better operating systems and/or better programming languages
I doubt this is worth pushing on now. If it’s useful, it’ll be useful when we’re closer to doing engineering rather than philosophy and math.
Do you know which of the open problems MIRI is likely to attack first?
In the immediate future we’ll keep tackling problems addressed in our past workshops. Other than that, I’m not sure which problems we’ll tackle next. We’ll have to wait and see what comes of Eliezer’s other “open problems” write-ups, and which ideas workshop participants bring to the November and December workshops.
potentially applying the method from the probabilistic reflection paper to the Halting Problem, since it seems to share the same self-referential structure
The participants of our April workshop checked this, and after some time decided they could probably break Tarski, Godel, and Lob with probabilistic reflection, but not the Halting Problem, despite the similarities in structure. You could ask (e.g.) Qiaochu if you want to know more.
I doubt this is worth pushing on now. If it’s useful, it’ll be useful when we’re closer to doing engineering rather than philosophy and math.
In the immediate future we’ll keep tackling problems addressed in our past workshops. Other than that, I’m not sure which problems we’ll tackle next. We’ll have to wait and see what comes of Eliezer’s other “open problems” write-ups, and which ideas workshop participants bring to the November and December workshops.
The participants of our April workshop checked this, and after some time decided they could probably break Tarski, Godel, and Lob with probabilistic reflection, but not the Halting Problem, despite the similarities in structure. You could ask (e.g.) Qiaochu if you want to know more.