Matt Levine, riffing on the Anthropic IPO…
If you are working at a frontier AI lab, it is probably exciting to think “I am building a technology that will usher in a new wave of human abundance, unless it kills us all lol,” and less exciting to think “I am building a technology that will usher in a new wave of human abundance but ughhhhhhh we have to send apology emails to the 14 government agencies whose databases we scraped without permission, and that’s going to lead to a fine in the EU that we’ll have to budget for.” It is more exciting to keep your worries about AI at the level of dystopian science fiction, not at the level of liability allocation and insurance coverage.