Pinned
My overriding thought going into 2026:
Problems you thought were out of reach no longer are.
Pick a problem 100x harder than you thought possible to tackle alone, and run with it.
Start now. AI will catch up and make it possible.
Sam Willis
3,480 posts
Sam Willis
@samwillis
Building Agents & composable sync primitives @ElectricSQL.
Sync Engines, Durable Streams & AI.
Started PGlite.dev, co-creator of @tan_stack DB.
- I'm absolutely blown away by the response to PGlite, 1k GitHub stars in under three days!
- PGlite (WASM Postgres) now gets over 1 million downloads a week! š¤Æš Absolutely incredible to see how people have embraced what we have built. Awesome work by the team!
- While you're all busy putting SQLite on the server, here I am putting Postgres in the client. Big update coming this week š
- With PGlite + @ElectricSQL sync you can build an issue tracker handling 100k issues + 500k comments, synced from Postgres to PGlite in the browser! Super fast UI updates with all writes going though the local database, and Postgres FTS in PGlite for search š„
00:00 - Perfect timing! We hit 5k stars on GitHub just as we launch the massing PGlite update šš
- Next time someone says Postgres doesn't scale... ChatGPT runs on Postgres with just a single master š¤Æ There was a great talk at PGconf how they have scaled one of the larges apps in the word - link in š§µ.
- Introducing PGlite - WASM Postgres running in the browser, Bun and Node! Working on this has been a lot of fun over the last few weeks. Unlike previous WASM Postgres projects this one isn't using a WASM Linux VM, it's pure Postgres š„³Hey @jarredsumner & @nikitabase, got it working! PGlite, WASM Postgres running in the browser, Bun and Node. Only 3.7mb gzipped. š¤Æ In-memory or persisted to the filesystem with Bun/Node and IndexedDB in the browser. š github.com/electric-sql/pā¦ x.com/jarredsumner/sā¦
- Just download them all... 2.6k cites in the US, ~78kb json (23kb gzip) in your CDN. Even all 19k "places" in the US is only ~150kb gziped. With proper cache headers this is in the browser cache next time.Simple interaction: one select changes the options of a second with a remote data source How would you build it? How do you deal with race conditions?
- PGlite, our Postgres in WASM project, has hit 1/2 million downloads a week! š¤Æ We (@ElectricSQL) have a lot planned for (native... š) PGlite this year, exciting times ahead!
- Exciting day, @ElectricSQL is now in Beta š Electric makes realtime replication between Postgres & *any* client side store possible! (And we've scaled it to 1 million connections š¤Æ) Here it is syncing 100k rows to a PGlite app:
00:00 - New docs for the @tan_stack DB live query api š We completed a major refactor/rewrite of both the query builder and query engine a couple of weeks ago, and the docs are now live - link in š§µ Live queries use IVM* to do the minimal work on each change.
- We've just shipped query driven sync for @tan_stack DB! Huge lift to get here - this is the DX I dreamed of when I started working on sync. Define where data comes from (anywhere!) as collections, and then query over those collections in your components. šŖšš TanStack DB 0.5 is here with Query-Driven Sync Your component's query IS the API call. No custom endpoints. No GraphQL resolvers. Just write your query and DB figures out exactly what to fetch. Details š§µš
- How fast are live queries in @tan_stack DB? This query with two joins and over 100k rows executes in under 1ms for the initial run. Updating a row in the source collection is reflected in the results in just *10 microseconds*! š
Live queries in @tan_stack DB are incremental - they are maintained, rather than re-run, when data changes. It does this with differential dataflow / DBSP. Usually this tech trades fast updates for a slower start. š§µ on how we made *both* initial and update runs lightning fast!
