An LLM agent read a paper and designed a working inference chip in 80 hours — no humans in the loop

An LLM agent read a paper and designed a working inference chip in 80 hours — no humans in the loop | BedrockNews