DeepSeek open-sources inference optimizations with 60–85% faster generation [pdf] github.com 86 points by aurenvale 49 minutes ago
Havoc 8 minutes ago Nice.Guessing the timing isn't accidental. Demonstrated openness vs harsh regulation
Jackobrien 6 minutes ago I see a world soon where there’s an extremely wide variety of small models for speculative decoding, unique to use cases, companies, and even individuals.
ricardobeat 6 minutes ago Presumably this has been in production for a while, and is one of the reasons they were able to dramatically lower prices a month ago?
Nice.
Guessing the timing isn't accidental. Demonstrated openness vs harsh regulation
I see a world soon where there’s an extremely wide variety of small models for speculative decoding, unique to use cases, companies, and even individuals.
Presumably this has been in production for a while, and is one of the reasons they were able to dramatically lower prices a month ago?