Have you ever received an AI response copied and pasted into a group or a review request? Developer gruhn.me reports that this is increasingly common, and asks: stop being a ‘meat middleman’.
According to gruhn.me, passing on the output of a model like Claude adds no value. ‘I can talk to Claude myself,’ he argues. ‘It’s faster and I control the context.’
The problem is that AI output requires extra effort to read. It’s verbose, often contains plausible absurdities, and is increasingly full of jargon. The author cites a phrase he received from Claude: ‘NATS control plane events: stream leader re-election / R3 quorum reformation during pod churn.’
‘I had to look up almost every word to understand this,’ he writes.
The guidance is clear: use AI to generate ideas, but don’t just pass on the result. Read, understand, validate, and then write a response in your own words. That effort is the value you can add.
In the case of code review, the author observes that it’s now possible to submit code with almost zero effort: copy the ticket description into Claude Code, don’t look at the code or read what Claude wrote. If there’s feedback, copy and paste that too. Iterate if necessary.
It works. But who did the implementation? The reviewers, using Claude Code, and you as a meat middleman.
The message is a warning for developers and teams: automation can speed up work, but responsibility and understanding remain human. Don’t delegate your ability to think to a model.
In a world where AI is everywhere, the ability to interpret, criticize, and communicate clearly becomes even more valuable. Being a mere transmitter of machine-generated text is a disservice to yourself and others.
So, next time you’re tempted to copy and paste an AI response, remember: read, understand, validate, and write in your own words. That’s what adds value.