Critiques of conferences | Rohan's Rambles

TL;DR: If you’re going to a conference and don’t have a very clear picture of what it going well looks like, I think you’ll probably be wasting your time. In general, the aims of conference organisers are not well correlated with your own aims, and in any case running a conference well seems difficult. Going to a conference for the ideas at talks/presentations probably isn’t a great motivation; going for the people is more sensible, provided you actually arrange to meet others.

I’ve been to six conferences over the last couple of years and their quality & atmosphere varied a lot. The first three were networking events for people involved in Effective Altruism (EA); one was about British politics and policy (“CF”); another focussed on tech & AI (“CX”); the most recent was an academic workshop-style programme for economics undergraduates. Out of all of them, I found the EA conferences the most worthwhile – it wasn’t that I preferred the content under discussion or sorts of people in attendance, but rather that the organisers had structured things in a way which made the most of the fact that they had gathered many keen Effective Altruists in one place. Conferences are a coordination mechanism: they bring hundreds or even thousands of individuals who’re interested in similar things together. This is a really valuable service! Rarely do you have the chance to talk at length to lots of other knowledgeable people outside your existing network about something you’re working on or interested in. The problem, however, is that (in my experience) many conferences fail to facilitate these kinds of conversations which they’re so well-suited to enabling.

Your goals are not their goals

In part, this is because what organisers want is not necessarily what is most useful for participants. Take the CF conference

Technically, a summit; I’m not quite sure of the difference.

, which had a little over a hundred participants, almost all invited directly (you could apply to go, which I did, but very few had got in via that route). It seemed to me that the main goal of the conference was to promote the newish organisation that was running it: their CEO had invited lots of her political former colleagues and journalist friends, their press team was very eagerly retweeting any mentions of it (& taking plenty of photos for the website), they emailed everyone with a favourable FT write-up of the event a couple of days later. For this marketing effort to be successful, they needed the great and good of Westminster

I wonder whether organisers of events like these use something like the party-organising tactic described in Scott Alexander’s Bay Area House Party sketch: “What you do is - you come up with your ideal guest list - who would you invite if you knew they were going to say yes. Actors, billionaires, all the coolest people in your social circle. Then you send them all an email saying - hey, Elon Musk is going to be at this party, you want to come? Of course they all say yes. Then, a few days before the party, you send out an email - sorry, Elon has to cancel, but we’ve still got [list of actors, billionaires, and all the coolest people in your social circle].”

– some MPs, ex-special advisors, think tank researchers, ideologically aligned senior civil servants – to attend, see how many others from those circles were also present, and leave with a positive impression of the org.

What I found strange about the CF conference, though, was the extent to which it infantilised all those Important Senior People in attendance. Working as a

They reimbursed me for all my expenses, provided accommodation, and even paid me a token amount, all for a total of maybe one hour’s work – so a pretty good deal for me.

at the conference, my main duty besides handing out name badges was to interrupt people having conversations in the foyer and herd them back into the lecture hall. As if this wasn’t school-like enough, I was handed a large bell to obnoxiously ring beside any groups defying the instruction to return to lessons from playtime. To me, this was absolutely crazy. If the organisers’ goal really was to help generate solutions to Britain’s economic woe, then surely bubbly, excited conversations between attendees were exactly the sort of thing they ought to be cultivating, not smothering? And even if they only wanted to promote their brand and the speakers they’d booked, would it have hurt to be a bit less pushy and controlling about it? The talks were vaguely interesting, but I don’t think I learned a huge amount from them – certainly not more than I would have done by reading blogposts for the same length of time. The fact that attendees spent most of their time flicking through emails rather than listening (let alone taking notes!) suggests that I’m not alone. When everyone at the conference is roughly as knowledgeable as each other (along different axes), having one person present their ideas to a group of 100 is a significantly worse way to circulate ideas and insights than just letting people mingle and chat freely. Yes, there’s a place for talks in a conference schedule: an opening speech to set the atmosphere, for example. But packing the programme with presentations back-to-back turns the conference into a huge missed opportunity. There’s no point having participants all together in the same place only for them to sit silently watching a lecture that just as well could’ve been a YouTube video.

Besides, the talk format almost inevitably comes with questions at the end, which tend to be an occasion for certain audience members to deliberately lose sight of the fact that a question is not the same as a testimony, and that prefacing one’s “question” with a lengthy biography of oneself is unlikely to be of great interest to the rest of the audience. I’m being facetious here – questions aren’t always bad, and things are better when technologies like Slido are used to allow the hosts to screen submissions for brevity and relevance. With keynote speakers whose opinions are of broad interest or importance, this talk-plus-curated-Q&A format can be a good approach. But, as demonstrated by the inconsiderately self-promoting question-askers, attendees will also want the chance to talk to and network with each other. Yet much of the time, conference organisers fail to provide the physical or social infrastructure

For example, setting aside spaces for 1:1 meetings, or using their coordination abilities to arrange topic-specific meetups.

to enable this, perhaps because they’re not really trying to make the event as useful as possible for attendees.

Who’s paying?

