I really like the focus on superintelligence alignment and theory. Many current intro programs are trying to be “very empirical” which sometimes results in transitioning into something like corporate upskilling for people who are going to do finetuning of enterprise LLMs, both in terms of program content and selection of participants. It’s not to say that being empirical is bad, it’s just that a clear large-scale methodological picture should be delivered as well, and people should also be filtered for seriousness about it.
I really like the focus on superintelligence alignment and theory. Many current intro programs are trying to be “very empirical” which sometimes results in transitioning into something like corporate upskilling for people who are going to do finetuning of enterprise LLMs, both in terms of program content and selection of participants. It’s not to say that being empirical is bad, it’s just that a clear large-scale methodological picture should be delivered as well, and people should also be filtered for seriousness about it.