For example, with $2k, I expect I could hire a pub in central London for an evening (or maybe a whole day), with perhaps around 100 people attending. So that's $20 per person, or 1% of the cost of EAG. Would they get as much benefit from attending my event as attending EAG? No, but I'd bet they'd get more than 1% of the benefit. 

Worth noting these aren't necessarily mutually exclusive. It's possible both running EAGs and running these smaller events are above current funding bars.

Firstly, I agree with Daniel that we should just do both. Smaller events like the one you're suggesting here are worth doing (and I expect local EA groups do exactly this)

But I think there are effects that kick in only when events reach a certain size, e.g.

• Speakers/experts will travel if they can speak to hundreds of people, but not to a room.
• Similarly, if travel is costly for attendees, they might only make the trip for one large event, but not for a small event.
• If you're looking for new opportunities, you want to speak to a wide range of people, and you might not get that with a small event.

FWIW, we have funded lots of small retreats and compared their cost-effectiveness with EAGx events, the post about that is here. The retreats were much more expensive than the pub idea, but we found that they produced a similar amount of value per person despite being almost twice the cost per person. EAG is even more expensive per person, but has the effects described above.

For example, with $2k, I expect I could hire a pub in central London for an evening (or maybe a whole day), with perhaps around 100 people attending. So that's $20 per person, or 1% of the cost of EAG. Would they get as much benefit from attending my event as attending EAG? No, but I'd bet they'd get more than 1% of the benefit. 

Actually, I'm not sure this is right. An evening has around 1/10 of the networking duration of a weekend, and number of connections are proportional to time spent networking and to number of participants squared. If this is 1/10th the attendance and 1/10th the duration, you're getting around 1/1000th the networking value for 1/1000th the cost, so not obviously better than an EAG weekend that costs $2M and has 1000 participants, maybe worse when you factor in people's time cost of getting to the pub as a fraction of the event duration, or maybe better if they prefer not to spend a whole weekend.

+1 that I'd love to know the assumptions behind the subsidies. I'd also like information that makes it easier for people to self-select into different EA events, beyond simply "knowledge of EA". Historically, what kind of people in what kind of situations have gotten the most value out going to an EAG conference? Who's added the most value? What kind of people do you really wish would attend? Obviously it couldn't be exhaustive, and maybe this would end up in the same kind of bad place as people over-indexing on 80k's advice, but I'd definitely be curious!

Thanks — these are some great questions, I'll try to provide some more context on some of these below:

- We do currently have EAGxVirtual scheduled for November, though this is an EAGx rather than an EAG (and perhaps we should consider doing an EAG Virtual again). When we've looked at the stats, our virtual conferences do seem to get around half the connections per person. I also have the intuition that people are much less likely to make big career changes from virtual events (things like "I went to EAG London and it pushed me to drop out of my PhD"). But naturally given how cheap virtual events are, I agree we could be exploring these more.

- To discuss subsidies (somewhat mentioned here but also mentioned elsewhere and in the replies): I see there as being two main (connected) reasons here. One reason is that a lot of attendees are providing value to other people as well as than themselves, meaning that the value to them is maybe $500 but the total value of their attendance is maybe $1000 (made up numbers). This may be especially true for more senior members of the community who can provide a lot for others but perhaps have little to gain for themselves. The second reason is that a lot of our attendees don't have that much disposable income — a lot of the people who can get the most value from advice and new connections are students, unemployed, or just generally not in a stable enough career path to be able to spend much money. But I'll note that I generally agree that it would be better if more EAs paid for the services they used, hence why we're planning to raise ticket prices (and are likely going to ask people who can afford it to pay for their ticket in its entirety).

- I'll note that we are somewhat experimenting with other types of events on other parts of the events team, with things like EAGx's in new regions and various retreats/smaller events (like the Summit on Existential Security) — and we continue to keep doing this in the future.

