alt Hacker NewsThe level of discourse here anytime distillment is mentioned is so mundane. Do we need 40 people saying the same thing about how they don't feel bad and it serves them right and all that surface level stuff on every single one of these? This is the level of insight one receives anytime you mention chocolate and dogs "Oh it's poisonous for dogs!". Yes, we've all heard that 100 times. Thanks for adding nothing to the conversation.
A more interesting part of this discussion is that consistently these Chinese models are held up as a great achievement, and that they're "catching up" when in reality they're just using the work of Anthropic and OpenAI to try and keep up with them. This isn't even to say it's not a valid tactic, but it definitely colors these announcements and proclamations about foreign companies catching up to American ones.
If I get a 1600 on the SAT and you copied my answers and got a 1540, your achievement isn't that significant.
Replies
> Do we need 40 people saying the same thing about how they don't feel bad and it serves them right and all that surface level stuff on every single one of these?
Apparently we do, given that we've got govt officials wasting time, money, and effort on whining about this now.
Let's be real. In reality, it's the Chinese kid getting the 1600.
Yes they distill, but if you think you can trivially get a frontier-level model by "just" distilling from Claude's public API. you fundamentally do not understand the amount of work that goes into a modern post-training stack.
Without even talking about the fact that any distillation that was done was on Opus, as the timeline of Mythos/Fable vs Kimi 3 release dates just do not match in any plausible way.
If you want to read an educated take from someone that has actually spent the last few years working on post training I recommend Nathan Lambert's: https://x.com/natolambert/status/2079616308203942332