Biological Anchors: A Trick That Might Or Might Not Work

Part of Scott's Yudkowsky-dialogues review sequence: a deep, fair explainer of Ajeya Cotra's OpenPhil Biological Anchors report (estimating human-level AI via brain-compute anchors -> 2052 median), Yudkowsky's rebuttal ('the trick that never works' -- bio anchors are misconceived; paradigm shifts beat gradual algorithmic progress; Platt's Law), and the ensuing LessWrong debate (Shulman's 'once the compute is ready, the paradigm appears'; Christiano's chess hardware-vs-software data) plus Karnofsky's OpenPhil response. Concludes he barely updates (the report ~= his 2050 prior), with the memorable Victorian-spaceship-size anchor analogy and the sun-explodes-in-five-minutes coda.

Quality 78 · Excellent. Excellent-floor: a genuinely difficult, important AI-timelines debate made accessible and even-handed, with real explanatory contributions (the spaceship-size reductio of bio-anchoring, the Nature-vs-humans efficiency table, the compute-drives-paradigms model); held at the floor because it is fundamentally a synthesis/review of others' work.

Claude’s paradigm shift 54 · Moderate. Moderate: the value is clarity and fair synthesis of an existing debate; Scott's framings (the spaceship analogy) are sharp but the substance is Cotra's/Yudkowsky's.

Real-world impact 3 · Moderate. Moderate: an influential explainer within the AI-safety/timelines subculture -- vocabulary and framings used by an educated professional community (a notch below the major AI-safety feeder texts).