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:...
speaking of covid-19, what do you think of the clapping towards the NHS doctors going on in the UK?
i know they do good work and all, but i feel like the general public are at least half-ignoring the people like the policy people and researchers. i think we should be clapping for them too, as they are, in principle, the ones that can put in practices and polices that can help prevent hospitals from being overfilled and actually solve this issue - even in the short term (as you can see with china, taiwan and south korea).
i would love to see your thoughts, thanks
p.s.
i donated a small amount to two of the organisations recommended on the 80k website and also listened to your podcasts about COVID-19 (among as well non-covid-19 topics). although i don't understand all the details of every episode, i do think you make for a rather entertaining and informative host and i do look forward to future episodes.
Hi PBS, I understand where you're coming from and expect many policy folks may well be having a bigger impact than front-line doctors, because in this case prevention is probably better than treatment.
At the same time I can see why we don't clap for them in that way, because they're not taking on a particularly high risk of death and injury in the same way the hospital staff are right now. I appreciate both, but on a personal level I'm more impressed by people who continue to accept a high risk of contracting COVID-19 in order to treat patients.