Will mechanistic interpretability have more academic impact than representation engineering by the end of 2025?

Will manifold markets meaningfully affect p(doom) by more than 3%?

Will agent foundations [eg Scott Garrabrant] end up affecting p(doom) more than 5%?

Will davidad meaningfully affect p(doom) by more than 3%?

Will mechanistic interpretability be essentially solved for GPT-4 before 2030?

Will MIRI meaningfully affect p(doom) by more than 5%?

Will mechanistic interpretability be essentially solved for the human brain before 2040?

Will mechanistic interpretability be essentially solved for GPT-3 before 2030?

Will janus/@repligate meaningfully affect p(doom) by more than 5%?

Will mechanistic interpretability be essentially solved for GPT-2 before 2030?