(updated)
|
min. read

Local-First Conf 2025 Reflections

Conrad Hofmeyr

Last week, I attended Local-First Conf in Berlin with 3 other members of the PowerSync team. We all had a blast. Here are some quick thoughts on the conference:

A unique gathering of like-minded people

Local-First Conf really stands out amongst the conferences I’ve attended over the years:

  • The density of talent is very high. The conference brings together a ton of very impressive developers and entrepreneurs.
  • At many other conferences, there are plenty of people who don’t seem particularly invested. Yet at Local-First Conf, it felt like everyone was passionate and invested in the common denominators of local-first software. You could almost feel the excitement in the air, and it felt like everyone I spoke to was brimming with enthusiasm and ideas.

And yet it was a “big tent”

The attendees were like-minded while also being diverse. The idea of “local-first” brings together:

  • Academic researchers and industry
  • Big companies, small companies and individuals 
  • Pre-traction vs post-traction outfits
  • AI skeptics and AI accelerationists
  • Local-first idealists and SaaS-minded utilitarians

This made the event especially interesting, and led to interesting debates. I enjoyed the panel discussion on “The big questions of Local First”. I think Aaron Boodman is a great spokesperson for the pragmatic view on sync engines as tools to improve user experience in web apps — as opposed to the broader ideals of local-first software around user ownership and longevity. I also applaud Aaron for taking a principled stance on excluding offline writes from Zero for now. Aaron argued that Zero should rather not have support for offline writes than offer a half-baked solution — despite this being met with surprise and generating some controversy. I commend their focus.

“What is local-first and what is not local-first?” has been a hotly-debated topic over the last couple of years, and I think the community has made some progress on coming to terms on this. The original local-first idealists have created a community of people that is very welcoming to divergent ideas, while the same organizing committee behind Local-First Conf is now also launching SyncConf, which will focus mainly on the idea of using sync engines to build better software. People who feel strongly about the original local-first ideals will find a home in Local-First Conf, whereas people who are interested in building better SaaS & PaaS using sync engines will gravitate towards SyncConf. I suspect many folks (including me) will want to attend both.

The tools are maturing

It’s exciting to see that the local-first space is heating up. The tools in the local-first landscape are markedly maturing compared to last year. PowerSync has expanded its capabilities significantly. Zero has evolved and tuned their ZVM system for incrementally materialized views for high performance. ElectricSQL has done great work on creating an incremental adoption path for sync with its TanStack DB initiative and partnership. Jazz is making great progress. 

I was very pleased to speak to lots of people at the conference who are already using PowerSync or have come across it and want to try it out. A few of them were surprised that someone from PowerSync was not speaking at the conference, and they asked us why this is the case. Unfortunately, I don’t know why. We tried to get a chance to speak, to no avail. I can understand why it seemed odd to folks that every significant player in the local-first landscape was speaking at the conference except for PowerSync. Fortunately though, I had a chance to talk to organizer James Arthur, who told me that this will be corrected at SyncConf, and hopefully at future installments of Local-First Conf too.

Research is progressing

What’s also awesome is that first-principles research on local-first concepts continues to make great progress. There were great talks from Ink & Switch folks, including on Keyhive and Beelay. I also really enjoyed Matthew Weidner’s talk on collaborative text editing without CRDTs or OT (he also wrote a post about it).

We are all on the same team

I’d like to echo what my colleague Kobie said: I’d love to see further collaboration amongst the rivals in the local-first tools landscape. Kyle Mathews of ElectricSQL told me that we should work together to grow sync and local-first into a larger movement, since it will benefit all of us. I agree wholeheartedly, and therefore I’m excited to support SyncConf alongside the ElectricSQL folks.

There’s a lot that we can learn from each other. I appreciated some of the wisdom shared at the conference. One of my favorite quotes from the talks is from Anselm Eickoff of Jazz: “With users you’re lucky if you see them fail once. With AI you can watch it fail repeatedly.” That cuts deep!

I would like to give huge kudos to the folks at Ink & Switch for starting this amazing movement, and to the organizers of Local-First Conf. Together, they have created something truly unique and brought together a fantastic group of people. I’m excited to remain a part of this collective and to contribute to growing it.