To my understanding this is a small low-energy chip designed for applications like basic object detection in a doorbell camera, or speech recognition for a home

To my understanding this is a small low-energy chip designed for applications like basic object detection in a doorbell camera, or speech recognition for a home assistant. I think it's a bit misleading to say "faster [...] than Nvidia's best AI GPU" and "outperofrms[sic] all major architectures" - it specifically means latency and energy efficiency. Its throughput/memory are so tiny that it's not at all a replacement for Nvidia's best AI GPUs (and not intended to be): > NorthPole's core array includes 192 MB of flexible memory (768KB of unified memory per core). Assigning 2/3rd of this memory to parameters, such as weights and biases, provides 128MB for network storage. https://www.science.org/action/downloadSupplement?doi=10.1126%2Fscience.adh1174&file=science.adh1174_sm.pdf https://www.science.org/action/downloadSupplement?doi=10.112...