Agreed, and on this forum we tend to focus on the tech/coding aspects. But as a knowledge worker in a different domain, I can also tell you that the same issue

Agreed, and on this forum we tend to focus on the tech/coding aspects. But as a knowledge worker in a different domain, I can also tell you that the same issue is happening for other knowledge areas that are not as auditable without expertise. While we do see this problem when relying on junior knowledge workers, there seems to be a more implicit trust of LLM outputs vs. junior knowledge workers. Also: senior knowledge workers are also subject to errors, but knowledge work isn't always deterministic.