[AINews] LLMs can read & write neural network weights: "Weight Space Programming" cuts Autoresearch out of the RSI loop

Researchers demonstrated that large language models (LLMs) can autonomously read and modify the weights of smaller neural networks, enabling training without human intervention. This new approach, called Weight Space Programming, could revolutionize recursive self-improvement by bypassing traditional manual coding and optimization. Early reproductions confirm its potential efficiency for simple tasks like games and reinforcement learning policies.