If the CF conference was aimed at generating positive publicity, what about the CX one? At a basic level, money – it was a profit-making venture. I don’t think their customers were the attendees, though. It’s telling that right at the top of the homepage for the GitHub Universe conference there’s a link to a prewritten email designed to “convince your boss” to fund the tickets from their career development budget. At tech conferences in particular, firms pay for their staff to attend, which in turn makes sponsoring the event an attractive proposition for others. This is great for the organisers, but means that their objectives are not entirely aligned with yours as a participant. Sure, you got a free ticket

Interestingly, CX had a practically unlimited number of free VIP passes they were sharing with panellists to give away to acquainatances, which again supports the hypothesis that they’re not putting together a package they expect attendees to be willing to fork out for.

, but it’s worth thinking about whether you’re the product being sold rather than the user buying a service.

Taking their profit-seeking motive as given, I still think there were several editorial and operational mistakes made by the CX organisers. First, and most unfortunately, they managed to bag a huge number of impressive speakers and yet have resoundingly mediocre discussions and fireside chats. Some of this was in the construction of the panels: a collection of clever people does not an insightful conversation make, especially when their interests and expertise are in completely separate areas. Even when the panel’s membership was more thematically coherent, having an ineffective MC running things often led to bland or rambling exchanges. As the producers and presenters of broadcasts & podcasts know, making a live factual programme takes skill: simply putting four people onstage and hoping that a facilitator’s false excitement and scripted interludes will be enough to prompt an interesting discussion is a bit like zapping some dilute primordial soup and expecting to create intelligent life.

It seemed like the CX attendees knew or realised this – the 10,000+ capacity hall that the organisers had booked out for the keynotes frequently had no more than 300 people inside, many of them focussed intently on taking photos or videos (presumably to prove to others that they had been to a “truly eye-opening discussion”). The adjacent (smaller) venue was busier and had a stronger sense of purpose – the business fair was popular, people were chatting over lunch outside, founders looking for funding delivered lightning pitches. It felt significantly more alive than the arena, so it was a shame that the venue was located several minutes’ walk away, and through a nearly half-hour wait for security checks. Like at CF, attendees’ revealed preference was that they wanted to spend time talking to other participants, not be crammed in to dimly-lit lecture halls as the organisers seemed to hope or expect.

Prestige cartels

Under this traditional conference model, speakers are essential – but if it’s not always obvious what attendees get out of the event, it feels like even more of a puzzle why the keynote contributors show up. Money no doubt plays a part, but that’s not why somebody becomes a public intellectual. There’s some kind of symbiosis between conference organisers and their speakers, in that both source prestige from each other’s status, a little like how PageRank works. Stephen Pinker doesn’t want to be seen at a low-status conference – its scrappiness might rub off onto him – but he’d be glad to go to one with other speakers at or above Pinker-level prominence, to affirm his continued relevance to the intellectual sphere. Strikingly, at both CF and CX the speakers knew each other socially as well as professionally

At CF, the attendees were drawn from roughly the same distribution as the speakers, so this applies to them too.

. Conferences like these are an opportunity for them to mingle in the circles they want to be in. It amused me to discover at CX that, although the big-name speakers left the conference venue practically as soon as they could after their session, several of them hung around the area for the rest of the day to have coffees and dinners with each other. There’s of course another attraction of conferences for public intellectuals: a forum for promoting not only their brand but also their ideas. In some sense, influenceable attendees are a draw for speakers in the same way that inspiring speakers are a draw for the attendees. Again, there’s something in the idea that conference attendees often pay quite a lot for the privilege (in time if not money), and don’t get a whole lot in exchange besides fodder for a LinkedIn post.¹

EA’s advantages

EA conferences have a very different structure which, I think, makes them much better for attendees. The main difference is that they’re centred around 1:1 half-hour conversations – although there are also talks, office hours and meetups, pre-event communications very heavily emphasise 1:1s. This means that there’s no shepherding of people into talks they don’t want to be at, or an undersupply of picnic benches to chat at, because the organisers arrange the event with 1:1s in mind. Requiring people to apply to attend the conference probably helps too

It also contributes to some undesirable in-group signalling and status-seeking dynamics, which aren’t my focus here but ought to be acknowledged anyway.

. The slight moat to enter increases inter-attendee trust and makes people more willing to offer advice or introductions, as well as possibly meaning that the organisers try harder to provide the atmosphere and infrastructure which is most useful for attendees.

Although you can’t really change the culture of a conference as an individual, I think it’s feasible to make them work better for you. One of the other things I like about EA conferences is that the application form almost always includes questions like “What do you want to get out of the conference?” or “Imagine it’s the day after the conference and it’s gone really well – what happened?” Forcing yourself to be explicit about your aims

for the conference before even signing up is a great way to avoid frustration when you get there. If your imagined best-case scenario feels meh or is obviously unachievable, then reevaluate whether you actually want to go. Conferences take up time! They should probably be bringing you something more valuable than photo opportunities and a space to practice small talk.

I remember reading on Twitter somewhere (though I can’t find the link now) that academic conferences are useful for attendees because it gives them something to put onto their CVs (which typically stretch to pages and pages), so there’s that too. Incidentally, the economics workshop this year was by a long way the best out of the non-EA conferences I’ve been to, because the organisers actually did set aside time for us to socialise and get to know the other people there (though I think they could’ve moved the balance towards interactive sessions even further – plenty of people lost interest during the mornings and afternoons of classes). ↩︎