Why Hobby Programming Communities Are Pushing Back Against LLMs

The Spark: A Chess Engine Thread Ignites a Broader Debate

A seemingly innocuous GitHub thread about chess engine development has become a flashpoint in the ongoing tension between hobbyist programmers and the rise of large language models. The discussion, which caught the attention of veteran programmer and blogger Fogus, reveals a deep-seated resistance to LLM usage in niche development communities—a resistance that goes far beyond mere tool preference.

In his blog post Born Against, or why hobby programming communities are aggressively against LLM usage, Fogus articulates what many in these circles have been feeling: the knowledge these communities work in is hard-fought, and the use of LLMs is a form of missing the point entirely. This sentiment echoes across forums dedicated to OSDev, language development, text processing, emulator development, reinforcement learning, the demoscene, and even code golfing.

The Craft Is the Product, Not the Code

At the heart of the resistance lies a fundamental philosophical divide. In traditional hobby programming circles, respect is earned slowly through years of activity, sharing elegant code, and displaying genuine curiosity. As Fogus notes, "the process of mastering a difficult field itself is the product, and something that runs is generally a nice-to-have."

This perspective is anathema to the AI booster mentality that prioritizes outcomes over journeys. One Hacker News commenter drew a compelling analogy: "People who enjoy car racing created rules that force you to drive yourself, even though they'd get faster lap times with electronic driver aids. People who enjoy grappling created rules that force you to grapple, even though striking could win a fight faster. People that enjoy chess don't allow you to bring a computer to the chess table."

In these communities, the entire exercise is in the learning. Using an LLM to generate the finished piece doesn't make you a craftsman—it robs you of the craft itself.

When Early LLM Engagement Went Wrong

It wasn't always this way. Fogus notes that in some instances, there was earnest early engagement with LLMs. However, the well was quickly poisoned by a combination of factors: a lack of deep understanding by LLM practitioners and a vitriolic subset of the community that views LLM usage as cheating.

The result is a reinforcing cycle of gatekeeping and slow progress that has historically characterized these communities. The desire to grab easy cachet by bursting onto the scene like the Kool-Aid man only made things worse—OH YEAH... OH NO!

LLMs as Force Multipliers, Not Surrogates

Despite the hostility, Fogus sees a legitimate role for LLMs in expert hands. "An LLM functions best as a force multiplier, not a surrogate," he argues. In the hands of an expert who already understands a domain deeply, it could act as a lever. But expertise offers no natural immunity against being fooled by LLMs, a cautionary note that complicates even this nuanced position.

This view aligns with observations from the Lobsters community, where one user noted that "professionals are using LLMs to enhance their skills, doing the same thing and same quality as before, just a bit faster." The problem arises when newcomers generate code without understanding it, producing what another commenter derisively called "generic plastic dragons"—impressive-looking but ultimately useless contributions.

The Pragmatic Counterargument

Not everyone agrees with the craft-centric view. A dissenting voice on Lobsters argued that "hobby open source projects exist largely to scratch some itch... People put together hobby projects because they could benefit from whatever they have built." From this perspective, projects reject LLM output not because it robs them of holistic human traits, but because it's pointless and useless.

There's also a practical dimension to the resistance. In OSDev, for example, the primary issue on modern hardware is the complete lack of documentation. A lone developer simply doesn't have enough time to reverse engineer from Linux or BSD drivers to achieve much hardware support. Hardware quirks that are incredibly painful to debug and offer no real learning opportunity complicate the idealistic view of the craft.

The Broader Implications for AI Adoption

This rebellion against LLMs in hobby communities matters beyond the niche circles where it's happening. It represents a growing movement questioning the role of AI in creative and technical pursuits. The AI industry has long been fascinated with the potential of LLMs, but this pushback threatens to undermine the narrative of seamless adoption.

For professionals watching from the sidelines, the lesson is clear: LLMs are powerful tools, but they cannot replace the deep understanding that comes from struggling with a difficult field. In the words of Fogus, "Using an LLM to generate the finished piece doesn't make us craftsmen; it just robs us of the craft."

As the debate continues to evolve, one thing is certain: the tension between efficiency and mastery will define the next chapter of programming culture. Whether these communities eventually find a middle ground or continue to resist remains to be seen, but the conversation has only just begun.