Hacker News Reader

- Fail at the above.

I don’t think this is what happened with DeepSeek. It seems that they’ve genuinely optimized their model for efficiency and used GPUs properly (tiled FP8 trick and FP8 training). And came out on top.

The impact on the NVIDIA stock is ridiculous. DeepSeek took the advantage of flexible GPU architecture (unlike inflexible hardware acceleration).

View on HN

This is what I still don't understand, how much of what they claim has been actually replicated? From what I understand the "50x cheaper" inference is coming from their pricing page, but is it actually 50x cheaper than the best open source models?

50x cheaper than OpenAI's pricing on an open source model which doesn't require giving that quality level up. The best open source models were much closer in pricing but V3/R1 are that way while being a results topper.

Reply on news.ycombinator.com