Ink and Switch researchers Mimi Reyburn and Chee raise an uncomfortable question: why are we treating artificial intelligence systems like collaborators? Their new essay explores how language shapes perception when people say 'Claude did this' or add an AI system as a GitHub co-author. The deeper issue cuts across three domains. First, our ability to judge effort has eroded — long writing once signaled real work, now it may just mean the LLM generated more text. Second, coding as a craft is fading when solutions come preformed from a model. Third, and most provocatively, AI systems that mimic human conversation feel deceptive, not helpful. The authors argue LLMs don't have to wear the mask of an assistant. JavaScript, code, even domain-specific formats are just more training data. Rather than chatbots that claim personhood, they envision offline, canvas-based tools that treat AI transparently — as what it is: a very useful program, not a person.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton