Debate: The Biggest Moat for Compute Is How Fast It Self-Cannibalizes
HanchungLee · x · 2026-08-26
Hanchung Lee and @codestar debated compute moats and commoditization on X. Lee's earlier claim: the biggest moat for compute is how fast it can self-cannibalize — Intel dominated via this for decades, as did AMD, Nvidia and other public semis, and that's the future for AI labs.
@codestar pushed back: a year or two ago everyone avoided AI areas that would become "commodities", but commodity isn't inherently bad — oil is a commodity too. Lee countered: don't conflate self-cannibalization of compute with commodity, and "commoditization" ain't "commodity".
Related event: Investors Debate AI Commoditization and Compute Moats(2 posts)→
More from AGI Musings
- On the Social Difficulty of Rejecting Authority: EA Community Consensus Debate — teortaxesTex · 2026-08-26
- MIT scholar rebuts: motivation to learn outside real work is scarcer than AI optimists think — mattbeane · 2026-08-26
- The best AI products don't show off intelligence — they quietly erase life's friction — yunta_tsai · 2026-08-26
- Compute is expensive, but wrong paradigm commitment costs more — cjmaddison · 2026-08-26
- Anecdote: Non-technical salesperson uses AI heavily — yacineMTB · 2026-08-26
- Concern: Normal people leveraging coding agents for regular work — yacineMTB · 2026-08-26