I thought they were as well, because there were nuances which made them harder for LLMs. When I tried on 1 dec, gpt4 couldn't solve day 1 part 2. And I tripped

I thought they were as well, because there were nuances which made them harder for LLMs. When I tried on 1 dec, gpt4 couldn't solve day 1 part 2. And I tripped over exactly the same thing ; when I parsed it correctly and then explained to the llm, it did solve it. But no where near as fast as me with my bag of horrible hacking aoc lib... It's interesting it turns out this year was not written with gpt in mind.