Interview, Podcast
Lee Cronin: Controversial Nature Paper on Evolution of Life and Universe | Lex Fridman Podcast #404
- Humans face a significant risk of permanent isolation from naturally occurring alien intelligence because their causal cones are unlikely to overlap with those of alien planets, potentially due to an inability to construct intersecting architectures in time or space-time.
- To mitigate this risk, plans exist to rapidly create lab-based alien life using alternative genetic or protein alphabets and controlled constraints, aiming to generate signals with high copy numbers that could be detected on other planets.
- Future detection strategies involve a "life meter" device capable of measuring assembly indices via mass spectrometry to identify molecules with an assembly index greater than 15, a threshold distinguishing biological samples from abiotic ones, or finding molecules with molecular weights exceeding 350 or more than 15 fragments.
- Assembly theory is predicted to serve as a universal metric for complexity, applicable to language, mathematical theorems, economic processes, microprocessor architecture, and culture, allowing for the identification of selection in non-Darwinian contexts such as hydrothermal vents.
- The framework suggests that while nature operates through shortest paths and compromises in joint assembly spaces, human abstraction may enable a transition beyond biological selection, potentially allowing causal cones to overlap with alien intelligence as humans build new technologies.
- Experimental goals include mapping the tree of life via mass spectroscopy, dating fossils by analyzing molecular decomposition and enantiomer deviation, and training models on electron densities to generate novel molecules for drug discovery.
- Current AI systems are characterized as inductivist tools that mine the past to produce shallow novelty, whereas genuine novelty is defined as configurations not predicted by past conditions, arising from the interaction of objects with different histories.
- The outlook posits that the future is undetermined and larger than the past, with the universe generating novelty through selection rather than adhering to deterministic laws, thereby refuting the Boltzmann brain hypothesis and the concept of a fictitious super-intelligence turning against humanity.
- Specific constraints for drug discovery and chemical reaction analysis will shift toward molecule evolution and quantum mechanical constraints rather than protein evolution, with the intention to create a chemical brain to understand evolved intelligence mechanisms.
- Risks regarding artificial intelligence are framed around the proliferation of fake data and users rather than existential threats from super-intelligent agents, with a stance against regulation that could halt progress or the notion that current hardware lacks the flexibility for domain integration.
- Material thresholds for life detection include the presence of molecules with a hundred different parts and a million identical copies, which are suggested to be impossible to produce via random processes in an infinite set of universes.