People love to show that skeptics were wrong about something, especially when national pride hangs in the balance. The South China Morning Post published the following headline on November 3rd: “Scientists find traditional Chinese medicine is based on a complex network of proteins – 3,000 years before modern science.”…
Every software project needs to represent the reality of the business he is embedded in. The way we can represent reality as limited rational beings is through propositions, i.e, declarative statements that affirm or deny something about reality. When a collection of such propositions is stored in a computer system…
Delegate investigative work to AI agents, from L1 to L3 reviews, with consistency, auditability and traceability.
Just wondering how many transactions or writes per second MySQL can handle? While it depends on many factors, fundamentally, about as many transactions as MySQL can commit to disk per second. A modern disk can do ~1000 fsyncs per second , but MySQL will group multiple writes with each fsync. An okay rule-of-thumb…
Introduction Readyset is designed to serve queries from cached views with sub-millisecond latency. This post focuses on the cold path—cases where a cache miss forces execution against the base tables. In these scenari...
MongoDB to integrate Voyage AI's industry-leading embedding and reranking models, delivering highly accurate and relevant information retrieval to power sophisticated AI use cases NEW YORK , Feb. 24, 2025 /PRNewswire/ -- MongoDB, Inc. (NASDAQ: MDB), the leading database for modern applications,
ClickHouse, Inc., a leader in real-time analytics, data warehousing, observability, and AI/ML, today announced it has raised $350 million in Series C financing.
This is Jessica. In my Prediction for Decision-making class last quarter we talked about human predictions and decisions relative to algorithmic ones. This led to a discussion about why it’s often hard to demonstrate the value of human knowledge once you have a decent statistical model.
Over the last 6 months, we collected ~10k hours of data across thousands of unique individuals. As far as we know, this is the largest neuro-language dataset in the world. [1] See here , here , here , here , and here (discussion only, no data available) for some of the larger datasets. See recent papers discussing the…
Today, we’re pleased to release pg_clickhouse, an Apache 2-licensed PostgreSQL extension to transparently execute analytics queries on ClickHouse directly from PostgreSQL.
Update 2025/07/21: Microsoft heard the complaints, and fixed the reliability added a better status page. Baby steps. Last week, Microsoft Fabric went down yet again for hours on multiple continents. Oh, you didn’t hear about it? Let’s talk about why. First, Fabric’s status page is fabricated bullshit. The link…
A Modern C++ Implementation of the Apache Arrow Columnar Format
A story about a bug with logical replication slot creation on Postgres read replicas
At Terrateam, we are big fans of Fly.io. The service is hosted there and it’s served us well. Just deploy your TOML file, get your infrastructure, do something else with the rest of your day. One of the interesting sides of Fly is that they invest heavily in server-side SQLite. They’ve written a number of blog posts…
The future of AI is hybrid, utilizing the respective strengths of cloud and client while harnessing every Windows device to achieve more. At Microsoft, we are reimagining what’s possible by bringing powerful AI compute directly to Windows devices,
If you’ve been in this trade for a while, you have probably seen dozens of debates on the merits and problems of SQL as a relational database query language. As an ORM maintainer, I have a few gripes with SQL, but overall it is workable, and anyway, it has so much inertia that there’s no point fantasizing about a…
Discover why we built Vectroid, a serverless vector database that delivers exceptional accuracy and low latency without compromising on cost. Learn about our approach to solving the traditional tradeoffs between speed, accuracy, and cost in vector search.
The design of SQLite was inspired by the design of TCL, both in the way it handles datatypes and in the formatting of its source code. The index use case for SQLite was in a Tcl/Tk application for an industrial company. From its inception, SQLite has always depended heavily on TCL. These days, SQLite no longer uses…
Arroyo has been acquired by Cloudflare to bring serverless SQL stream processing to the Cloudflare Developer Platfrorm, integrated with Queues, Workers, and R2. The Arroyo Engine will remain open-source and self-hostable.
Embedding sizes and architectures have changed remarkably over the past 5 years
I was chatting with a close friend of mine and he sent me a link to his new SaaS that he's developing.
Since this article will probably have more reach than my usual rambling, It'll be written in English (but let's not make a habit of it).
Building MVT tiles from scratch
Don’t waste time trying to discover which files are still open when you want to check or repair an APFS disk and Disk Utility throws an error. Here’s the solution.
Testing code fidelity with TLA+ specs at MongoDB. Explore conformance checking experiences and agile modelling in distributed systems.
The OpenSearch Software Foundation today announced the general availability of OpenSearch 3.0. This major release delivers a 9.5x performance improvement over OpenSearch 1.3, building on benchmarking data that showed earlier iterations of OpenSearch operating 1.6x faster than its closest industry competitor.
Introduction to Kumo Relational — the relational foundation model for predictive AI on relational data
why people pretending CSV is dead are wrong
The following is an update on the security vulnerability identified in December 2025.
