Shawn Wang: uncertainty

Transcript context

…Right. I think in the future, it should hopefully go more as this is a task and I return it. But we need to teach that to the model to train that, which is far from now. Every medium long-term direction that could be relevant here is thinking into latent space. I know some early works are doing that. And that's a way probably to move to first you think, and then you don't have to write all the tokens. Like it's in your head. It doesn't have to be as constricted than a plain text BLM. And once you have done your thoughts, you can just write the final answer or take an action. Just a commentary on that. Anthropic actually cheats at this right now. If you look at the system prompt in Claude Artifacts, I actually have a thinking section that is explicitly removed from the output, which is, I mean, they're still spending the tokens, but before training it, at the prompting level, you can simulate this. And then at iClear, there was the pause token, the backtrack token. I feel like all these are token level stopgap measures. I feel like it's still not the final form. We still need to have, at the architecture level, some kind of variable inference length thing that lets you actually think in latent space, like you're talking about. I don't know if there's any papers that you're thinking about. No, but that's interesting because that's what we said at the beginning of the discussion. If you remember, we are lacking flexibility for pre-training architecture transformers, where we spend the same amount of compute per token. And so because of that, how can you mitigate this? By generating more tokens, so more thoughts, more compute, because you have only access to this dimension. Ideally, you want an architecture that will enable, naturally, to make this emerge, basically.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.