The MICDE 𝗦𝘁𝗮𝘁𝗲 𝗼𝗳 𝗔𝗜 & 𝗙𝘂𝘁𝘂𝗿𝗲 𝗼𝗳 𝗜𝗻𝘀𝘁𝗶𝘁𝘂𝘁𝗶𝗼𝗻𝘀 event was incredible. It will take many hours to write a deep summary, and I might never get to it, so here are some highlights instead:
Chandra Sripada Suitably sophisticated AI systems are not "alien minds"; exhibit deep convergences with the human mind-brain because they both reach similar solution spaces due to structure of highly constrained optimization ; AI may yield "tragic knowledge"—insights that deepen understanding while destabilizing meaning, particularly regarding conceptions of free will and moral responsibility
Rada Mihalcea AI replicas created through simulations show 85% correlation with human psychological behavior; multi-agent systems can recreate and favor implicit biases; We must focus on advancing the foundations of AI to ensure systems are built to be robust, private, and aligned with human interests
Karthik Duraisamy Not "LLMs" anymore (scaffolded systems with agency); AI’s primary strength in science is finding pockets of reducibility that it can exploit patterns and iterate, even if it cannot abstract or solve irreducible problems; Many limitations can be overcome via feedback (human and otherwise)
Timothy Welch Conflict between the democratization of production capacity and a potential restriction of the drive to develop crystallized intelligence; By removing the economic facet of intellectualism, AI has the potential to return us to being "learners once more," focusing on the joy of discovery. Prior Post
Graham Hardig Dynamic machine intelligence is "out of the bag" and cannot be uncreated; Fearing AI to the point of stagnation is a national security risk; democratic nations must innovate at competitive speeds to ensure their values guide the technology’s use
Jerry Davis AI could enable the rise of zero-person unicorns—corporations controlled entirely by algorithmic entities that exist solely to maximize shareholder value without any human employees; The "Promethean spark" of being able to speak software into existence is liberating for small businesses and individuals
Scott Page High cognitive ability has always correlated with human happiness and life expectancy; a major concern is whether AI will erode the sense of agency that is central to human well-being; Moving toward a world of "incredible heterogeneity" where local knowledge, history, and diverse human perspectives matter more than identical "commodity" skills
Richard Lewis Nothing in science suggests that any human cognitive capacity will remain out of reach for AI; We are entering a new era for mechanistic theories of central cognition
Jason Owen-Smith Critical risk that AI will make individual faculty more productive while making the broader knowledge system "dumber, more homogeneous, and less novel"; Graduate/post-doc training must be refactored around judgment, discernment, and the ability to pivot, rather than narrow technical expertise that AI can now replicate
Chandra Sripada Suitably sophisticated AI systems are not "alien minds"; exhibit deep convergences with the human mind-brain because they both reach similar solution spaces due to structure of highly constrained optimization ; AI may yield "tragic knowledge"—insights that deepen understanding while destabilizing meaning, particularly regarding conceptions of free will and moral responsibility
Rada Mihalcea AI replicas created through simulations show 85% correlation with human psychological behavior; multi-agent systems can recreate and favor implicit biases; We must focus on advancing the foundations of AI to ensure systems are built to be robust, private, and aligned with human interests
Karthik Duraisamy Not "LLMs" anymore (scaffolded systems with agency); AI’s primary strength in science is finding pockets of reducibility that it can exploit patterns and iterate, even if it cannot abstract or solve irreducible problems; Many limitations can be overcome via feedback (human and otherwise)
Timothy Welch Conflict between the democratization of production capacity and a potential restriction of the drive to develop crystallized intelligence; By removing the economic facet of intellectualism, AI has the potential to return us to being "learners once more," focusing on the joy of discovery. Prior Post
Graham Hardig Dynamic machine intelligence is "out of the bag" and cannot be uncreated; Fearing AI to the point of stagnation is a national security risk; democratic nations must innovate at competitive speeds to ensure their values guide the technology’s use
Jerry Davis AI could enable the rise of zero-person unicorns—corporations controlled entirely by algorithmic entities that exist solely to maximize shareholder value without any human employees; The "Promethean spark" of being able to speak software into existence is liberating for small businesses and individuals
Scott Page High cognitive ability has always correlated with human happiness and life expectancy; a major concern is whether AI will erode the sense of agency that is central to human well-being; Moving toward a world of "incredible heterogeneity" where local knowledge, history, and diverse human perspectives matter more than identical "commodity" skills
Richard Lewis Nothing in science suggests that any human cognitive capacity will remain out of reach for AI; We are entering a new era for mechanistic theories of central cognition
Jason Owen-Smith Critical risk that AI will make individual faculty more productive while making the broader knowledge system "dumber, more homogeneous, and less novel"; Graduate/post-doc training must be refactored around judgment, discernment, and the ability to pivot, rather than narrow technical expertise that AI can now replicate