The LLM critics are right, and the author keeps using LLMs

Sources: theocharis.dev, HN discussion
Summary: A 2026-07-15 essay (Hacker News front page 2026-07-17, 206 points) accepts that LLM criticisms about environmental cost, copyright, open-source trust, and geopolitics are valid, then argues the author still uses LLMs heavily because they amplify existing thinking, reporting $9,838 of token spend in June 2026. The thread debates whether small reviewable diffs with tests are the real trust boundary rather than whether code was AI-assisted.
Comments: HN commenters argue the trust problem is diff size and reproducibility, not authorship: a small change with clear tests is reviewable while a large opaque diff is not, regardless of who wrote it.
Why it matters: It captures the current practitioner split between accepting AI-tooling criticism and adopting it anyway, and the review-cost argument that follows.