I love LLMs, I hate hype
What happened
George Hotz, Comma.ai founder, published 'I love LLMs, I hate hype' on July 12, 2026, arguing that AI progress derives primarily from Moore's Law and general compute — not from frontier labs that overstate contributions to justify massive fundraising.
Context and impact
The post garnered 289 points on Hacker News and represents a significant counter-signal to prevailing AI hype narratives. Hotz frames LLMs as genuinely useful tools comparable to compilers, while rejecting both doomsday scenarios and claims of revolutionary breakthroughs owned by specific companies.
Details
- Hotz: LLMs = compilers of this era — productive, not revolutionary
- AI progress = Moore's Law + compute, not lab-specific innovation
- Critique of frontier lab fundraising model
- Rejection of AI doomsday narratives and AGI hype
- HN: 289 points, July 12, 2026
Open original source
geohot.github.io