A moderate's view on the recent events and trajectory of AI.

I was underestimating how much we are likely to scale inference-time compute in the near future, but this is different than RSI and I have not seen enough to convince me that an intelligence explosion is near.