That seems inaccurate. GPUs are used for stuff like real-time games, video, and ML inference: the hardware is explicitly built for heavy streaming. I would agr

That seems inaccurate. GPUs are used for stuff like real-time games, video, and ML inference: the hardware is explicitly built for heavy streaming. I would agree that productive GPU data frameworks in streaming modes are nascent, e.g., https://medium.com/rapids-ai/gpu-accelerated-stream-processing-with-rapids-f2b725696a61 https://medium.com/rapids-ai/gpu-accelerated-stream-processi... .