AI upstart Anthropic has agreed to create a $1.5 billion fund it will use to compensate authors whose works it used to train its models without seeking or securing permission. News of the settlement ...