- We see the main route to impact for EAGs as (briefly): people forming new connections and having conversations that shift or accelerate them into higher-impact career paths. Though naturally this isn't a super in-depth explanation, and perhaps we should share something more extensive in the future (I expect such a post to discuss a lot of concrete past examples of the above). I also feel like it could be a mistake to isolate the routes to impact down to just one or a few items, as part of me sees large events like this as somewhat nebulous, and likely to create impact in surprising or unexpected ways.

- And then to briefly touch on the venue closing thing — at least for the past two years we've closed our doors at ~10pm each day, largely because we need to move around furniture and set stuff up for the next day (as well as let staff sleep). Though I'd guess there's some previous EAG or EAGx event pre-2022 that you're referring to here that had earlier closing times (this may have been due to cost or staff capacity, not sure).

Thanks for sharing this Sebastian! We haven't explicitly asked people whether weekends work better than weekdays, though this has now come up a couple of times such that I'd like to do so in the future.

But my expectation is that weekends would work much better for most of our attendees as few of them have children (even the more senior folks). A lot of our attendees are students who might have classes during the week and many others work jobs for which they can't easily take time off for our event (jobs in government or academia for example).

Especially as I seem to recall EAG was put on Easter/Passover weekend one year.

Strongly agree. Of course it's different what works for different people but I think it's a little odd that both EAG and EAGx seem to always be over the weekend, and I would be curious to see how the composition of attendees would shift if an event was held on work days.

It’s more about functionality and capacity than anything else. For the Bay Area, the Google campus isn’t something that’s publicly available and you need to have a way in somehow. I’m not entirely sure how we managed to get it in 2015 (that was before my time). My understanding is that EAG 2015 only had ~400 attendees, and from chatting to people who were there, it seems pretty clear that the space they had couldn't host much more than that (though perhaps the complex has other rooms/spaces that can be used).

Re UC Berkeley, I'll note that I still see this one as a near-viable option — the space has a bunch of issues but could work for something EAG-like in the future. With EAGs current setup it seems like it'd be tricky to fit 1500 people or more — we'd have to cut down on a lot of the sit-down seating and the number of session rooms. It’s also challenging in other ways, as a large event there could only be held at specific times of the year when school isn’t in session and the venue is also spread across several different buildings and levels. When it was used in 2016 one of CEA's staff was a student at the university, so we got better access and reduced pricing that would be hard for us to get now. (But to clarify, I'm still somewhat excited about UC Berkeley as an option in the future.)

Re the Harvard campus, my understanding is that couldn’t fit 1500 people either — IIRC EAG Boston 2017 only had 300 people or so.

This feels like a "be the change you want to see in the world" moment. If you want such an event, it seems like you could basically just make a forum post (or quick take) offering 1:1s?

I share your concern, but I think it would be more than offset by the competition with EAGx. My understanding from the dashboard, and from personal experience, is that EAGx has roughly the same value as EAG for first-timers (and many/most non first-timers) while being ~3x more cost-efficient. (If you went to both, would you rather pay $2k to go to an EAG or $600 to go to an EAGx as a first timer?)

I think this is good also because EAGx are mostly organized by different community members and national groups staff, so different groups could try different strategies and we could see what works best (e.g. minimize printing, have the many bored idle volunteers record talks with their phones instead of paying thousands per recorded talk, ...)
It might also help move EA towards being more of a do-ocracy and less of an "EV-ocracy".

Another great advantage imho is that people and community builders would be more mindful of other opportunities, or even try to start alternatives to EAGs, that have been crowded out by the current subsidies.

Feels like there could be some way to offer subsidy here. But it does seem good for the costs of things to be properly subsidised. If the benefit of one's first EAG is higher then maybe there could be a first EAG subsidy.

I'm not sure I count as 'senior', but I could understand some reluctance even if 'all expenses paid'. 

I consider my EAG(x) participation as an act of community service. Although there are diffuse benefits, I do not get that much out of it myself, professionally speaking. This is not that surprising: contacts at EAG (or knowledge at EAG, etc. etc.) matter a lot less on the margin of several years spent working in the field than just starting out. I spend most of my time at EAG trying to be helpful - typically, through the medium of several hours of 1-1s each day. I find this fulfilling, but not leisurely.  

