This has always been the case. Some of these folks are experienced competitive programmers that have a whole automated setup for downloading inputs and uploading results the moment their solution stops running. When you watch their livestreams, you realize that they spend almost no time reading the flavor text and just extract what they see as relevant and rely on pre-defined code samples from other competitions (sometimes even their own personal programming language) to solve these as quickly as possible.
So what? Let them generate their breadth-first search or knapsack problem code and such. After participating multiple times over the years, you probably have used almost every data structure and algorithm out there. I also re-use and copy-paste my previous approaches over and over again.
The day an LLM can solve Day 15 of 2018, will be the day I switch to gardening full-time.
Until then I simply see AoC as a relaxing yearly event where I can learn a new programming language.
