AMA | Feb 2026

What It Covers

Sean Carroll addresses the February 2026 political crisis in the United States, including ICE killings in Minneapolis and authoritarian trends, while maintaining optimism about democracy's eventual triumph. The AMA covers dark energy measurements from DESI, black hole information puzzles, consciousness and computational functionalism debates following Ned Block's episode, neutrino mass eigenstates, and practical advice for career changers pursuing theoretical physics.

Key Questions Answered

• •Dark Energy Evolution: Recent DESI and DES observational results suggest the dark energy density may change slightly over time rather than remaining constant as Lambda CDM predicts, though data remains inconclusive. This does not imply the universe stops accelerating or will Big Crunch - even zero dark energy would only slow expansion asymptotically. Negative energy density would be required for collapse, but no theoretical motivation exists beyond checking possibilities. Most credence should remain on cosmological constant models.
• •Black Hole Information Mechanism: Current research establishes information conservation through AdS/CFT correspondence, but the specific mechanism for information extraction from real-world black holes remains unclear. Arguments from the 1980s-90s about extraction difficulties across different spacetime slicings have not been fully resolved. A paradigm exists showing information can escape, but detailed mechanisms for non-AdS black holes in our universe lack complete understanding. The field needs concrete explanations beyond theoretical frameworks.
• •Computational Functionalism Rejection: Consciousness likely requires more than input-output mapping - the physical processes matter, not just computational results. Unlike algorithms where method is irrelevant to outcomes, brain metabolism, organic chemistry, and free energy conversion may be essential. Current LLMs lack consciousness partly because they don't experience time passage or boredom. Future AI consciousness would require rich inner phenomenology and subconsciousness, not just sophisticated responses. The algorithm itself matters, not merely the answer.
• •Democratic Resilience Strategy: Despite authoritarian trends, most Americans oppose current policies, including many Republicans who regret their support. The fundamental work involves voting, organizing, and spreading democratic messages broadly - no magic bullets exist. Like cleaning a room, fighting authoritarianism never ends; even in a hundred years, anti-democratic forces will persist. The fight currently goes badly, but even-money odds favor eventual democratic victory through sustained effort and system engagement.
• •Neutrino Mass Eigenstates: Electron, muon, and tau neutrinos are interaction eigenstates that mix the three mass eigenstates (light, middleweight, heavy). Charged leptons don't get described as mixtures because neutrinos are much lighter - charged leptons decay into neutrinos, not vice versa. An alternative valid description would label neutrinos by mass and acknowledge muon decay emits a superposition of mass eigenstates. Charged leptons remain in mass eigenstates because we never create them from neutrino flavor eigenstates.
• •Physics Career Transition Reality: Learning physics to develop a pre-existing theory is misguided - successful modern theories are deeply mathematical from inception, not pictures needing mathematical backing. Prospective physicists must write equations to test whether ideas produce plausible results. Self-study through textbooks, online courses, MOOCs, and AI tutoring provides viable paths. Test readiness by attempting appropriate-level textbooks - if progress seems possible with effort, pursue it; if incomprehensible, reconsider. Allow five years for serious evaluation.
• •Simulation Hypothesis Critique: The standard simulation argument assumes random selection among all observers across simulation levels, which lacks justification. Observers should reason from their actual universe's properties, not hypothetical observer distributions. The universe shows no aspects suggesting simulation design. If tiny Planck-scale black hole remnants stored information, the enormous number of distinguishable remnant types would create detectable effects in particle physics experiments through virtual loop diagrams - effects not observed, arguing against remnant models.