If you're an EA org or a "rationalist" you may very likely maintain an (anonymous) suggestion box. So does The Unjournal. Are people using it? I guess not, or at best minimally. (Aside: please do give us feedback.)
Why not: Writing up useful criticism takes time. Making it anonymous can help make it less awkward or risky, but it's still not rewarded (and precludes the possibility for reputational benefit).
One could offer rewards for this feedback, but the promise may not be credible. It's a lot of work and bandwidth to follow up. And it's not easy to have org A compensate "anonymous person B" without giving up the anonymity.
There's also the time-inconsistency of informatino disclosure problem: once you has a good idea in hand, you might want to cheap out and not pay. .
I had this general idea for decades, but now maybe we can actually try it.
My proposal (I've been calling these suggestions "complainments"): ... somethinglike
- The org, individual, or business commits a monthly or yearly reward pool in advance, to be paid under published rules.
- Contributors submit through an independent administrator, which keeps their identity from the orgs but still they can be contacted.
- The administrator (org) assesses usefulness and implementation, or requires the organization to report on and score each suggestion.
- Rewards come out of the pool, and payouts are reported.
Money set aside in advance: harder to be stingy later. The administrator's fee would be separate and disclosed, so it gains nothing by rejecting claims.
And it can work better now, because "AI tools could eventually help with triage, grouping duplicates, and helping contributors clarify their points, with people making the payout decisions" (Quote: Claude 5.5 chiming in)
There are plenty of unresolved issues still to consider, with ~workable solutions I think (unused funds, duplicates, disputes, valuation, collusion, etc.).
The Unjournalwhich I co-direct, might volunteer as a pilot subject, with an independent administrator handling feedback about us. (But nothing is funded or committed yet.)
Has this been tried? Does it seem credible? Failure mode? Would you want to lead, help test, fund, or run a pilot?
More detail and short response forms: https://complainments.netlify.app/
(That page is AI generated, I'll improve it as I have bandwidth)
The core of the idea is promising. I think as you acknowledge, there's unresolved issues. My first is a lottery effect that attracts poor quality ideas. This ties to open of the issues you already anticipated, "We'd also need a rule for a month with nothing worth rewarding."
I'd suggest your primary solution here is longer time windows. I'd suggest rather than simply rolling forward an unclaimed month, have a two part reward. One, a low and fixed number, for a suggestion accepted as worth pursuing, and then the a second, from an reward pool accumulated overtime for implemented suggestions.
Organizations with a list of suggestions have to start off with triage, so the first part pairs with this. Suggestions would be rated. For example: not interested, curious, investigating, amazing. Curios would be a sort of backlog, and you'd provide a fixed small reward for getting an investigating or amazing rating. This would happen sort of real-time. Really, at the pace of the evaluators. These would draw from the reward pool, but you wouldn't expect them to exhaust it.
The reward pool remainder would be divided up yearly, or something like that. Any implemented suggestion would be evaluated on something like a 10 point-scale, and then the pool divided on a weighted basis.
You could play with variations: sliding windows, progressive evaluations (looks good at first, looks great 2 years later). But I think the generality of stretching the time-window is the core of the fix. An organization might want to consider additional rewards outside of the reward pool, so that there's a fixed contribution, but if 10 great ideas all land at once, they have the opportunity to fix that. Weakens your core idea a little, but I think non-catastrophically as the core-commitment can still stick.