We're Not Platonists, We've Just Learned The Bitter Lesson

Responds to a Twitter claim that intelligence-explosion arguments require Platonism. Argues: (1) concepts exist as 'bundles of useful correlations' (the Mike-Tyson-stronger-than-grandmother decomposition - 'strength' is a useful shorthand without a Platonic essence); (2) 'intelligence' is such a concept - g/IQ, SAT verbal-math correlated 0.72, holds across species via neuron count; (3) 'intelligence' has been the FRUITFUL frame for AI - the Bitter Lesson (Sutton), Jelinek's Law ('every time I fire a linguist...'), OpenAI's 'giant blob of intelligence' beating clever hand-coded paradigms; (4) this suffices for 'intelligence explosion' to be coherent (the steroids-in-a-box 'strength explosion'; reading/writing/iodine as prior intelligence explosions). Closes with Near-Mode vs Far-Mode (Beware Isolated Demands For Rigor callback).

Quality 74 · Strong. Top-of-Strong / standout explanation. A clean, oft-cited defense of intelligence-realism (the bundles-of-correlations framing + the strength analogy are excellent pedagogy) tied well to real AI history (Bitter Lesson, Jelinek). Held below Excellent because it largely re-deploys Scott's own existing toolkit (categories-were-made-for-man, Near/Far, isolated-demands-for-rigor) plus established AI lore against a Twitter objection - high clarity, modest originality.

Claude’s paradigm shift 48 · Moderate. Moderate - the components (fuzzy correlation-bundle concepts, intelligence-realism, the Bitter Lesson) all pre-exist; the synthesis defending intelligence-explosion coherence is a sharp packaging, not a new frame.

Real-world impact 2 · Minor. Within-discourse conceptual/AI essay; the bundles-of-correlations framing is influential in the intelligence-realism debate but has no material reach.