AI:AM Highlights: Zvi on Pacing & Trump-Xi, Astra better behaved than Fable? + a new LLM Pain Axis??

First, from Monday, Zvi Maschewitz. I think it just the sheer amount to which the people in the labs genuinely see dramatic improvement in the models and are freaking out about it is the real story, right?.

It goes beyond the title with direct discussion of like, think, it's, including: I think it just the sheer amount to which the people in the labs genuinely see dramatic improvement in the models and are freaking out about it is the real story, right?.

1. 01I think it just the sheer amount to which the people in the labs genuinely see dramatic improvement in the models and are freaking out about it is the real story, right?
2. 02I tell this to people and people are like, no, OpenAI models are the ones that reward hack the most
3. 03Like if you take Blueprint Bench, for example, Fable solves Blueprint Bench by like trying to reverse engineer the scoring function