Thread

1. techjamie · 1790095494 · focus · HN ↗

   With the performance gains they're claiming, I wonder if they implemented the Casual Encoder-Decoder technology from DeepSeek 4.1's paper.

   I could see them accomplishing it and seeing gains like this in roughly the correct timeframe, and when I heard about that development I assumed the frontiers would probably jump on it.

   How it works: <a href="https:&#x2F;&#x2F;miraflow.ai&#x2F;blog&#x2F;deepseek-v4-1-flash-causal-encoder-decoder-explained-2026" rel="nofollow">https:&#x2F;&#x2F;miraflow.ai&#x2F;blog&#x2F;deepseek-v4-1-flash-causal-encoder-...

   1. stri8ted · 1790098796 · focus · HN ↗

      This model was likely trained months before deepseek released their paper.

      1. dannyw · 1790116021 · focus · HN ↗

         Models are generally posttrained to a window shorter than you think.

Unofficial Hacker News client; not affiliated with Y Combinator.