Investor Critique: Etching Transformers Into Silicon Is Inherently Limiting

JosephJacks_ · x · 2026-07-24

Investor Joseph Jacks sharply criticized the trend of etching fixed learning architectures like Transformers into silicon chips.

While acknowledging it might produce a nice exit for investors, he argues it is inherently very limiting. Unlike Bitcoin mining, where the arithmetic is fixed, the learning arithmetic in AI is continuously evolving. Hardcoding current architectures into hardware ignores the rapid evolution of AI algorithms.

Original post →

More from AGI Musings

AGI Musings channel →