Glad you posted this, Luke. I really like the idea of EA being more ambitious, and I think if the community onboards some of the suggestions in this post, itâs more likely to happen.
One specific comment: I am also concerned by overselling (discussed in section 6), but I think itâs worth being aware of how âEA oversellingâ might look different from regular overselling.
In my mind, I can imagine grant applications overselling while saying things that sound pretty humble, like:
âI think thereâs a small, but plausible chance that this organisation succeeds in its mission.â
âWe have an extremely competent director, though one risk is that theyâd have more impact on a different project.â
âWe understand that the base-rate of success for similar projects is quite low, but our team has a strong inside view that the project will succeed. Our credence in success is 70%.â
These kinds of statements genuinely do show self-awareness, and may be of interest to funders, so I think itâs a good thing that the community has a norm where this language is rewarded, and not punished. But depending on their context, they might only be a little humble, or even worse, could even have the effect of making the overselling harder to spot. Take the above example, where someone says itâs plausible but unlikely theyâll succeed in their mission. If their mission involves literally saving humanity from a key extinction risk, claiming itâs plausible theyâll succeed might actually be wildly overconfident (or at least misleading).
I also think EA overselling might come more from ommission:
Not sharing highly relevant, but negative, informationâe.g., criticisms others have made.
Providing a vision for your project that implies more alignment with the funder than there is in reality (e.g., the funder cares about X, but you care about X, Y and Z, yet in the application you de-emphasise Y and Z).
Underselling what you and the team would otherwise do if you didnât work on your project.
In any case, I think your suggestion is right. To get around this, itâs just really important funders and fundees have a strong, high-trust relationship. But I agree, thatâs hard to do if thereâs little communication between them.
Glad you posted this, Luke. I really like the idea of EA being more ambitious, and I think if the community onboards some of the suggestions in this post, itâs more likely to happen.
One specific comment: I am also concerned by overselling (discussed in section 6), but I think itâs worth being aware of how âEA oversellingâ might look different from regular overselling.
In my mind, I can imagine grant applications overselling while saying things that sound pretty humble, like:
âI think thereâs a small, but plausible chance that this organisation succeeds in its mission.â
âWe have an extremely competent director, though one risk is that theyâd have more impact on a different project.â
âWe understand that the base-rate of success for similar projects is quite low, but our team has a strong inside view that the project will succeed. Our credence in success is 70%.â
These kinds of statements genuinely do show self-awareness, and may be of interest to funders, so I think itâs a good thing that the community has a norm where this language is rewarded, and not punished. But depending on their context, they might only be a little humble, or even worse, could even have the effect of making the overselling harder to spot. Take the above example, where someone says itâs plausible but unlikely theyâll succeed in their mission. If their mission involves literally saving humanity from a key extinction risk, claiming itâs plausible theyâll succeed might actually be wildly overconfident (or at least misleading).
I also think EA overselling might come more from ommission:
Not sharing highly relevant, but negative, informationâe.g., criticisms others have made.
Providing a vision for your project that implies more alignment with the funder than there is in reality (e.g., the funder cares about X, but you care about X, Y and Z, yet in the application you de-emphasise Y and Z).
Underselling what you and the team would otherwise do if you didnât work on your project.
In any case, I think your suggestion is right. To get around this, itâs just really important funders and fundees have a strong, high-trust relationship. But I agree, thatâs hard to do if thereâs little communication between them.