Скачать или смотреть Agentic AI In Chip Design

Large language models (LLMs) like ChatGPT are just the starting point for generating content with AI. The next phase will be about harnessing LLMs with agents, providing automated feedback and improvements in performance and accuracy. Mahir Arora, backend engineer at ChipAgents, talks with Semiconductor Engineering about the impact this can have on EDA and chip design, allowing smaller teams to compete with larger teams, and helping junior engineers come up to speed much more quickly than in the past. For example, an AI agent can be asked to map how data flows through a system, and that agent will set up a plan and self-correct it until all the possible options are exhausted, which in turn can provide insights to accelerate the verification process.