Hacker News Reader

It is labour intensive to optimise a prompt though. Particularly for smaller models.

Setting parameters for any ml model is easy, but we'd call it labour intensive if we expected people to do it manually despite having evals. Instead we have ways of searching for and optimising settings. The methods for that are obvious for small cardinality discrete values or continuous variables. Less so for arbitrary text.

It is not labor-intensive if you want a prompt to work in 80%-90% of cases and humans are good at that. But it is labor-intensive if you want to make it to work at 99%. Then you need to go through many cases and "optimize" the prompt, which is the advantage of optimizer.

this made my day