This is a highly technical paper but I suspect it will be one of the more important studies on how AI is going to impact employment. The central claim is that AI’s impact depends on what types of tasks are automated (expert vs non-expert) and that this will then determine where that job plots on the graph of employment vs wage growth. Intriguingly, the suggestion is that there is an inverse relationship between employment share and wage growth. I’m going to explain this in more depth in Open Kitchen tomorrow so follow along
there for a longer form breakdown but in short: if AI automates non-expert tasks, the number of jobs shrinks as there are less qualified workers to do them, which then forces up wages for those experts who remain; if AI automates expert tasks, the number of jobs expands (greater talent pool) but the wages decrease with increase labour supply. These are critically important insights which we need to internalise as we go ahead with automating workflow in TA.
Must read, or get your LLM of choice to summarise it for you, or put into
NotebookLLM and listen to it.