Lauren Tan and Gergely Orosz on writing your own words with AI agents

Lauren Tan says she does not let AI agents write for her, including docs, tweets and prompts, because she believes writing is thinking.

In an X post, she argues that clear writing improves how she directs agents, and that words are the interface to capable models. She says she will keep writing every word until brain-computer interfaces replace the keyboard.

Gergely Orosz says many of the most AI-enthusiastic people at AI labs write their own posts instead of letting AI do it, to keep thinking. He says many less AI-enthusiastic people let AI take over their writing, and that this shows in their thinking.

Both posts state opinions and personal practice. Neither reports a measured effect of writing by hand on output or on anyone's thinking.