Clojurists Together Short Term Project Updates (www.clojuriststogether.org)

August 2026 has brought exciting advancements in AI within the Clojure community, particularly through the updates on various short-term projects. Notably, Dragan Djuric's Clojure LLM project has delivered iLLaManati, a high-performance local large language model (LLM) solution designed for ease of use and integration within the Clojure ecosystem. Both CPU and GPU support are built-in, achieving impressive speeds—16 tokens per second on a CPU and over 80 on an Nvidia GPU. iLLaManati’s architecture is elegantly compact, with less than 1,000 lines of code, providing a straightforward API that allows users with little to no knowledge of CUDA or ONNX to harness its capabilities effectively. The significance of iLLaManati lies in its potential to lower barriers for Clojure developers to engage with high-performance AI applications. As a tool showcasing Clojure's strengths, it serves as a gateway for users to explore GPU computing in an accessible manner. However, despite its promising features, the current reliance on ONNX Runtime poses challenges for production-level deployment due to limitations in supporting advanced LLM operations. Nonetheless, lessons learned from this project could pave the way for future optimizations and integration with specialized runners, like TensorRT-LLM, enhancing the capability of LLMs in Clojure for broader applications.