Hacker News Reader

It looks like that’s not the case anymore. That’s what this is all about.

It still is. The DeepSeek R1 is 670B parameters, requiring more than 300GB of video ram (or unified memory), even if using 4-bit quantized models.

How much video RAM was needed for similar performance a year ago?

Roughly the same, as our collective goalposts have shifted, and are still shifting.

What was amazing output last year is today's slop.

NV is in no immediate danger; medium/long term, to anyone knowing what "75% gross margin" means it's no secret that there will be serious threats. You don't need DeepSeek for that particular realization.

Which open source model last January gave you similar results with the same amount of video ram?

Nomic Falcon, Mixtral, YALM, Bloom, Grok-1 and Jamba, just to name a few of the foundation models.

Reply on news.ycombinator.com