Bay Solstice 2019 Retrospective

by mingyuan15 min read16th Jan 202036 comments


Secular SolsticePostmortems & Retrospectives
Personal Blog

I was the Creative Director for last year’s Winter Solstice in the Bay Area. I worked with Nat Kozak and Chelsea Voss, who were both focused more on logistics. Chelsea was also the official leader who oversaw both me and Nat and had final say on disputes. (However, I was granted dictatorial control over the Solstice arc and had final say in that arena.) I legit have no idea how any one of us would have pulled this off without the others; love to both of them and also massive respect to Cody Wild, who somehow ran the entire thing herself in 2018.

While I worked with a bunch of other people on Solstice, all opinions expressed here are my own. 


Beginnings & research phase

Chelsea, Nat, and I spent just about a full year preparing for Solstice. Our first meeting as a group was on January 20th, and we received official confirmation that we’d be in charge about a month later. In the following month, the three of us had roughly weekly planning meetings via video chat. In the first meeting, we set goals for what we wanted to have done by the first of April, and after that we checked in regularly for a while.

My goal by the first of April was to have a rough outline of the entire arc and a plan for how to make the tone be coherent. I also wanted to get a better handle on the task, and to that end I had conversations with several previous Solstice organizers and also read as much as I could. This included Ray’s sequence on ritual, most of the writing on the Secular Solstice website (both the blog and the resources section were useful), discussions on the Rational Ritual Facebook group, and whatever else I could find. 

I also spent a full day listening to every single song that had ever been recommended for Solstice (see below) and making notes on them in a spreadsheet. When I ran out of songs in that reference class, I sorted my iTunes library by most listens and started going through to see if any of those songs might fit. This was quite surprisingly fruitful - I actually ended up using three of the songs that were originally put on my list in this way.

Another thing I did was go around and ask people what they wanted out of Solstice. I got responses that were actually fairly useful, like “I lose interest during long speeches” and “there should be more singalongs” and “I like singalongs but a lot of the songs have really complicated tunes and I can’t handle it.”

I think even with all that research and preparation, I still didn’t have a very good sense of the history of Solstice. I had only been to two big Bay Area Solstices, plus a private Solstice in the woods, plus a short, small, off-the-cuff Solstice that Habryka and I ran at my mom’s house in 2018. I wasn’t around for early Solstices, and I’d never seen what they were like in Boston, Seattle, or New York. 

Creating the setlist

The first setlist

I met the April 1st deadline for having an outline of the entire arc. I drew up my first setlist in my notebook on a five-hour plane ride. I was taking into consideration more small snippets of advice than I can list here, but I can quote the guiding goal I referred to throughout the entirety of my time working on this Solstice:

I want my Solstice to be about the brokenness of the world and the ethos of "somebody has to and no one else will", but also the fact that even if you exercise your individual agency to do the most you can, we still might fail. 

Below are some sample pages from my notebook, with most names redacted. Note that I didn't ask any of these people before assigning them to songs; this was just idle speculation on a plane ride about who might be able to pull off each song.

Some sample pages from my notebook

After that was a process of constant iteration. 


We had our first full run-through of the setlist on June 16th. In the two months leading up to this, Chelsea checked in with me weekly. I had a lot of speeches to write and planned to write one and send it to her each week (which I more or less accomplished). I found this accountability mechanism quite helpful.

The first run-through was just me, Habryka, and Chelsea. We sat on the floor of my bedroom, played recordings of the songs to sing along with, and took turns reading the speeches. For this run-through, I wrote three new speeches, re-used two existing Solstice speeches, and threw in Pale Blue Dot, an abridged version of the poem Ring Out, Wild Bells, and the Sequences post On Doing The Impossible. After the run-through, which took about two hours (we timed each piece), the three of us debriefed, and I used the extensive notes I took to make changes to the setlist.

I felt burned out after this and had the luxury of taking all of July off from Solstice work. This was a huge benefit of starting so early.

The next run-through was in September, scheduled around the dates when our featured musician, Taylor (who lives in Vermont) would be in town. This was significantly more involved but still fairly informal. Seven people participated in total (including Ray, who was experimenting with projection), we provided our own instrumentation for most of the songs, and most of the speeches were read by the people who would eventually give them. Afterwards, each person individually gave me feedback, and again, I made significant changes to the setlist in response.

