Hacker News Reader

somnialon DeepSeek V4 Flash on a Single AMD MI300X

throughput scales superlinearly with number of GPUs when networked well and deployed with wideEP, so 1x won't compare.

also it would be interesting to figure from the DSpark paper whether their numbers are consistent with the GPUs still being H800s, since they never actually say...

somnialon The gap between open weights LLMs and closed source LLMs

this is a blog post from a company that hosts open weights LLMs (https://www.doubleword.ai/). I think its possible it might have been tongue in cheek

somnialon The Economics of Speculative Decoding

https://fergusfinn.com/blog/economics-of-speculative-decodin...

good point tho - plus for Deepseek the shared expert increases the overlap slightly

somnialon Speculative KV coding: losslessly compressing KV cache by up to ~4×

true, but no reason the predictor model couldn't use linear attention (i.e. mamba, GDN etc) to predict KV caches