Conclusion: r/LocalLLaMA still has brilliant open-weight research, but finding it requires wading through endless benchmark drama, non-local Discussion Points and repetitive hardware flexes.