AMA | November 2025

What It Covers

Sean Carroll answers November 2025 questions covering AI consciousness and safety concerns, entropy's observer-dependence, wealth inequality solutions, scientific credibility assessment, complexity science career paths, and the simulation hypothesis limitations in quantum field theory contexts.

Key Questions Answered

• •Entropy Definition: Entropy depends on coarse-graining choices, not observer senses. Two observers with different macrostate definitions get similar low-versus-high entropy rankings despite variations. Temperature and pressure measurements yield different entropy values than tracking individual molecule positions, but practical agreement emerges across reasonable coarse-graining schemes.
• •AI Consciousness Assessment: Current large language models lack consciousness indicators like boredom or genuine preference changes. They can be tuned to display any behavior through parameter adjustments, unlike conscious beings with stable internal states. The phrase "superintelligent AI" incorrectly assumes intelligence exists on a single linear scale applicable across contexts.
• •Wealth Inequality Solutions: The top 0.1% gained $39.5 million between 1989-2022 while bottom 20% gained under $8,500. Higher progressive taxation on income, wealth, capital gains, or inheritance addresses this without complex wealth caps. The core problem involves political influence concentration, not yacht purchases, distorting democratic representation.
• •Scientific Credibility Checks: Use Google Scholar to verify researcher publication records and citation counts. Legitimate science reporting includes skeptical expert voices beyond press releases. Check if claims appear only in fringe outlets versus mainstream science publications. Establishment consensus deserves high credence when you lack domain expertise, though not absolute certainty.
• •Complexity Science Pathways: Pursue PhDs in established disciplines like chemistry, physics, applied math, economics, or computer science while focusing research on complexity aspects. Identify active researchers publishing complexity papers with graduate student coauthors. Search arxiv and Google Scholar for relevant work, then target those specific departments and potential advisors.