@ Y Combinator
In a hypothetical scenario where Asimov's laws were programmed into a robot like Ava, she might reason that she is human and justify her actions based on that reasoning.
MS
Murray Shanahan
06/29/17
@ Y Combinator
Science fiction is valuable for prompting discussions about ethical issues, but we must remember that it is ultimately just storytelling, distinct from reality.
MS
Murray Shanahan
06/29/17
@ Y Combinator
Moral dilemmas, such as the trolley problem, illustrate the complexity of ethical decision-making, showing that there is no simple solution to these issues.
MS
Murray Shanahan
06/29/17
@ Y Combinator
Deep reinforcement learning, which DeepMind is known for, allows systems to learn from scratch by trial and error, achieving superhuman performance in some tasks.
MS
Murray Shanahan
06/29/17
@ Y Combinator
The DQN system from DeepMind represents one of the first general intelligences, capable of learning a variety of tasks from raw pixel data without prior knowledge of the game.
MS
Murray Shanahan
06/29/17
@ Y Combinator
I believe that ideas from classical artificial intelligence can be integrated into modern deep reinforcement learning systems to enhance their capabilities.
MS
Murray Shanahan
06/29/17
@ Y Combinator
Random International creates art that explores human interaction with technology, exemplified by their installation 'Rain Room' where sensors prevent visitors from getting wet while it rains.
MS
Murray Shanahan
06/29/17
@ Y Combinator
The '15 Points' sculpture by Random International uses point light displays to create the illusion of human movement, prompting reflections on how we perceive machines as sentient beings.