Everything can be phrased as distribution modelling.
That might a big claim, since Beren thinks there’s a real difference in type, and one example is that he thinks alignment solutions for model based agents coming out of GPT-N can’t work, due to amortized optimization. Thus, a non-vacous restriction is there.
That might a big claim, since Beren thinks there’s a real difference in type, and one example is that he thinks alignment solutions for model based agents coming out of GPT-N can’t work, due to amortized optimization. Thus, a non-vacous restriction is there.