Darko retweeted
My latest experiment: DuckDB × TiDB File System.
A laptop can commit a write, close the database, and a worker on another machine can reopen it and continue. The native database and WAL stay on shared storage.
That is the foundation for a DuckDB service: shared database state with replaceable compute. Run a job, hand it off, or restart a worker without copying the database between machines.
DuckDB handles SQL, transactions and recovery. The filesystem handles storage and coordinated access.
Experimental boundary: one writer OR multiple read-only clients, using advisory locks.
WIP: github.com/siddontang/duckdb…
Darko retweeted
Software is dead in that same way restaurants are dead because i can grow my own vegetables and own an air fryer
You’re building a backend that handles 150K requests per second 😱
Which lang are you choosing?
A. Go
B. Node.js
C. Rust
D. Java
Unpopular opinion:
Writing code was never the hardest part of software engineering...
Understanding what happens when 10 million people use it simultaneously is!
Can AI solve that?
With Stream it's npm install @stream-io/feeds-react-sdk and a few lines of code. Your feed is live in minutes.
Then come fan-out, ranking, real-time updates, notifications, moderation, and scaling to millions of users. Months later your team is maintaining infrastructure instead of shipping product.
Hot take:
the next $10B AI company might not build a model.
it might build something else.
what do you think it’ll be?
This is why we built Stream: feeds, notifications and moderation as an API, so your team ships product instead of infrastructure.
> getstream.io/activity-feeds
Hot take: you probably don't need to build your own recommendation system.
just few lines of code...
Real-time notifications, out of the box.
Likes, comments, follows and mentions arrive instantly, get grouped automatically ("Sarah and 6 others liked your post") and track read/unread for you.
One flag. That's it.
getstream.io/activity-feeds
One feed doesn’t have to mean one experience.
With Feed Views, you can take an existing feed and create completely different timelines from it.
Viral posts - rank by popularity, comments & shares
Nearby - prioritize proximity
For You - bring in your own ranking signals
Same feed. Different ranking. Different experience.
And you can configure it all from the dashboard.
Building a great feed shouldn’t require a deploy every time you want to test an idea.
So we built the Feeds Playground.
Change the user. Apply a view. Add filters. Tune ranking. Hit Run.
See the result immediately.
Experiment first. Ship when it feels right 🚀
Darko retweeted
Replying to @atanasovsky
@atanasovsky, a Senior Software engineer at Stream, built a working demo of a personalized "For You" feed in about the time it takes to grab lunch, using Stream Activity Feeds under the hood. Full demo below
You can now build the core of a personalized social network in minutes.
- For You feed.
- User interests.
- Personalized ranking.
- Content discovery.
- Follow suggestions.
- Infrastructure that scales.
All without building the recommendation + feed infrastructure yourself.
Your users interact → Stream learns what they care about → their feed gets more relevant.
You can even seed interests during onboarding, so the first feed is already personalized.
Build the product. We'll handle the feed infrastructure.
Demo 👇
intereset-tags-demo.vercel.a…
Docs:
getstream.io/activity-feeds/…








