Twitxy Saved

Jacques

Verified account This account is verified. Learn more

@JacquesThibs

Montréal, Québec jacquesthibodeau.com Joined May 2008
1.7K Following 6.5K Followers 20K Posts

Latest posts 5

Jacques Verified account @JacquesThibs
this was my covert way of saying google would come back "from the dead" btw
Jacques Verified account @JacquesThibs Jul 17
Kimi K3 is obviously good and a big accomplishment, but I am indeed of the view that:

- There’s lots to pull out of pretraining via great post-training
- Using traces from other better models allows you to hit the ceiling much faster
- Therefore, an exceptional post-training process allows you to leapfrog stronger base models with weaker post-training (as Ryan effectively shares an example of)

My guess is that Google is much more careful about post-training on OpenAI/Anthropic traces, which is one reason they are behind the Chinese labs as of now.

I think it’s also true to say the Chinese labs are currently paying less attention and effort to safety-related stuff. Google has much more to lose if they release a model that does/says some crazy shit.

I know, I know, “distillation cope”, but I do want to say I’m fine being proven wrong and I also don’t deny the Chinese labs have strong technical teams with extreme talent density. I’m also not anti-open-source.
View quoted post on X
Jacques Verified account @JacquesThibs
One component of writing that is often a signal of good writing is the ability to say the quiet part out loud (the best tweeters have this trait!). I expect this is the sort of thing that AI-written text can lack, and therefore can feel devoid of insight or spark.