Weeknotes 2026.06

Posted on So 08 Februar 2026 in Blog

It has been another though week. I am still not feeling 100 percent recovered from the cold I mentioned two weeks ago.

That was the reason I took some days off and went on a day trip where I decided at the bus stop or train station where to go next. Most of the time I took the next bus or next train that entered a station and decided on the way where to get off.

It was quite fun to take trains and buses I usually do not take.

On Friday, I attended a multimedia talk by Christoph Strasser. It was very entertaining, even though I already knew some of his stories from his Sitzfleisch podcast.

Content:

Story of the week

Recall.ai has published a blog post about their scaling issues and their investigation into the reason. They figured out that the postmaster process was not able to handle their workload.

Postgres Postmaster does not scale

Postgres and the world of data

Just use Postgres

To some extent such a claim is controversial. I guess it depends on your team size and your workloads. In certain workloads it makes sense to use a different database, but every database system is another tool that needs to be learned and managed. In my opinion, it is always a tradeoff between a usable system and a system that is too complex.

My rule is: make Postgres the standard database in the datacenter and try to scale it up. And if the workload demands it, use a specialized database for specific workloads.

TigerData put together which databases could by replaced by using Postgres and its ecosystem.

It’s 2026, Just Use Postgres

Postgres contribution statistics

Phil Eaton has looked at which companies are behind the contributors to Postgres 18. Meassuring contribution is not easy, but there are some interesting numbers. The most commits are from EnterpriseDB and Microsoft. The most people commiting code are from Postgres Professional and Amazon.

Companies behind Postgres 18 development

Hackorum

Kai Wagner announced a new project: Hackorum. This is a modern interface for the pg-hackers mailinglist. I guess it is a step in the right direction to attract more people to the Postgres project.

Hackorum - A Forum-Style View of pg-hackers

FOSDEM 2026

Jimmy Angelakos has written a blog post about his presentation at FOSDEM 2026: FOSDEM 2026 — Defining "Drop-in Replacement" and Beyond.

And Kai Wagner has a writeup about his FOSDEM experience: PGDay and FOSDEM Report from Kai

Fast spatial joins

FloeDB shared some insights on how they speed up spatial joins with H3 indexes.

How we made geo joins 400× faster with H3 indexes

Postgres Release Monitor

Security and Privacy

Notepad++

Notepad++ obviously has some trouble. They got owned by a state-sponsored hacker group and redirected the updates to a malicious executable. But that is not the worst thing. They are using a shared hosting provider. As a project I wouldn't trust a shared hosting provider delivering my software updates. I wouldn't want too many people to mess with my critical server, but I guess that is a decision a project has to make.

Digital Sovereignty

It seems like there is a little movement and the European Commission is preparing an internal communication tool based on the Matrix protocol. For now it seems like this will be used as a backup to Teams.

Commission trials European open source communications software

Data breach ticker

AI

Wayme World Model

Waymo announced the Waymo World Model. This is a generative model used for autonomous driving simulation.

The Waymo World Model: A New Frontier For Autonomous Driving Simulation

OpenClaw disaster

OpenClaw had a bad week. There have been a lot of naming issues around the project. But that wasn't the bad part.

A lot of security problems popped up this week, and in our company it was banned. Okay, the last part is probably not a big deal for the project ;-)

Around the world

Is Site Reliability Engineering the future?

Swizec claims that SRE is the future of software engineering.

Anyone can build a greenfield demo, but it takes engineering to run a service.

The future of software engineering is SRE

Scaling issues

Brendan Gregg is known for his expertise in computing performance. He has written about his process in joining OpenAI. What stood out the most for me was the following quote:

The engineers I met were impressive: the AI giants have been very selective, to the point that I wasn't totally sure I'd pass the interviews myself.

It seems like there are many excellent people working at OpenAI, if someone like Brendan Gregg isn't totally sure about his interview process.

Why I joined OpenAI