Not sure, we don't have any particular measure for impact-adjusted plan changes per $, it's more just what in theory I think we should be aiming for. In practice we mostly just track connections, to what extent people find EAGs valuable, attendance, and other easy-to-track metrics.

Yeah that's right, I wouldn't want to raise the prices so much such that more senior folk/experts are put off (especially as they might be providing value, and as such it might feel weird to them to have to pay for that). Right now I expect we'll have a variety of ticket prices with optional discounts for those who need them, so I'm not too worried about more senior folks getting priced out here.

I think that's roughly right (I might put it as "impact-adjusted plan changes per $").

My understanding is that EAGxCambridge got an unusually cheap venue and as such was unusually cheap overall (it also only had ~500 attendees compared to 1000–1500 for EAGs). Other EAGx's often have quite small budgets because they're often held in countries that are much cheaper to hold events (and they also have fewer attendees).

Historically, EAGx's have also been scrappier and not been of as high production value — but I expect the gap to close here somewhat moving forwards.

It's not obvious to me whether larger venues cost more in a super-linear way, this is something I'm still confused about, but I expect it to not be the case overall.

No problem, all good! Re those venues:

• Fort Mason kinda works but has a similar issue of having lots of spaces kinda spread out over a wider area (some of which are kinda hard to find). The main networking area it has is the Festival Pavilion, which is huge but is also just an empty warehouse and would probably need to be built out a bit with lighting and furniture. This venue is weird and annoying enough that I'd only go for it if it was notably cheaper, but my understanding is that it's sort of similarly expensive to some of our other options.
• Moscone is great but yeah it's too big. To host an event there you need to commit to something like >1000 hotel rooms at their partner hotels. As you mention, you can apparently split the venue or like sub-let out rooms if a bigger event is going on. We tried investigating this for 2024 but it seemed risky (you'd sort of be at the larger events whim, and perhaps only getting confirmation very close to the dates themselves). I would be keen to investigate this further, but my guess is that it's not likely to be worth it (unless our event got much bigger).
• Yerba Buena is too small as you suggest.
• South San Francisco Conference Center is also a bit too small (but closer to the necessary size than Yerba Buena). They only have three proper rooms (though these can be subdivided with air walls), the largest of which can only fit 1200 people. I would be down to use this venue if it was sufficiently cheap (we'd have to limit capacity and change the event structure a fair bit), but my sense is that it's not quite worth it. There is an adjacent hotel (Hilton DoubleTree) that we could run a combined event with to have more session spaces, but my understanding is that this wouldn't let us go up to 1500 people either. And then I do think the location of these two venues isn't optimal — they're in South San Francisco and are difficult to get to by public transit, but I'd say this is a minor concern overall.

Re your second bullet, a large part of it is that one of the two people in a new connection may be financial unempowered. That is, they're a student or other early career person who isn't really able to spend much money on anything (this represents a lot of our attendees and is where philanthropy can jump in to fill the gap).

Sorry yeah, they cost way less, and if we were purely optimizing for connections per dollar then virtual conferences might be all that we do. So we are going to think about doing more of these moving forwards, though I do think it would be a mistake to optimize solely for connections per dollar.

It's still a bit unclear to me at this point whether a 1500 person event would cost less per person than three 500 person events. My current sense is that it is, but I don't have massive confidence in our investigations here.

Re agglomeration effects, I think basically yes it's better to run a larger event all things equal — the main benefits that come to mind are staff time and making it easier for people to coordinate and meet in larger groups (i.e. everyone you want to meet is going to the same event).

I don't think it's quite that it's a side hustle for them — it's mostly just that it's only a minority of attendees are working for EA orgs that are likely to be okay with them taking time off for an EA conference. If you're a biology student planning on working in biosecurity in the future, my guess is that you won't easily be able to move or skip your classes. Similar things might apply for people working in government or people who are skilling up outside of EA (e.g. as a law clerk).

There are EAs at lots of different companies/universities, but they don't work for CEA itself. Perhaps we should be asking around to see if we could still pull this off, but I do expect coordinating with an external party to be kinda tricky and unreliable (e.g. when we used UC Berkeley, the university really didn't want to talk to anyone at CEA other than the student in question). 

I also have reservations about working with something that's not an explicit venue-for-hire — I expect their staff to not be experienced at things like working with AV, catering, loading and unloading policies, and so on (example annoying scenario: we randomly get kicked out at 8pm because that's when the security team close everything down but this was never communicated to us even when we asked about it).

I'm not too sure about the Google campus, and perhaps we should look into this further. I do think lots of these options would involve multiple different spaces separated by ~10 minute walks. This might be fine for veteran EAs but might come across as wacky to any more external experts we were trying to engage with.

For the Harvard campus, my understanding is that the Science Center couldn't stretch to much more than a thousand. But yes in theory we could use multiple venues across the campus — this would likely involve coordinating and contracting with multiple parties (often the venues are managed by different groups) which is possible but not trivial. Klarman Hall appears to just be a large theater — generally schools are great at offering those but tend to be lacking in large networking/catering spaces.

However if any of the above options would be sufficiently cheap, then these downsides might be worth it, and I take your point about exploring more unconventional venue setups if it would save money (this is something that we're planning to explore in the future).