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...
Often folks hit us up because they are thinking of starting an incubator and want advice.
Typically their motivation is either that (a) they have a list of specific things they want built that no one is building, or (b) they think an ecosystem needs more new projects generally to absorb more talent and deploy more funding effectively.
Here are six questions we often ask prospective teams, to help them figure out what to do. If you're incubator-curious...
Summary:
First, I give several different angles on how I feel about reinforcement learning:
* Theoretical case: RL is a black-box source of agency — this should give us classic misalignment worries, especially compared to agency-via-scaffolding
* Recent incidents (huggingface etc) and more mundane forms of misaligned behaviour in personal use give me bad vibes about the direction-of-travel of recent AI progress
* I’m worried things might get worse:...
After a few years of missing the mark, this year I've exceeded my goal of giving 10% of my income away by a substantial margin (I never took the Giving What We Can pledge, but I still aspire at this point in my career to exceed the 10% bar).
It's bittersweet, because I think that the reason I succeeded is that it seems like there's more funding gaps than there were a few years ago - insofar as that's about there being more good giving opportunities (which I think it partly is) that's exciting but I also think it's partly due to there being more promising opportunities that aren't getting funded by large funders or other small funders, for a variety of reasons. And that's unfortunate.
I got into EA more than ten years ago when the focus was more squarely on giving. I think that the relative switch over to careers has been a great thing and to this day I think the overwhelming majority of my impact will come through my choices about my career rather than my giving. But I'd encourage "old school" EAs like me who haven't been as focused on personal giving to re-engage with the question of:
Thanks Claire! Are you able to point to some newer opportunities you think are especially promising?
Thanks for sharing this! @Julia_Wise🔸 and I also decided to give more in 2025 (or early 2026), from a combination of pressing funding gaps and wanting to pull giving forward because of Anthropic donors.