Every story and comment from the Hacker News API, in a layout borrowed from 2001.
A comment on DeepSeek 4 Flash local inference engine for Metal. Parent comment
48 gb is enough for a capable LLM.
Doing that on consumer grade hardware is entirely possible. The bottleneck is CUDA and other intellectual property moats.
There are no comments on this comment yet.
48 gb is enough for a capable LLM.
Doing that on consumer grade hardware is entirely possible. The bottleneck is CUDA and other intellectual property moats.