How should we approach the art of programming? Is it a disciplined practice of mastering a language and meticulously crafting code line by line, or is it fundamentally an iterative journey of trans...

Improving DuckDB DevX DuckDB, an in-process SQL database, has been gaining traction since its development began in 2018, with a significant surge in interest starting in 2022 according to Google T...

Fquery meets SQLModel Keep your Graphs and dataclasses and make them even more powerful with SQLModel and DuckDB! Fquery started as a graph query engine for nested JSON objects (think jq) that’s i...

React for Entities and Business Logic Introduction Object models have been a cornerstone of software development for decades, providing a foundation for writing business logic in various programm...

A Universal Knowledge Graph Schema I’ll keep this one short. There have been a bunch of controversies around wikipedia and it’s lack of viewpoint diversity and undue influence of state level actor...

Transpilers in the Emerging Agentic Era As we transition from the Large Language Model (LLM) era to the agentic era, anticipated to unfold in 2025, a new architectural framework is taking shape. I...

A Comparison of different approaches to accelerating Python ML code Performance bottlenecks in Python have led to several innovative solutions: Triton, PyTorch, Mojo, CuPy, and Taichi. Each of the...

Compressing Wikidata Using DuckDB Wikidata publishes a 40GB blob called latest-truthy.nt.bz2 every two days. This is the largest publicly available human knowledge base as a graph. For someone int...

Towards Explainable AI Large language models consume vast amounts of training data and spit out billions of parameters that have hard to pin down meanings. Embeddings in a higher dimensional space...

Mojo Lists vs Python Lists Mojo seeks to provide a familar programming model to python programmers, while allowing them the option of using more performant low level constructs. The stdlib is simi...