Teortaxes: once NS-solver-level training is mastered, remaining chip bottlenecks fall fast
teortaxesTex · x · 2026-09-18
Quoting willdepue's "fact check: true", teortaxesTex argues: if the claim holds — that model training at the level of an NS solver is already mastered and many 100K+ GPU clusters are imminent — then once the model exists, the remaining semiconductor bottlenecks will be quickly solved, pushing past industry-max parity.
A big-picture take on frontier training capability racing ahead of compute supply chains; context is limited and the view is speculative.
More from AGI Musings
- Top AI Researchers Predict 10x Productivity Gains in 2 Years, Economists Still Forecast Only 4-5% GDP Growth — soumitrashukla9 · 2026-09-18
- Merchants of Cope: AI takeoff is mass-producing cognitive dissonance, and soothing it pays — jankulveit · 2026-09-18
- 'AI killed SaaS' but everyone still buys your CRM: the irony goes viral — jacob_posel · 2026-09-18
- John Arnold op-ed: job training is the popular AI-disruption fix that barely works — robseamans · 2026-09-18
- John Arnold: West Coast sees an AI economic transformation, East Coast still forecasting 2.1% GDP growth — robseamans · 2026-09-18
- TMLR desk-rejection rate jumps from 6% to 53% as AI-assisted submissions flood in — vykthur · 2026-09-18