Comment by gehsty Comment by gehsty 11 hours ago 0 replies Copy Link View on Hacker News Only in the size of model it can run, not speed of token generation.