@SemiAnalysis_ reply (Wed Oct 21 23:47:26 +0000 2026): As agentic workloads grow, high interactivity cases, especially h | Hanami
@SemiAnalysis_ reply (Wed Oct 21 23:47:26 +0000 2026): As agentic workloads grow, high interactivity cases, especially how quickly a model generates tokens for each user, is becoming increasingly valuable.
We're also seeing AI coding tools accelerate kernel development, where correctness can be tested and performance can be measured. Faster iteration gives talented engineers more room to explore and refine new ideas.
We're excited to see what the next wave of kernel innovation brings. (8/8)
Original @SemiAnalysis_ reply (Wed Oct 21 23:47:26 +0000 2026): NVIDIA's GB300 trace shows one K3 layer block going from 23 kernels to 13, and from 130 to 69 microseconds.
Combining work of overlapping communication with other tasks and fetching weights early all help shorten the step.
Kimi-K3 CuTe DSL decode: https://t.co/51CJXTYBDH
image source: Inferact blog (7/8)
Original: unavailable