IURIX Blog

Fix forward doesn't undo a payment: the limit of AI speed when the bug lives in the data

The Pragmatic Engineer's winning essay shows an engineer merging 11x more PRs with agent-only review. Yet the serious incidents he reports all happened in the data. There, reversibility doesn't come from speed: it has to be designed into the pipeline.

Read More →

The Terraform trap: automating components is not a data platform

Scripts that create buckets and jobs in seconds feel like a platform. What is almost always missing is the data product abstraction, and building it is a design decision, not a tool purchase.

Read More →

When 20% becomes 12%: what ChatGPT's metrics teach about data

OpenAI says ChatGPT has 20% of Google's search volume; Ahrefs estimated 12%. The gap is in the definition, and that's where leaders need to look.

Read More →

The CEO as a single point of failure: governance lessons from the Nvidia case

The investigation into Jensen Huang is still open, but it already shows four risk lessons for any company: key people, everyday security, partners and credibility.

Read More →

Too many interests, too little progress: how Brás Cubas broke free from wanting to do everything

In the 'A Tal Liberdade Moderna' newsletter, Brás Cubas describes how the urge to follow every path kept him stuck for years. The author explains why he only got moving after choosing a single line of work, without giving up his other passions.

Read More →

From forward to publish: the pipeline that turns emails into a fact-checked newsletter

The author presents a system built with AWS, GitHub and Claude that recreates articles received by email, requires evidence for every claim and keeps human review before publication.

Read More →

I Forward One Email. An AI Pipeline Turns It Into a Newsletter.

How I built a reviewed, fact-checked publishing system on AWS, GitHub and Claude, and what broke along the way.

Read More →

Waiting for disaster: Alex Kantrowitz's critique of the brakeless AI race

Alex Kantrowitz of Big Technology argues that the AI industry sees clear signs of an approaching catastrophe, whether in hacks, military conflicts or bioweapons, yet will likely act only after it happens.

Read More →

One's career by Machiavelly and Jung's doctrines, a diary of a geek.

A career strategy for moving toward Principal Data Engineer or Director, combining calculated pragmatism with an understanding of corporate psychology.

Read More →

Most AI Landing Page Prompts Are Generic. I Spent Months Building Ones That Actually Work.

Most AI landing page prompts return the same layout with new colours and a new heading. Vijeet Deliwala on building prompts that actually work.

Read More →

From Zero to Enterprise: Building QONTEO's AI/Biometrics Platform

How the full technology stack of QONTEO, a B2B AI and biometrics platform, was built from a blank slate in 2020 for enterprise clients.

Read More →

AI & Machine Learning Integration

Learn how artificial intelligence is revolutionizing business processes and customer experiences.

Read More →

Lists vs Generators: The PySpark Memory Trap

Why the difference between list comprehensions and generator expressions decides memory use in PySpark ingestion, transformation and presentation.

Read More →

Cybersecurity Best Practices 2026

Essential strategies to protect your organization against evolving cyber threats.

Read More →

The Future of Cloud Computing

Discover how cloud technology is reshaping enterprise infrastructure and enabling digital transformation.

Read More →