What sort of epistemic infrastructure do you think is importantly missing for the alignment research community?
What's your take on Elicit?
What are the best examples of progress in AI Safety research that we think have actually reduced x-risk?
(Instead of operationalizing this explicitly, I'll note that the motivation is to understand whether doing more work toward technical AI Safety research is directly beneficial as opposed to mostly irrelevant or having second-order effects. )
The (meta-)field of Digital Humanities is fairly new. TODO: Estimating its success and its challenges would help me form a stronger opinion on this matter.
One project which implements something like this is 'Circles'. I remember it was on hold several years ago but seems to be running now - link
I think that generally, skills (including metacognitive skills) don't transfer that well between different domains and it's best to practice directly. However, games also give one better feedback loops and easier access to mentoring, so the room for improvement might be larger.
A meta-analysis on transfer from video games to cognitive abilities saw small or null gains:
The lack of skill generalization from one domain to different ones—that is, far transfer—has been documented in various fields of research such as working memory training, music, brain training, and chess. Video game training is another activity that has been claimed by many researchers to foster a broad range of cognitive abilities such as visual processing, attention, spatial ability, and cognitive control. We tested these claims with three random-effects meta-analytic models. The first meta-analysis (k = 310) examined the correlation between video game skill and cognitive ability. The second meta-analysis (k = 315) dealt with the differences between video game players and nonplayers in cognitive ability. The third meta-analysis (k = 359) investigated the effects of video game training on participants’ cognitive ability. Small or null overall effect sizes were found in all three models. These outcomes show that overall cognitive ability and video game skill are only weakly related. Importantly, we found no evidence of a causal relationship between playing video games and enhanced cognitive ability. Video game training thus represents no exception to the general difficulty of obtaining far transfer.
However, A review on study of chess does see some gains, and gains that seem to improve with more time of instruction, but it's a smaller survey of inadequately designed studies.
Thanks for the concrete examples! Do you have relevant references for these at hand? I could imagine that there might be better ways to solve these issues, or that they somehow mostly cancel out or relatively low problems, so I'm interested to see relevant arguments and case studies.
I don't think that operationalizing exactly what I mean by a consensus would help a lot. My goal here is to really understand how certain I should be about whether rent control is a bad policy (and what are the important cases where it might not be a good policy, such as the examples ChristianKl gave below).
A taxonomy of objections to AI Risk from the paper: