# @kotekjedi_ml — 2026-08-11

♥1019 ↻95 · https://x.com/kotekjedi_ml/status/2087147093735714919

As you might guess, this suggests that distilling reasoning traces may have been possible for a long time without ever breaking the cryptography.

An anecdote: we find that prefilling Kimi-K3 reasoning with a few tokens of Opus reasoning measurably shifts its response toward https://t.co/KQgIM274UE

![photo](../../../media/HPcJN44a8AAuf_X.jpg)
![photo](../../../media/HPcJOKhagAATn0v.jpg)
tags: author:kotekjedi_ml, cot, has-image, kind:image, kind:tweet, model:kimi-k2, on:kimi-k3, year:2026
cited on: kimi-k3