2025 Feb 28. Put it in your calendar.
It's been three years since HYTRADBOI. Long enough that I've mostly forgotten how stressful it was to run a conference and it's starting to seem like a good idea again.
talks
The format will stay the same. 10 minute, pre-recorded, heavily-edited talks. Asynchronous chat. Join from a different time-zone, watch talks on your lunch break, answer questions in your pajamas, pause the talk to go feed your baby. Conferences work better async.
The original conference was sparked by frustration that database and programming language folks live in two totally separate worlds, but the actual talk selection ended up being almost entirely database people because that's who I know. This time I'm adding an explicit track for programming languages, interpreters, compilers etc and I'll be reaching out to a lot of PL folks.
Last time most of the talks were about the speakers current project. This time I'll still accept some project/idea talks, but I also want to make room for more education talks eg:
- Distillation: Share the hard-learned knowledge that bridges the gap between textbooks and cutting-edge practice. Pay down research debt.
- Survival guides: 'As a specialist in X, here are the most important things that generalists should know.'
- Unexplanations: X is widely misunderstood, here is the correct version.
The implicit contract for most industrial conferences is that speakers fill the schedule in exchange for getting to promote themselves or their current project. This is fine if you're a cto and you just want to know what vendors are available but the more important role of conferences is to spread knowledge, and I don't think product sales pitches are an effective way to do that. But I am keenly aware that asking speakers to do more work in exchange for less immediate reward is a hard sell. (I do think that there is a lot of career value in giving a talk that becomes a standard reference (eg), but it's definitely a lot more work.)
So I'm open to recommendations for speakers, but I'll mainly rely on reaching out to individual people who I think are directly motivated by teaching and explaining. I'll also work with the speakers much closer, going through at least one draft and helping with the editing.
To balance that out I'll also hold totally open submissions for 3 minute lightning talks. I'll accept any video that arrives before the deadline, is vaguely on-topic, and meets the format requirements (length, video encoding, correct subs etc). It'll be a buffet.
infrastructure
Last time the infrastructure was... not perfect. That's the main thing that has put me off running a sequel. But I think I can avoid much of the pain.
Last time I had a range of ticket prices. The goal was that people making six-figure salaries or who can expense their employer would pay full price, but everyone else would still be able to attend. Eg last time I had a fair few attendees with email addresses from indian universities. I searched for phd stipends at those universities and got a lot of numbers around $400 / month, which I imagine makes a $64 ticket a pretty gnarly expense.
I couldn't find any service that would both handle ticket sales and also connect ticket-holders to a reasonable chat app (ie threads, code blocks). So I rolled my own and it was a disaster - I really underestimated how hard it is to reliably email invites to people, even when using a commercial email service. I spent most of the conference manually resending chat invites.
The plan this time is to just have two ticket prices:
- Regular - $64
- Discounted - $0
Really, this is just asking for donations, but in a way that makes it easy for people to expense their employers.
Both kinds of ticket will immediately take the ticket buyer to a registration page for the chat, and also send an email invite as backup. Crucially, having a $0 option means that I don't need to connect ticket sales to chat registrations - I can just use an open registration link. That immediately deletes half of the infrastructure failure I dealt with last time.
The other half of infrastructure failure was the chat server falling over. I self-hosted a matrix server because I wanted to enable some features that weren't available on EMS. But I failed to increase the default max tcp connections of the proxy 512, and also failed to load test with enough clients (in my defense, I load-tested for 5x the number of attendees I expected). This time I'll just use a hosted chat server.
A lot of people also struggled with the UX on matrix (eg constant warnings about encryption keys). I'll use something else this time. I have a soft spot for zulip, which makes it very easy to follow multiple threads, but I'll also consider discord for the familiarity.
Last time I hosted videos on bunny which is very cheap and easy to operate. Unfortunately they also sometimes serve spurious 404s which prevented some attendees from watching the talks. I'll still use bunny for hytradboi.com because it's the only affordable option if something ends up on hacker news, but for the conference itself I'll consider one of the more expensive options. Recommendations welcome.
I used rev for captions. The quality wasn't nearly good enough and I ended having to hand-edit txt files for 34 talks. This time I'll probably use descript for the entire editing process. My wife has been using it for editing her podcast and I've been impressed by the UX. I'll probably still have to manually correct the captions but the workflow will be much nicer.
contact
If you want updates on tickets / speakers / lightning talk submissions etc:
- Subscribe to the atom feed or email list for this site (~monthly updates).
- Follow hytradboi on mastodon or twitter.
I'll announce details for lightning talks in a few months.
If you have recommendations for speakers, or suggestions for infrastructure options, email info@hytradboi.com.