How OpenAI Used Its Own LLMs to Design Its Jalapeño Chip
OpenAI used its own large language models (LLMs) to design the Jalapeño chip, a custom ASIC for its AI research. The LLMs generated a high-level design of the chip, which was then refined by human engineers. The Jalapeño chip is designed for AI research and has a 1.3-billion-transistor, 7-nanometer process node, and will be manufactured by Taiwan Semiconductor Manufacturing Company (TSMC). The chip's performance is predicted to be 15 times faster than OpenAI's current TPU-based systems. The company plans to use the Jalapeño chip for its future AI research, including the development of its GPT-4 model.
Read the full article at spectrum.ieee.org →