We don't want to post again "This might be the last AI Safety Camp"

By Remmelt @ 2025-01-21T12:03 (+38)

This is a linkpost to https://manifund.org/projects/11th-edition-of-ai-safety-camp

This is a crosspost, probably from LessWrong. Try viewing it there.

null
John Salter @ 2025-01-21T19:47 (+10)

Two questions I imagine prospective funders would have:

  1. Can you give some indication as to the value of stripends? It's not clear how the benefits trade off against that cost. It's tempting to think that stripends are responsible for >80% of the costs but bring <20% of the benefit.

  2. What would your attendees have been doing otherwise?

     

Linda Linsefors @ 2025-01-23T13:14 (+5)

If I calculated correct, in the fully funded version, stipends would be 76% of the cost. Not quite >80% but close. I think I agree that stipends is not much more than than 20% of the value. 

Basically I agree with you that stipends are the least cost effective part of AISC. This is why stipends are lowest on the funding priority. 

However it is possible for stipend to be less necessary than the rest, but still worth paying. They are in the budget because, if someone wants to fund it, we would like to hand out stipends. 

I think giving stipends to low participants from low income countries are probably cost effective, but it's probably better to prioritise runway for future camps rather than stipends for everyone else. If you know any donors who would like to earmark their donation this way, or any other way, tell them to contact us.