we talk a lot about how the only vaguely applicable uses of GenAI are exclusively solving problems that were produced by the very pressures and ideologies that made GenAI.
automating useless busywork is only impressive if you don't realise it's useless busywork to begin with. drafting an email with GenAI for it to only be turned into bullet points by someone else's GenAI is only impressive to boneheads who don't understand the point of communication in the first place. one squogillion lines of code that could be replaced by smart library choices and some deterministic boilerplate are only impressive to someone who has had their understanding of software engineering replaced by a block of aged cheese in a botched medical intervention.
this is the core reason why GenAI does not translate to productivity, it just translates to "more". but it's also why it is such a difficult argument for those of us who have not self-lobo'ed to make, because we have to prove, to people whose entire identity is pointless busywork, that doing more of their job won't help anything because the fundamental nature of their job doesn't help anything.