LlamaIndex Founder: Humans May Stop Reviewing AI Code in 1-2 Years

dotey · x · 2026-07-30

LlamaIndex founder Jerry Liu discussed agent loop engineering with a group of founders. The conversation highlighted that while most teams are not yet deeply using fully automated loops in Codex or Claude Code, there is a strong consensus that humans will no longer review code in 1-2 years. The ultimate question evolves into whether humans will need to review any output at all. Additionally, developers still face a skill issue in maximizing AI output and reducing slop.

Original post →

More from AGI Musings

AGI Musings channel →