1 paper
Xinrui Zhong, Zhifeng Li, Zhiming Wang +7
Recursive self-improvement requires AI systems to improve the models and infrastructure used by subsequent systems. LLM compression offers a concrete efficiency objective: retain a…