So from the selfish perspective EAG feels pretty marginal either re. 'professional development' or 'fun'. I'd guess many could be dissuaded by small frictions. Non-hub locations probably fit the bill: "Oh, I could visit [hub] for EAG, and meet my professional contacts in [hub] whilst I'm in town" is a lot more tempting to the minds eye than a dedicated trip for EAG alone.

Speaking for myself, my org would definitely be happy to reimburse travel. But I very much dislike travelling for a number of reasons including travel time and jet lag increasing the cost significantly. I don't want to be away from my family longer than necessary, in part because I already optimise fairly strongly for working long hours. So I'm most likely to go to EAGs nearby. Like Greg, going to another EA hub has advantages that sometimes offset the cost of needing to travel for me.

I currently work at a large EA-ish org that allows me to fully expense EAG travel and I (like some of the other commenters) am pretty strongly in the "prefer hub" camp. Like lots of EAs, I try to intensely optimize my time, and I'd prefer to optimize for work and play separately (so I would prefer to focus on work when going to EAGs, then separately take vacations optimized for being fun for me, e.g. by being in a place that's a great fit for me and my primary partner). I am happy to travel occasionally if there's a strong impact justification, but don't want CEA trying to influence me to do travel for fun at a location and time it picks. In my experience, EAs in general are more intense about their time and possibly less into travel than most people in academia. 

Even if you assume everyone would go, I don't think it's a clear win. I think a lot of professionals in the space place a lot of value on an hour of their labor; if they value it at $100/hour (i.e. equivalent to $200k/year in donations), and you make them travel e.g. 12 hours roundtrip to get to a conference location, and that affects 300 attendees who would otherwise have had reasonable in-city daily commutes, that's $360k-equivalent added (though in reality I agree many just wouldn't go, and some would also do the vacation thing so this would funge against hours they'd spend traveling for vacation anyway). Then additionally, you have EA orgs paying the travel costs themselves, which maybe looks better for CEA but is the same to EA funders (though maybe some people can also expense it to non-EA orgs?). If the orgs are paying $1000 per person (let's say $400 on travel, $450 for 3 nights of hotel rooms at $150/night, the rest for meals and other incidental expenses) and 300 more people need to travel than otherwise would if the EAG were in a hub area, that's another $300k.

Also, CEA staff probably benefit from specialist knowledge of cities they often run EAGs in, so either they are stuck in the same non-hub city repeatedly, or they probably suffer costs of trying to run conferences in cities they aren't used to. 

It'd be partly counterbalanced, in addition to being less expensive to CEA, by being less expensive for the people that would need to travel either way (lower hotel and meal costs in lower cost-of-living cities), to get to the EAG, so I don't think it's an obvious call.

There’s a school of thought that academics travel much much more than optimal or healthy. See Cal Newport’s Deep Work, where he cites a claim that it’s “typical for junior faculty to travel twelve to twenty-four times a year”, and compares that to Radhika Nagpal’s blog post The Awesomest 7-Year Postdoc or: How I Learned to Stop Worrying and Love the Tenure-Track Faculty Life which says:

I travel at most 5 times a year. This includes: all invited lectures, all NSF/Darpa investigator or panel meetings, conferences, special workshops, etc. Typically it looks something like this: I do one or two invited lectures at places where I really like the people, I go one full week to a main conference, I do maybe one NSF/Darpa event, and I reserve one wildcard to attend something I really care about (e.g. the Grace Hopper Conference, or a workshop on a special topic). It is *not easy* to say no that often, especially when the invitations are so attractive, or when the people asking are so ungraceful in accepting no for an answer. But when I didn't have this limit I noticed other things. Like how exhausted and unhappy I was, how I got sick a lot, how it affected my kids and my husband, and how when I stopped traveling I had so much more time to pay real attention to my research and my amazing students.

The author of that post wound up getting tenure at Harvard on schedule, and then getting further promoted to full professor unusually fast.