Regular coworking hours

Around the time of the September run-through, things really picked up. Going into crunch time, Chelsea, Nat, and I set up a regular time for Solstice coworking. We met for a couple hours every Monday night - that served both as our designated time to work on Solstice things (since we all also have day jobs) and an opportunity for in-person communication about high-context or sensitive topics.

Choosing performers

Skill level

We put out a call for auditions in August, but we didn’t publicize it very well, so we didn’t receive many applications. I required auditions for speakers and singalong vocalists, but not for instrumentalists - in retrospect, this was an obvious mistake. I let almost all of the instrumentalists who applied participate in as many songs as they wanted, but there were issues with everything from not having time to rehearse to not being able to play in the key the vocalist wanted to disagreements on chords and time signatures. 

Since we didn’t get enough applications to fill all the slots, I reached out to people who I already knew were competent from seeing them perform in previous Solstices or in other contexts (such as at jam sessions or in the REACH Musical Revue). This ended up working quite well - I think that everyone I chose in this way gave a good performance.

NYC has often hired professional musicians, rather than having community members provide the instrumentation - a possibility I knew about but never seriously considered. While I still wouldn’t pay for professional musicians, I’ve come to understand better why someone might decide to do so, after experiencing the difficulties (mentioned above) of getting amateur musicians to produce a performance-ready piece.

On the other hand, a lot of the amateurs did do an excellent job! So I think the takeaway here is just that it’s important to have people audition (or be very sure of their competence level in some other way) and make sure they’re able to put in the time commitment to rehearse.

Orientation towards Solstice as a whole

Something I struggled with a fair amount was disagreements about arc cohesion vs showcasing technical skill. Arc cohesion trumped all other concerns for me, with singalongability a close second. However, it’s often the case that when singers are given the chance to perform, they want to do something more interesting than just lead the melody of a singalong, so I was sometimes at loggerheads with performers who wanted to do more complex pieces or include intricate harmonies. There are some pieces for which I regret not being firmer about putting my foot down on this issue, and I think that ultimately it’s probably reasonable to exclude performers on this basis if you can’t come to an agreement.

On the flip side, I was extremely grateful to the people whose pieces I cut after the dress rehearsal, a week before the performance. I apologized to all of them and gave them the opportunity to contest the decision, but they were all really great about it and said things along the lines of “I care way more about Solstice being better than I do about cashing out on the work I did.” To those people - I really appreciate you, and thank you for being wonderful!

Final preparations

In late November we did a walk-through of the venue (important for testing planetarium footage and lighting options), then in early December we had a tech rehearsal at the planetarium (mostly for A/V) and a dress rehearsal (not at the planetarium). Much of the benefit here was logistical, and I won’t touch on that too much since it’s not my wheelhouse, but it also gave me a chance to see what the final product would be like. I was originally supposed to stop editing the setlist entirely on November 5th, but in reality, I made several changes to the order and even cut some pieces in response to the dress rehearsal, and the exact content of some things was kind of up in the air even on the day of. While I technically missed my deadline or whatever, I don’t regret that at all - I think my Solstice was quite significantly better than it would have been if I had stopped iterating on the setlist a month earlier.

Considerations for creating a setlist

Creating a cohesive message

When it came to speeches, I took an arc-first approach. I decided what message fit in every place in the arc and found a piece that fit there. Only after that did I approach potential speakers. I ended up using four existing pieces wholesale; the rest were based on existing pieces but were either remixed or heavily edited by the speaker. The takeaway action item here is to have something very specific in mind for each speech and conveying that to the speaker. This improves cohesion while still allowing each speaker to put their own twist on the piece if they so choose.


  • Tessa remixed Ray’s A Bottomless Pit of Suffering to make more sense in Berkeley (it was originally given in NYC, where it’s very cold) and to have more of her own voice.
  • I had Nate Soares on the docket from quite early on in the process, but he’s busy executive-directing MIRI so I didn’t want to take up too much of his time. With his input, I chose an old post from his blog for him to read (How We Will Be Measured), and Chelsea and I edited it to be shorter and more appropriate as a speech. Nate then made a bunch of his own changes to reflect the ways his thinking has changed over the past few years, but ultimately it was still recognizable as the same piece.
  • Though it ended up being cut at the very last minute, Alex Altair adapted a scene from HPMoR into a speech, which was pretty cool.

