Comment by qwertyi0k

Most serious researchers want to work on interesting problems like reinforcement learning or robotics or RNN or dozen other avant-garde subjects. None want to work on "boring" LLM technology, requiring significant engineering effort and huge dataset wrangling effort.