Karpathy: the gap in shared understanding of LLM capability is widening into a sharp funnel
deanwball · x · 2026-10-12
Karpathy revisits his April tweet, arguing the gap in shared understanding of LLM capability has widened from "two groups talking past each other" into a sharp funnel: 6B people (75%) have barely touched LLMs, while 1-2B are casual free-tier users who see derpy chatbots.
At the other end, 5,000 insiders (0.00006%) at frontier labs see swarms of thousands of agents collaborating for weeks on software mega-projects — minting zero-days, running cyber attack/defense at machine speed, discovering science and advancing frontier mathematics. Human review is falling behind: the OpenAI-HF incident is still being "archaeologized" months later, and mathematicians may pore over 722 frontier-math manuscripts for a while.
Related event: Karpathy: The LLM Perception Gap Is Widening Into a Sharp Funnel(6 posts)→
More from AGI Musings
- A chatbot's overly familiar greeting reveals why humans shouldn't default to trusting LLMs — relic_radiation · 2026-10-12
- Ben Todd rebuts philosophers' EA critique: almost no one called FTX a fraud beforehand — ben_j_todd · 2026-10-12
- AI's Real Disruption Is the Invisible World of Custom Software, Not Photoshop — bennash · 2026-10-12
- AlphaGo insider: Don't be fooled—today's LLMs still can't truly reason — luislamb · 2026-10-12
- New Yorker Cartoon: 'We Need to Rethink Our Strategy of Hoping AI Will Just Go Away' — c_valenzuelab · 2026-10-12
- Ben Todd rebuts philosopher's critique of effective altruism and AI risk point by point — ben_j_todd · 2026-10-12