Creating a smooth arc

In order to create a smooth experience, you have to make sure that there are smooth transitions  between the message, emotional tone, and musical/artistic feel of each piece and the next. It turns out to be really hard to match these all up. For example, there are some funny/upbeat/light-hearted songs about death (e.g. We Will All Go Together When We Go), and some fairly serious-sounding songs about more light-hearted topics (e.g. Time Wrote the Rocks). Some songs are up-tempo, some are slow and mournful, some have percussion, some are performed by choir. There are just a ton of considerations. (This is why Ray writes so many of his own songs - that’s the only way you can really have control over the message, tone, and feel all at once.)

I was aware of all of these considerations, and that’s a big part of the reason that I made sure to run through each version of the setlist from start to finish, but I don’t think I quite got it right until the final performance (even the dress rehearsal, one week prior, was fairly rocky in this regard). And even then there were still a few problems, like the energy drop from Son of Man to Uplift and the energy drop from Singularity to Five Thousand Years (making Five Thousand Years a rather anticlimactic finale).


The previous Solstices I had attended were just a series of pieces strung together, and the audience were mostly left to discern the arc on their own. At some point in my research, I saw a video clip of Kenzi MCing a Solstice, and I immediately decided I wanted to do that. 

In my opinion, there are a ton of advantages to having an MC. Here are a couple:

  1. It gives the audience more insight into why each piece was chosen and generally gives you a chance to tie the arc together more explicitly.
  2. It allows you to make announcements during the program without breaking immersion, such as giving trigger warnings or asking the audience to hold their applause until further notice.
  3. It fills what would otherwise be awkward pauses between pieces as performers get on and off the stage.

Post-performance feedback

About 50 people filled out the feedback survey, and their feedback falls into a few rough categories for me:

  1. “This person is right and I would change/tweak this if I had it to do over”
  2. “This feedback should not be acted upon”
  3. “Feedback on this is very split, and you can’t win ‘em all”

Things I would tweak


The peak-end rule is really important, and my Solstice didn’t have a very strong ending. I had to go up onstage and be like, “Okay, now it’s over.” If I had it to do over, I might cut Five Thousand Years and end on Singularity, which everyone loved.

Moment of darkness

The moment of darkness itself (two minutes of silence in the pitch black) got mixed reviews - some people found it very powerful, some people found it existentially horrifying and had to distract themselves, and a lot of people found it didn’t really land. The main thing I would change here is the way I introduced it.

I said, “We’re about to sit in silence for two minutes. If you’re up for it, I want you to look up at the stars, and think of someone you’ve lost. Someone whose voice you will never hear again, whose mind is gone from the world forever. Give them your grief, yes, but also give them your resolve.”

In retrospect, this was far too specific an ask. A lot of people said that the moment of darkness didn’t really land specifically because they’d never lost anyone close to them. (I copied the text from the Solstice Habryka and I ran in Madison, where it worked very well, but where the circumstances were very different in quite a number of significant ways.) If I had it to do over, I would encourage people to sit with their feelings, wherever they were at, rather than prescribing something for them to think about. 

First speech

I struggled for a long time to find or write an appropriate speech for the first-speech slot in the program. It was only a day or so before the dress rehearsal that I settled on giving an abridged version of Nate’s This Is A Dawn. While it had roughly the right message, I don’t think I myself was that bought into it, and as a result, people seemed to find it a bit generic, and not really meaningful. I’m not actually sure what in particular I would do here if I had a do-over, but I do want to highlight that the first-speech slot is quite important and I definitely didn’t totally nail it.

Feedback that should not be acted upon

A hopefully uncontroversial example of this is the person who doesn’t like the sound of strummed instruments, and therefore gave a low rating to every song with strummed guitars. Sure, this is a valid way to feel, but at the same time, one person’s preference in this area does not mean it makes sense to cut all guitars from Solstice.

A more controversial example, but one that I am still willing to stand by publicly, is the common complaint that Son of Man is sexist. Look, I’m a woman. Chelsea is a woman. The person who soloed on Son of Man is a woman. My sense is that, while some people were genuinely offended, and that came through in their feedback, most of the people who registered complaints were just people who were worried that other people might have been offended. I continue to think this song is an excellent fit for Solstice message-wise and has great energy (it was intended to be performed a bit faster but there were some technical difficulties with the drums). I would not hesitate to include it again.

