On Sunday I fly out of Denver to attend local-first conf happening next week. I‘ve attended a ton of traditional tech conferences like MWC, and this is the first conference I’m genuinely looking forward to attending.

Recap of 2024

The inaugural Local-first conf was in 2024, and there were some pretty big announcements:

• Aaron Boodman from Replicache announced Zero
• Ralf from our team announced PowerSync’s Open Edition
• Anselm Eickhoff announced Jazz

It’s awesome to see so many people working really hard to fundamentally change the way that apps manage state, and just as cool to see most (all?) of the vendors adopt some form of open source licensing model. Last year we also saw continued efforts to firm up what the term "local-first" actually means and what it means for developers. Unfortunately, it looks like there will once again be much hand wringing over the term. Personally I find this conversation quite distracting… here's my mental model:

1. The term has already been clearly defined by the inventors, let’s try to stick to the original definition.
2. IRL however, local-first is used to describe a broader range of technologies where state is synced to the device rather than loaded into ephemeral memory. Triplit wrote about this, but unfortunately try to redefine the original term which I think just causes more confusion.
3. "Sync engine" is a much more useful umbrella term than “local-first”. Vendors not adhering to the strict definition should try to stick with that or something like it
4. Don’t be a pundit on a crusade when someone uses the term incorrectly. Definitely don’t be toxic about it.
5. While “sync engine” is a better general purpose term than “local-first”, an unsolved problem is that "local-first app" rolls of the tongue whereas "an app built with a sync engine” is a bit dry.

Sorry about the distraction.

Progress since 2024

It has been a super busy year since the previous conference. At PowerSync we’ve seen very solid growth in user numbers, and I’m sure other projects have seen the same. This is fantastic! Some notable highlights over the past year:

• ElectricSQL shelved their legacy product and pivoted into a minimally invasive and highly pluggable architecture
• Zero delivered with a highly successful launch and have done wonders to increase developer awareness of local-first
• PowerSync helped launch the weekly open-contribution Local-First News digest which, at the time of writing, has just over 400 subscribers
• PowerSync also brought local-first to MongoDB and MySQL users on the server-side, and React Native Web, Swift, Kotlin/JVM and Node.js. We also added broad support SQLCipher across many of our SDKs for encryption at rest on the client, and we added OPFS support for our web SDK making multi-tab sync on Sadfari a good experience.
• The local-first landscape initiative was launched to compare all the emergent tools in the ecosystem

What to expect at 2025

It looks like the usual gang of speakers again this year with some exciting showcases such as Keyhive, an access control project for secure & decentralized collaboration. Also Patchwork: "version control for everything”. I’m very psyched for this one, since version control of spreadsheets is a big pet peeve. I also have three personal objectives:

1. Meet other members of the local-first community
2. Hang out with my fellow PowerSync team members (hi Simon, Benita and Conrad!)
3. Test this mental model I recently came up with:

I haven’t included vendors like automerge and convex on my sync engine map, because there are two hills I’ll die one:

1. Most P2P implementations don’t provide network layer privacy and peers disclose their IP addresses to each other.
2. Bring-your-own-database is a core architectural principle for me. Hosted data is especially vulnerable to rug pulls.

Local-fist 👊

If you're at the conference hit me up on the socials - DMs open