They get impressive performance, but it's not really the first "LLM-native processor". They built the thing a while ago before LLMs were that hot, and rebranded

They get impressive performance, but it's not really the first "LLM-native processor". They built the thing a while ago before LLMs were that hot, and rebranded it as an LPU to pivot into Generative AI: https://www.eetimes.com/groq-demos-fast-llms-on-4-year-old-silicon/ https://www.eetimes.com/groq-demos-fast-llms-on-4-year-old-s...