I recently read Nabeel Qureshi's principles. The 59th (!) stood out to me:
There’s a lot of alpha in being willing to do “menial” work (take notes, send out agendas, order pizza, manually inspect raw data, whatever). Beware over-delegation and being too far from the details.
This was written in 2023, but it feels even more important today, as we are able to delegate more and more of the details to LLMs. This can be good for efficiency, but you risk missing the alpha that is hidden in the menial work.
As with all outsourcing decisions, it comes down to judgement. If you are building products, there is danger in doing anything that puts an LLM between you and your users.
An example is customer feedback - while it is tempting to just ask Claude to summarise, there are often weak signals which you can only pick up on by reading it through yourself.
AI is changing the way we work, but the devil is still in the details.