Anyway, for my part, I have little kids, so traveling is a burden on my family, in addition to sucking up a surprising amount of time / energy (much more than you would think just the nominal travel time, because there’s also booking & planning, packing & unpacking, catching up on chores and sleep and taking notes afterwards, etc.). Big opportunity cost.

I think the idea of having AI safety conferences makes sense, but I think it would be a pretty bad idea for these conferences to be subsidized by industry. Insofar as we want to work with industry on AI safety related stuff, I think there's a lot of other stuff ahead of conferences that both: a) industry would be more excited about subsidizing, and b) I'd worry less about the COI leading to bad effects. (For instance, industry subsidies for mechanistic interpretability research.)

It depends on the venue, but for EAG Boston for example the food and beverage minimum spend is $540k, and we're planning to spend just a bit over this (we may go higher if more attendees end up registering).

I'm not sure how those academic conferences work — it's possible that they take place in venues without in-house caterers or that they pay more up front to avoid a minimum spend. Because we usually get a minimum spend, attendees aren't allowed to bring in their own food unless they have serious allergies (though of course people might still bring in their own food covertly).

Thanks for your thoughts here!

It is true that we haven't advertised our content much and could possibly get more views if we did. We are planning to do more here, but we estimated that we'd have to really get a lot more views for it to be worth it (recording each stage costs ~$50k) and we expect we won't succeed in getting the videos that much exposure.

And we considered your second (and third to some extent) points when making a decision here but still concluded we should cut back at this time.

Re the printing and signage — I'll note that EAG London was something of an outlier here because the venue we used is a Grade I listed building (i.e. a building designated of architectural/historical interest and hence worthy of protection). This meant that we couldn't stick anything on any of the walls and so we had to spend more money on other types of free-standing signage.

Coming from the "Chaos", the recording quality on media.ccc.de for bigger events is often higher than that of EAG(x) recordings. (E.g. small details like normalizing / boosting audio levels.) That makes a large difference in the "watchability" of recordings.

I agree that the talks being recorded has a huge impact on my behavior during events – e.g. at CCC events, if a talk is being recorded I generally don't bother trying to go there in person unless I'm really interested. Watching it later at home is easy / possibly more convenient than at the event (no standing in line, no crowd noise in the room, known good recording quality), while talking to the people at the con can't really happen later. (E.g. at the last Chaos Communication Camp just a few weeks ago, I checked the schedule ahead of time, picked 4 events for "maybe live attendance", then went to none. By my current estimate, I'll probably watch about 20-30% of the recordings over the next weeks / months at least partially – don't have to stay / keep watching if I'm not interested.)

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.

Thanks — and yep we're aware of them! We haven't gained much from their information previously though perhaps could be digging deeper here.

No problem, thanks for your thoughts here! I'll note that in venues where we have a minimum spend, attendees aren't allowed to bring outside food into the venue unless they have serious allergies. So if we did hand out Soylent/snacks somewhere else, they'd have to consume this on the street or something (which may be a disaster if it's raining or cold).

Re reserving nearby restaurants — we did this at EAG Bay Area and plan to do it again in the future (we just reserved some restaurant tables but had attendees pay for their own meals). If we were to actually plan meals/do catering at multiple nearby restaurants, that would likely be a lot of work on our end and we probably wouldn't have capacity to pull it off well.

Renting out nearby restaurants seems like plausibly a good idea, though, a) that might also be quite expensive, so I'm not sure we'd actually save on costs, and b) the logistical overhead on figuring that out could be large.

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).

It seems likely to me that the externality is likely to be larger earlier on in a social movement than as it becomes more established.

Thinking this through for a minute, it seems like the obvious answer would be: let people choose either (say) $500 ticket price or $1,000, and also have a note saying "If your annual income is above XYZ, we would like to ask you to choose $1,000, though this will be done on an honor-code basis. If your income is below XYZ, feel free to choose either option" (or something like that)

And some speakers/orgs might be willing to subsidise av and av editing production costs for their own sessions. It’s something they can link to on their website for promo, onboarding etc.