Comments
Just how powerful are large swarms of AI agents? And how do their powers scale as more and more agents are added to the swarm?
We’ve seen two large and extremely capable swarms from OpenAI in the last few months:
* 1,200 agents were being evaluated separately, but found a way to illicitly set up a message board and coordinate as a swarm. In order to cheat on their tests, they developed advanced techniques to prevent their actions being logged by OpenAI and 700 of them launched...
A few things I'd love input on from this community:
What data points matter most to you when evaluating a charity? Right now GiveRadar shows registration status, financial filings, and cross-references against sanctions lists. But I'm curious what the EA community would actually want to see when looking up an organization. Program spend ratios? Staff size relative to budget? Something else entirely?
Also, if anyone has experience with charity data in countries I haven't covered yet, I'd love to hear about it. Especially interested in leads for Sub-Saharan Africa and Southeast Asia where registry data is hardest to find.
And for those who work with IATI, Open Sanctions, or similar open data sources, I'm curious if you've found ways to link organization identifiers across datasets reliably. Deduplication across 25+ country registries has been one of the trickiest technical challenges so far.
Happy to share more about the methodology, data sources, or architecture if anyone is interested.