"The LLM critics are right. I use LLMs anyway" - the voice that reassembles

A post from theocharis.dev makes it to Hacker News: "LLM critics are right. I use them anyway." The assumed compromise position becomes an editorial style in itself.

The fact

Published on July 16, 2026 on theocharis.dev and relayed on the front page of Hacker News, the post "The LLM Critics Are Right. I Use LLMs Anyway" does not contest the harshest criticisms made of the models (hallucinations, lack of understanding, dependence) - it concedes them. Then explains why its author uses them daily nonetheless. It's the third voice in this line to pass in one week, after geohot's post ("I love LLMs, I hate hype", July 12) and a post from raymyers.org cutting through Anthropic's communication (July 13).

Our reading

What these texts have in common is not content, but a stance: refusing the "fan / critic" divide and occupying the slot of the practitioner who counts the costs. Three markers: (1) none of these authors cite a benchmark, (2) all talk about their real workflow, (3) all explicitly distinguish the model from the harness. It's the narrative shift we follow in the hype-fatigue-2026 thread: what makes an LLM useful is no longer a capability, it's a use. And the use is judged by what it produces, not by what it promises.

To watch

The transition of this stance from individual authors to structured teams - when companies' technical reviews adopt this tone, hype fatigue will no longer be a marketing signal but a methodological norm.