Links for Tuesday

Four links for a Tuesday.

• Thinking about AI in the context of monsters? — The monster is the figure through which the West has always thought about what is at once familiar and alien. It is not a literary flourish: it helps examine why these systems unsettle us most when they resemble us, rather than when they fail.

• Dawkins wants to know whether ChatGPT is conscious — The value of the conversation lies not in the model's answers but in the interviewer's discomfort. Dawkins does what almost nobody does: he asks in earnest, and stays to listen to what does not convince him.

• Is entrepreneurialism a new form of pathology? — A history of how a country turned economic insecurity into moral character. Reading it helps explain where some of our industry's most celebrated reflexes come from.

• Science as a publication game? — When the metric becomes the goal, the system learns to produce metric. It is a problem of incentives and therefore a problem of design — the same thing that happens to us when we mistake activity for progress.