Jorge Gómez Sancha

As coding agents take over more of the development workflow, humans need better tools to understand, observe, and operate their production data. @juliavallina and @nmediavilla19 rebuilt the Tinybird UI for that

Launch Week Day #2: Python SDK tbrd.co/launch-week

𝚗𝚙𝚡 𝚜𝚔𝚒𝚕𝚕𝚜 𝚊𝚍𝚍 𝚝𝚒𝚗𝚢𝚋𝚒𝚛𝚍𝚌𝚘/𝚝𝚒𝚗𝚢𝚋𝚒𝚛𝚍-𝚊𝚐𝚎𝚗𝚝-𝚜𝚔𝚒𝚕𝚕𝚜 skills are updated for the python sdk too

𝚝𝚋 𝚞𝚙𝚍𝚊𝚝𝚎

Launch Week Day #1: CLI 4.0 tbrd.co/launch-week

Supabase - Convex drama aside, the point is this: your db (convex for transactional, tinybird for analytical) should feel like part of your app, not something you constantly wire up and hope stays in sync. That's why we built a TypeScript SDK for Tinybird. Not a wrapper of the APIs, any LLM can do that. It's about making analytics native to your codebase. Define once, use everywhere, compiler catches your mistakes so prod doesn't have to. github.com/tinybirdco/tin…

We shipped a bunch of performance optimizations in @tinybird nextjs app thanks to this 👇

We built Learn ClickHouse. 80+ lessons covering ingestion, querying, optimization, and distributed architectures. Not just tutorials. Production-ready patterns from teams running ClickHouse at scale. Free. No fluff. 👉 learnclickhouse.com

Looking for some deep technical content to read over the holidays? Interested in how Tinybird works under the hood? 🤓 Here's a thread with our 12 best engineering blog posts of 2025 🧵

Kafka continues to be the power horse of streaming data ingestion to Tinybird, and we are making it easier and easier.

Here are 9 real-world, "Flink-shaped" use cases teams are shipping with Kafka + Tinybird. All without a single stateful job. And a couple of CLI commands to set up 🧵🔥

Whether through Kafka or via our Events API, you can map JSON fields to columns at ingestion time by adding JSONPath expressions in the Schema. No need to materialize downstream if you know what you need. Easy!

@chronark @tinybird Any time!

@jorgesancha @tinybird 💚 let me take you up on that tomorrow I got to deal with our clickhouse outage first

hey @tinybird, what's the entry price for enterprise?

Huge value here

• Kafka for 250 KB/s? • Flink for a meal-planning AI Agent sending 2 msg/s? Absolutely absurd. It's high time we stop recommending these overkill big data solutions for small data problems. Thankfully, nature seems to be healing. 🌱 Two recent blog posts with similar names caught my attention - Kafka's 80% problem and Flink's 95% problem. They had the courage to call out the elephant in the room - that most big data solutions are overkill for the majority of use cases. Despite the blogs coming from vendors (likely biased), I see a lot of truth in what's said. Nevertheless, the trend they're riding is undeniable. Ever since the zero interest rate environment ended, we've only seen a decline in these types of ultra-complex distributed infra software products. When speculative money and high growth stopped being abundant, organizations seemed to realize the obvious - we need simple solutions for simple problems. Most problems are simple, but the solutions being marketed are anything but. Kafka has 300+ configs. Flink has more than 700... 💀 Let's stop pretending every dataset needs a highly available distributed cluster. It's no question we're seeing the rise of "Small Data" 💡 SQLite and especially DuckDB showed the world that simple, embedded, dev-friendly infra works great for the majority of use cases. Not only that. They even OUTPERFORM the distributed alternatives. 👉 For example, Zach Wilson at the Data Engineer blog recently posted a benchmark of DuckDB vs Spark showing DuckDB was 5x faster (!) for 500 million (!) row tables. ...how many businesses have more than 500m rows? Stop over-engineering for the sake of following what's being marketed. The best solution to a problem is the most practical one. 💡

@felixhhaas So good! Would love to see some of those projects, with no expectations or luggage of any kind My 14 year old "interned" for a bit this summar at @tinybird and her first assignment was "Build a personal website with @lovable to track and share your experience". Nailed it!

Imagine a world where imagination is enough 🚀 1500 kids from 80 countries just built their first products with Lovable. Everything from games to AI selfie booths. It’s insane what kids can do when you give them the right tools. The next generation isn’t learning how to code. They’re learning how to imagine. Thanks for sharing and making it possible @FabianHedin + @Thom_Wolf

Meta cutting legacy AI roles to go all in on superintelligence. Soon even robots will be updating their LinkedIns... 😅

@javisantana @AmpCode ampcode.com/news/amp-free

@javisantana About ads, I was pretty impressed to see @AmpCode's approach (and that they were first) and its a clear glimpse of what's to come IMO

Nexts steps for OpenAI: - email - spreadsheets and docs - ads