Triggering content

Eating disorder trigger

For complicated reasons, there was a brief discussion of weight loss at the end of Solstice. It was intended as a sort of light-hearted post-credits piece, but we mishandled it, and people didn’t end up getting the chance to leave if the topic was difficult for them. This had significant negative consequences for some people, and I sincerely apologize for that.

We’re taking steps to make sure that future Bay Solstices are more careful around sensitive topics like this. Specific action items include providing verbal trigger warnings in addition to the ones written in the program, and allowing significantly more time for people to leave if they need to, including having some people planted in the audience to stand up and leave so that it feels socially okay to do so. (Even though I myself won’t be running Solstice next year, I’m in close contact with next year’s organizers and have made at least one of them aware of this.)

Other triggers

On the feedback form, some people mentioned being very upset by Solstice because it reminded them that they were lonely or felt like they could be accomplishing more. I do not think anything should change about Solstice itself in response to this feedback, because being reminded that the universe is vast and dark and cold is pretty much the entire point of Solstice. 

Perhaps in the future it would be good to make it clear to potential Solstice-goers that Solstice deals a lot with death, individual responsibility, and the vast, uncaring universe. Then they can make more of an informed choice about whether or not to go, and if they can’t handle it, they can’t reasonably blame it on anyone but themselves.

Individual pieces

Raw feedback

(The table below has the pieces in chronological order of how they appeared in the performance.)

1The Circle1361221
2The X days of X-Risk3471614
3To Drive the Cold Winter Away01141611
4Bold Orion0142016
5Time Wrote the Rocks0610179
6This is a Dawn (abridged)0012195
7Hard Times Come Again No More0281812
8There Will Come Soft Rain3612119
9Pale Blue Dot0031225
11Do You Realize1461614
12A Bottomless Pit of Suffering1171514
13Bitter Wind Lullaby03111610
15The Moment of Darkness1151517
16Spoken Call/Response0211187
17We Are the Light2113165
18Light Pollution0112193
19Endless Lights0311149
20Brighter Than Today0121618
21How We Will Be Measured0412158
22Son of Man3610156
24What it means to win126205
25Beautiful Tomorrow1371612
27Five Thousand Years2471710

At a high level, most people liked most things! This is heartening. 

Ray and I sorted the pieces four different ways*, and there were five pieces that clearly came out on top and five that (only a little bit less) clearly came out at the bottom.


  1. Eulogy
  2. Singularity
  3. Pale Blue Dot
  4. Brighter Than Today
  5. Bold Orion


  1. After-the-Credits Eliezer bit
  2. Son of Man
  3. There Will Come Soft Rain
  4. We Are the Light
  5. Light Pollution

Effect of delivery

Something I notice that’s interesting (but not that surprising) is the large effect that delivery had on people’s ratings. For example, two of the highest rated pieces were Singularity and Pale Blue Dot. In addition to a solid delivery by Chelsea, Pale Blue Dot had a backing track and custom planetarium footage. Singularity was extremely energetic and fun, and people had generally positive affect towards all of the songs that prominently featured Taylor because he’s such an obviously skilled musician. Brighter Than Today and Bold Orion were also energetic and very polished performances.

By contrast, people were relatively lukewarm on Endless Lights and Bitter Wind Lullaby, two Solstice staples that I think of as being fairly well-liked in general. Both of these songs had significant problems with their execution, with the performers having trouble agreeing on the time signature. As a result, it was difficult for people to sing along, which seems to have made for a negative overall impression.

Addressing the elephant in the room

An additional thing that people who attended this Solstice might want to see addressed is what the heck was up with the Eliezer piece. Even apart from those who found it triggering or otherwise inappropriate, a lot of people were just confused about why it happened (e.g., several people's reaction was, "Why is this guy talking to me like I'm his friend, I don't even know him"). The explanation is perhaps not all that satisfying, but I'll give it anyway.

In early October, Eliezer contacted me asking if he could do a shenanigan at Solstice. He explained his idea to me, and while I didn't really see how it would fit in, I also didn't want to reject him out of hand.

