“Don’t be a meat proxy”: Why Relaying AI Answers Verbatim Adds No Value

Software engineer and blogger Niklas Gruhn published a post titled “Don’t be a meat proxy” on 3 August 2026, criticising the growing habit of copying an AI model’s answer verbatim and passing it on to colleagues in chats, code reviews or group messages.

What a “meat proxy” is

Gruhn describes a familiar scene: you ask a question in Slack, leave feedback under a merge or pull request, or argue with friends in a WhatsApp group — and the reply comes back as “Claude said:” followed by the model’s giant response, word for word. He admits he has done it too, but says that after being on the receiving end too many times, he no longer sees the value: anyone can talk to Claude themselves — faster, and with control over the context. A “meat proxy” in between is simply not needed.

Reading AI output is extra work

According to the post, reading model-generated text is an effort in itself: it is verbose, frequently contains all too plausible nonsense, and is increasingly jargon dense. He quotes a sentence about infrastructure he recently received — “NATS control-plane events: stream leader election / R3 quorum re-form during pod churn” — and admits he had to look up almost every word to make sense of it.

What to do instead

The recommendation: prompt AI by all means, but do not just relay the output. Read it, understand it, validate it, and then write a response in your own words — that rewrite is a decent certificate that you have done the prior steps, and the effort is the value you add. He applies this to code review in particular: shipping code is now close to zero effort — paste the ticket description into Claude Code, don’t look at the code or read what the model wrote, paste reviewers’ feedback in as well, and iterate. It works, he writes, but then asks who actually did the implementation: the reviewers did, using Claude Code, and you as a meat proxy.