Hobby programming communities are increasingly exhibiting hostility towards the use of Large Language Models (LLMs). This sentiment is observed across diverse niches such as chess engine development, OSDev, Lang Dev, and the demoscene.

The core of this resistance stems from the belief that the knowledge within these communities is hard-won, and LLM usage is seen as circumventing the essential learning process. For many, the journey of mastering a difficult field is the primary product, with a functional outcome being secondary.

Early engagement with LLMs in these circles has often been soured by a lack of deep understanding from LLM practitioners and a vocal segment within the communities who perceive LLMs as a form of cheating. Respect in these traditional circles is built over years through shared code, curiosity, and domain knowledge, with an emphasis on understanding the 'why' and 'how' rather than just the 'what' works.

Ultimately, LLMs are viewed as potential force multipliers for experts, not as substitutes. In communities where the learning process is paramount, using an LLM to generate finished work is seen as robbing individuals of the craft itself.