Predicting AI right-to-left: lab sociology and compute distribution beat tech reasoning
akbirthko · x · 2026-11-09
akbirthko argues that as intelligence commoditizes, reasoning about AI systems' shape must go right-to-left — from frontier lab sociology, compute distribution and physical constraints, not forward from technical roadmaps.
The quoted thread from teortaxesTex makes a historical case for skepticism about remaining hard barriers: Rosenblatt, Hinton, Sutskever, Kingma, Vaswani, DeepDream's Mordvintsev — none of it was a paradigmatic leap. The king's road was just ironing out connectionism's logistics: continuous relaxation, iterative problem-solving, applied-math tool combinations, and pruning of too-clever branches. (XKCD 'Fields Arranged by Purity' comic attached.)
More from AGI Musings
- Takeoff debate: concave curves don't prove diminishing returns block superintelligence — austinc3301 · 2026-11-09
- Exponential AI progress is exponential investment with sublinear returns — world-changing, not god — dhadfieldmenell · 2026-11-09
- Takeoff debate: concave curves are by construction, not evidence of diminishing returns — austinc3301 · 2026-11-09
- Derya Tabak: People are still massively underestimating AI — abundant superintelligence changes everything — DeryaTR_ · 2026-11-09
- Counter to Sutton's "LLMs aren't bitter lesson enough": pretraining won via Hayekian aggregation — akbirthko · 2026-11-09
- fkasummer: 2028 hinges on a new paradigm for intelligence — transformers, backprop all up for grabs — fkasummer · 2026-11-09