Hacker News Reader

You're confused. The compression algorithm was something different. TextSynth is an LLM inference server, similar to (but older than) llama.cpp.

Creating a llama.cpp like software is not using LLMs to develop software neither.