I’ve come to think of AI as being a pretty-good-intern on my team. I don’t trust it with anything important and it definitely shouldn’t do independent work. I find AI helpful for organizing my thoughts in a way I might not, but key point is these are my thoughts, my work, my analysis.
the experience has led me to believe people who are using consumer-facing LLMs for research work are either a) wildly inefficient workers b) researchers producing fraudulent work. and their work is not actually being reviewed for accuracy by their supervisors/peers. that's a lot of fraud.
Dec 21, 2025 19:13
