Why assume Roland assigns epsilon probability to miracles? If he’s sufficiently uncertain about atheism so as to want to look for miracles, I’d say going out and looking is the best possible thing he can be doing. It’s too much to expect that he go from being a Christian to having epsilon probability of God right away. If a few years after being a church-going Christian he’s now at 5% probability of God, I wouldn’t say he’s doing anything wrong. Even Jeffreyssai doesn’t say every decision should be made in less than a second. Just don’t take thirty years.
It could be argued that because of the importance of miracles (a single one would prove some form of religion, religion is a very important issue) it’s worth keeping an eye out for them even at such low levels of expectation that you’d give up looking for, say, the Higgs boson.
If thirty years from now he’s still looking for miracles, that would be a problem, but the thought-mode that one should look for evidence if one is uncertain wouldn’t be the issue. The issue would be whatever it was that was keeping his probability distribution at 5%.
If a few years after being a church-going Christian he’s now at 5% probability of God, I wouldn’t say he’s doing anything wrong.
I disagree. Changes of opinion about conclusions should be swift and decisive, which doesn’t mean that in the same movement you should wipe out from your mind the understanding and experience gained from the previous, invalidated position. Changing your mind swiftly, while keeping the background that allows to regain mastery in the disbelieved position in case it returns to plausibility, seems to bring the best of both practices. This is the attitude I have towards Robin Hanson’s modesty argument: you should change your conclusion towards the universally accepted one, but still be ready to change it back in an informed manner.
The costs of acting on the overly inert and therefore wrong conclusion project on everything you do, while the cost of still thinking of and remembering the background that lead to the very unlikely conclusion are low enough to keep them. At one point you stop actively researching your unlikely idea, at the next point you stop thinking about it, and some clicks further you forget it entirely. Just don’t be overconfident about the conclusion, thinking you know how overwhelmingly, 10 to the minus 1000 unlikely it is, when in fact it turns out to be correct.
Why assume Roland assigns epsilon probability to miracles? If he’s sufficiently uncertain about atheism so as to want to look for miracles, I’d say going out and looking is the best possible thing he can be doing. It’s too much to expect that he go from being a Christian to having epsilon probability of God right away. If a few years after being a church-going Christian he’s now at 5% probability of God, I wouldn’t say he’s doing anything wrong. Even Jeffreyssai doesn’t say every decision should be made in less than a second. Just don’t take thirty years.
It could be argued that because of the importance of miracles (a single one would prove some form of religion, religion is a very important issue) it’s worth keeping an eye out for them even at such low levels of expectation that you’d give up looking for, say, the Higgs boson.
If thirty years from now he’s still looking for miracles, that would be a problem, but the thought-mode that one should look for evidence if one is uncertain wouldn’t be the issue. The issue would be whatever it was that was keeping his probability distribution at 5%.
I disagree. Changes of opinion about conclusions should be swift and decisive, which doesn’t mean that in the same movement you should wipe out from your mind the understanding and experience gained from the previous, invalidated position. Changing your mind swiftly, while keeping the background that allows to regain mastery in the disbelieved position in case it returns to plausibility, seems to bring the best of both practices. This is the attitude I have towards Robin Hanson’s modesty argument: you should change your conclusion towards the universally accepted one, but still be ready to change it back in an informed manner.
The costs of acting on the overly inert and therefore wrong conclusion project on everything you do, while the cost of still thinking of and remembering the background that lead to the very unlikely conclusion are low enough to keep them. At one point you stop actively researching your unlikely idea, at the next point you stop thinking about it, and some clicks further you forget it entirely. Just don’t be overconfident about the conclusion, thinking you know how overwhelmingly, 10 to the minus 1000 unlikely it is, when in fact it turns out to be correct.