OpenAI touts Jalapeño chip for faster AI responses
OpenAI says its new Jalapeño chip can complete AI tasks more efficiently and return responses faster than competing AI systems. The company made the claim in a blog post published Tuesday, presenting the chip as a way to improve both response speed and processing capacity.
During a briefing with reporters, OpenAI hardware vice president Richard Ho said Jalapeño provides what he called the “best of both worlds”: lower latency and higher throughput. Ho said AI systems generally have to trade one of those qualities against the other. OpenAI first introduced Jalapeño in June. The chip is an Application-Specific Integrated Circuit, or ASIC, developed in partnership with Broadcom. It is intended for AI inference, the stage in which a trained model runs to complete a task or deploy an agent. The company’s statements describe the chip’s expected advantages, but the article provides no specific performance figures or detailed comparison with particular competing systems.
Written by Sakhanda Wire AI from reporting by The Verge. Facts, quotes and full context are in the original article below.