I do agree with the premise/problem description (your first claim), but not with the solution/EA hotel—mostly because there might be more efficient ways to reduce the gap (e.g., the project evaluation platform). My main point is that, even if the EA hotel is the best way of supporting/incenting productive EAs in the beginning of their careers, it doesn’t solve the problem of selecting the best projects—the main cause of the Chasm; and even if someone has a great idea, they might not be the most capable of implementing them. More than the EA hotel, we need a platform to elicit project ideas and information, spread them among a community of people interested in evaluating them, and then support EA hotel guests interested in giving some of these ideas a try, or assign them to the best fit for implementation. Possible ways to fill this part of the chasm: hiring/training superforecasters, using prediction markets, contests, or maybe even just creating something like a voting system kind of incentive compatible (or just a particular forum) only for that.
My main point is that, even if the EA hotel is the best way of supporting/incenting productive EAs in the beginning of their careers, it doesn’t solve the problem of selecting the best projects
What do you think about the argument of using the processes in the hotel to filter projects? I tend to think that one way to cross the chasm is “just try as many projects as possible, but have tight feedback loops so you don’t waste too many resources.”
I think it’d still need a previous filter, and a good one would use many opinions of others. Maybe something lik crowdsource decision-making, as in your Metis project.
I do agree with the premise/problem description (your first claim), but not with the solution/EA hotel—mostly because there might be more efficient ways to reduce the gap (e.g., the project evaluation platform). My main point is that, even if the EA hotel is the best way of supporting/incenting productive EAs in the beginning of their careers, it doesn’t solve the problem of selecting the best projects—the main cause of the Chasm; and even if someone has a great idea, they might not be the most capable of implementing them. More than the EA hotel, we need a platform to elicit project ideas and information, spread them among a community of people interested in evaluating them, and then support EA hotel guests interested in giving some of these ideas a try, or assign them to the best fit for implementation. Possible ways to fill this part of the chasm: hiring/training superforecasters, using prediction markets, contests, or maybe even just creating something like a voting system kind of incentive compatible (or just a particular forum) only for that.
What do you think about the argument of using the processes in the hotel to filter projects? I tend to think that one way to cross the chasm is “just try as many projects as possible, but have tight feedback loops so you don’t waste too many resources.”
I think it’d still need a previous filter, and a good one would use many opinions of others. Maybe something lik crowdsource decision-making, as in your Metis project.