I wish there was a way to do granular disagreement on this. While this is a clear case for forecasting and estimation, all issues boil down to how big quantities are.
There is no way for me to comment on specific lines or specific numbers. And technically speaking I don’t think this can be too difficult since LessWrong has a docs-style editor. I suggest that inline comment threads are hidden by default and users can turn them on if they want to. Perhaps they get brighter the more upvoted they are.
Secondly, I wish estimation was a first-class citizen of the forum. Imagine if for each of these models, users could add their own numbers and we could see a community median across all supplied values. For many articles, I think that would be really valuable/game-changing.
I reckon that it’s gonna happen by 2030 and I’m happy to bet to that effect
Because I think crowdsourced forecasts without any incentives are basically random. It’s going to give a wrong semblance of information.
Same for estimations that are forced to be numerical even when that’s not appropriate. I think EA should have less forecasting and estimation, and not more.
(I do appreciate the group making this post—see my top level comment)
I unironically think that almost everything in EA should be numerical somewhere. Maybe we have a text layer to explain the numbers, but deep down everything is about the sizes of quantities.
If you think crowdsources forecasts without incentives are random, what do you think about crowdsourced discussions?
I down x-ed your post. I get that feels harsh, but I strongly think you are wrong.
I wish there was a way to do granular disagreement on this. While this is a clear case for forecasting and estimation, all issues boil down to how big quantities are.
There is no way for me to comment on specific lines or specific numbers. And technically speaking I don’t think this can be too difficult since LessWrong has a docs-style editor. I suggest that inline comment threads are hidden by default and users can turn them on if they want to. Perhaps they get brighter the more upvoted they are.
Secondly, I wish estimation was a first-class citizen of the forum. Imagine if for each of these models, users could add their own numbers and we could see a community median across all supplied values. For many articles, I think that would be really valuable/game-changing.
I reckon that it’s gonna happen by 2030 and I’m happy to bet to that effect
I’m willing to bet* the forum will support it—and I think it’s a bad idea.
*Not, like, on the actual platform
Why do you think it’s a bad idea?
Because I think crowdsourced forecasts without any incentives are basically random. It’s going to give a wrong semblance of information.
Same for estimations that are forced to be numerical even when that’s not appropriate. I think EA should have less forecasting and estimation, and not more.
(I do appreciate the group making this post—see my top level comment)
I unironically think that almost everything in EA should be numerical somewhere. Maybe we have a text layer to explain the numbers, but deep down everything is about the sizes of quantities.
If you think crowdsources forecasts without incentives are random, what do you think about crowdsourced discussions?
I down x-ed your post. I get that feels harsh, but I strongly think you are wrong.
Much more informative—in accordance with what my other comment saying I find the reasoning in the OP much more informative than the numbers.
Feels honest rather than harsh. But thanks for the sympathy, and it’s nice to know where the votes come from.
Obviously, I also strongly disagree with you :)