Isaac Su argued on August 1, 2026, that workers should not credit LLM tools when answering questions, posting pitch documents, or submitting pull requests, because credit and accountability should stay with the human who sends the work. The post cited examples such as someone saying an LLM tool found 17 support requests for Feature X since release, or saying an LLM tool wrote unit tests for a pull request.

Su described several possible reasons people disclose LLM use: they may think credit is owed to the tool, feel that LLM assistance makes the work impressive by association, send output without reviewing it, or work in an environment where LLM use has been mandated by managers. He singled out the first reason for challenge and treated the others as signs of weaker review or workplace pressure.

Su’s central claim was that a tool can take only as much credit as it can be held accountable for. Because an LLM cannot be held responsible for mistakes, crediting it can dilute the worker’s own accountability for the result. He compared routine LLM disclosure to a writer naming a spellchecker or a pilot announcing avionics software, asking why recipients needed that information.

Su concluded that people should take full credit for profound or remarkable work and full responsibility when work fails or becomes “slop.” The linked Hacker News discussion had a score of 34 and 40 comments, while the article listed an HN discussion link and a “20” marker at the end.

Isaac Su argued on August 1, 2026, that workers should not credit LLM tools when answering questions, posting pitch documents, or submitting pull requests, because credit and accountability should stay with the human who sends the work. The post cited examples such as someone saying an LLM tool found 17 support requests for Feature X since release, or saying an LLM tool wrote unit tests for a pull request.

Su described several possible reasons people disclose LLM use: they may think credit is owed to the tool, feel that LLM assistance makes the work impressive by association, send output without reviewing it, or work in an environment where LLM use has been mandated by managers. He singled out the first reason for challenge and treated the others as signs of weaker review or workplace pressure.

Su’s central claim was that a tool can take only as much credit as it can be held accountable for. Because an LLM cannot be held responsible for mistakes, crediting it can dilute the worker’s own accountability for the result. He compared routine LLM disclosure to a writer naming a spellchecker or a pilot announcing avionics software, asking why recipients needed that information.

Su concluded that people should take full credit for profound or remarkable work and full responsibility when work fails or becomes “slop.” The linked Hacker News discussion had a score of 34 and 40 comments, while the article listed an HN discussion link and a “20” marker at the end.