Murray Shanahan
Showing 1–2 of 2 transcripts.
- a16z42 min
a16z Podcast | Artificial Intelligence and the 'Space of Possible Minds'
Murray Shanahan, Azeem Azhar, Tom Standage, Sonal
Leading experts including Murray Shanahan, Tom Stanis, and Azim Azhar examine the theoretical and societal implications of AI, distinguishing between current machine learning techniques and broader concepts of intelligence and consciousness. The discussion highlights critical challenges in regulating "black box" deep learning systems, analyzing how algorithms might alter human behavior and redefining employment trends toward service sectors resistant to automation. Ultimately, the panel concludes that future AI trajectories involve a complex "tree of possibilities" where outcomes depend on variables of embodiment and goals that may remain entirely alien to human motives.
- Y Combinator53 min
Ex Machina's Scientific Advisor - Murray Shanahan
Academician Roger Penrose details the historical persistence of the frame problem in artificial intelligence, tracing his own career shift from symbolic logic to computational neuroscience and back to a proposed synthesis of symbolic concepts within modern deep learning. Drawing parallels with the evolution of chess and Go under AlphaGo, Penrose critiques current AI's data inefficiency and lack of semantic understanding while offering insights from his consultation on the film *Ex Machina* regarding the relationship between physical embodiment and consciousness. He concludes that while hardware capable of matching the human brain is imminent, achieving true Artificial General Intelligence still requires fundamental conceptual breakthroughs that transcend specialized algorithms or Asimov-style ethical constraints.