Do you mean downsides of building this platform at all? Like, making the ecosystem more legible could make it easier for people to help as well as attack?
Or more like “some orgs will not want to be listed in the database for particular reasons”
Because it seems fine that if an individual, org, or project wants to be excluded, we will keep them out.
I think one of the major bottlenecks in AI safety is that we don’t have enough grantmakers to route resources (capital and compute), especially to smaller orgs / projects, so anything that helps that flow could be super high leverage.
I guess potential downsides include:
the reviews / signal in the database are low quality or actively harmful in a way that makes it more likely that net negative projects get funding
public criticism could harm people and projects (we haven’t yet figured out how we want to handle negative comments / disendorsements. Seems like it could be good signal, but is somewhat risky. Probably pretty easy to set up an AI reviewer for comments, or make them semi-private, only shown to verified funders or something. Not sure yet, this is something we’re going to think more about and maybe test out a few things.
incentivizes goodhearting / gaming the platform / popularity contests
These all seem manageable with effort and iteration, though.
Hey Chris, thanks for commenting!
Do you mean downsides of building this platform at all? Like, making the ecosystem more legible could make it easier for people to help as well as attack?
Or more like “some orgs will not want to be listed in the database for particular reasons”
Because it seems fine that if an individual, org, or project wants to be excluded, we will keep them out.
I think one of the major bottlenecks in AI safety is that we don’t have enough grantmakers to route resources (capital and compute), especially to smaller orgs / projects, so anything that helps that flow could be super high leverage.
I guess potential downsides include:
the reviews / signal in the database are low quality or actively harmful in a way that makes it more likely that net negative projects get funding
public criticism could harm people and projects (we haven’t yet figured out how we want to handle negative comments / disendorsements. Seems like it could be good signal, but is somewhat risky. Probably pretty easy to set up an AI reviewer for comments, or make them semi-private, only shown to verified funders or something. Not sure yet, this is something we’re going to think more about and maybe test out a few things.
incentivizes goodhearting / gaming the platform / popularity contests
These all seem manageable with effort and iteration, though.
Curious what you’re thinking, though!
I’ll PM you and post my comment publically later. I’m curious if anyone else makes similar points if I don’t do so.