I think it is quite plausible that £2m is too low for the year. Not having enough funding increases the costs to applicants (time spent applying) and you (time spent assessing) relative to the benefits (funding moved), especially if there are applicants above the bar for funding but that you cannot afford to fund. Also I had this thought prior to reading that one of your noted mistakes was “underestimated the number of applications”, it feels like you might still be making this mistake.
I haven’t thought about this much, but a natural strategy is to try to have a budget sufficiently large that you know you’ll definitely be able to fund all the good projects, and then binary search down to the amount that only funds all the good projects.
I’d also be interested to find out what happens if CEA announces they’re budgeting like 5 million for this, and see if there are any good projects that appear when that much money is potentially available in the community. Naturally CEA neednt give it all away.
(But right now I’d expect most of the best projects are just 1-3 people’s full time salaries for a small team to work together, so each grant being <200k at most.)
Added: On the margin I’d expect the most useful thing EA Grants could do would be to offer multi-year grants, so people in the community can consider major careeer changes based on what’s most effective rather than what’s most stable.
I haven’t thought about this much, but a natural strategy is to try to have a budget sufficiently large that you know you’ll definitely be able to fund all the good projects, and then binary search down to the amount that only funds all the good projects.
I’d also be interested to find out what happens if CEA announces they’re budgeting like 5 million for this, and see if there are any good projects that appear when that much money is potentially available in the community. Naturally CEA neednt give it all away.
(But right now I’d expect most of the best projects are just 1-3 people’s full time salaries for a small team to work together, so each grant being <200k at most.)
Added: On the margin I’d expect the most useful thing EA Grants could do would be to offer multi-year grants, so people in the community can consider major careeer changes based on what’s most effective rather than what’s most stable.