1. Any low hanging fruits that could easily be solved by an LLM easily probably would have been solved by someone already using standard methods.
2. Humans and LLMs have to spend some particular amount of energy to solve problems. Now, there are efficiencies that can lower/raise that amount of energy but at the end of the day TANSTAAFL. Humans spend this in a lifetime of learning and eating, and LLMs spend this in GPU time and power. Even when AI gets to human level it's never going to abstract this cost away, energy still needs spent to learn.
1. Any low hanging fruits that could easily be solved by an LLM easily probably would have been solved by someone already using standard methods.
2. Humans and LLMs have to spend some particular amount of energy to solve problems. Now, there are efficiencies that can lower/raise that amount of energy but at the end of the day TANSTAAFL. Humans spend this in a lifetime of learning and eating, and LLMs spend this in GPU time and power. Even when AI gets to human level it's never going to abstract this cost away, energy still needs spent to learn.