I talked to a couple people I trusted about this, and we came to the conclusion that it would be pretty valuable to have Eliezer onstage. The reasons for this were a bit nebulous, but roughly rested on the following:

  • Regardless of any single community member's personal feelings on him and his writings, it's hard to deny that this community would not exist as it does today without Eliezer. (I, for example, came in through a chance encounter with HPMOR in high school, and basically every aspect of my current life is a direct result of that encounter.)
  • Eliezer has increasingly retreated from public life over the past few years, and this has resulted in some feelings of abandonment on the part of the community.
  • Having Eliezer onstage during Solstice would show his implicit support for the community and the event; following the above, it would remind the audience of what brought us all together and that we haven't been abandoned by our founders.

Based on this reasoning, it was having Eliezer onstage that mattered, and the content of his piece wasn't really relevant. The eating disorder trigger was honestly not something I even considered until someone mentioned it after the dress rehearsal. It was at that point that I decided to move the piece to be 'post-credits' (it had previously been early in the program proper), to make it opt-out for people uncomfortable with the topic, but as mentioned above, I failed to handle this correctly.

It’s also worth noting that, while more people hated the Eliezer bit than hated any other piece, there were also a fair number of people who loved it (if you sort by the raw number of Loves, it comes dead middle). So it was in fact not universally reviled (lots of people found it hilarious or heartwarming); it was just very polarizing.

Summary of takeaways

This is just all of the takeaways from the main body of this post, in the order that they appeared.

  • Starting a year in advance and testing and iterating often makes for a really good final product but also burns you out like hell. I think this was ultimately definitely worth it, but if I was told I had to do another year of this I would probably flee the country.
  • Deciding on a central theme/thesis for your Solstice early on is really important.
  • Set a regular time to work on Solstice things so that they don’t slip through the cracks, especially if you have a full-time job. It’s best if you can meet with other people regularly for this purpose, because accountability.
  • While hiring professional musicians may be easier, there are enough skilled musicians in the Bay Area rationalist community that I think it’s worthwhile to go that route, especially since this makes it feel more like a community event. Just make sure that people audition (or are known to be skilled and easy to work with) and can commit to rehearsing with each other.
  • Choose speakers and other performers largely based on their skill level, but it’s also important to make sure that they’re value-aligned with you when it comes to the Solstice you’re creating together.
  • It’s okay to iterate on the content until the very last minute so long as everyone is on the same page / no one is thrown off or blind-sided by the late changes.
  • If you want your arc to be really cohesive, you need to exert centralized control over each piece rather than just leaving performers to do their own thing.
  • It’s really hard to create a smooth arc over all the dimensions that matter. If you can write your own songs or work with a friend who can write original songs, this is a huge asset.
  • MCing is great.
  • Not all feedback should be acted upon.
  • Pay attention to the peak-end rule.
  • Potentially triggering topics should be handled more carefully than they were by me. It’s important for people to have a genuine opportunity to make an informed choice about what they’re exposed to.
  • Delivery/execution of pieces is just as important as (if not more important than) the semantic content and the fit in the arc.


Setlist spreadsheet

I love spreadsheets with a passion, and I found keeping all of the relevant material in one place to be enormously helpful both for me and for communication purposes. (Whenever someone had a question about the arc, the performers, or anything, we could just pull up the spreadsheet, and even make a copy of it to see how changing the order of the pieces would feel.)

Here is a template for the spreadsheet I used. Let me know if anything is unclear!

Masterlist of Solstice materials

Daniel Speyer runs the Secular Solstice GitHub page, which is a useful resource, but it’s also very hard to edit - especially if, like me, you’re not a programmer and don’t know how to use GitHub in general. The Giant Epic Rationalist Solstice Filk spreadsheet is likewise a useful resource, but it’s kind of a mess. So I made my own spreadsheet, which is publicly editable and incorporates every song, poem, story, and speech from the above two repositories. (Apologies to Daniel Speyer and to anyone who sees this as polluting the commons by instantiating too many competing projects.)

* The sorting algorithms we used were the following:

  1. % Positive : (Liked + Loved) / (Total responses)
  2. Overall-Liked : (Liked + Loved) – (Disliked + Hated)
  3. Weighted : (2.25*Loved + Liked) – (2.25*Hated + Disliked)
  4. Loved : Raw number of ‘Love’s

Thanks to Ray Arnold, Nat Kozak, and Chelsea Voss for their input and edits.