Personal Research Blog | Hendrik Erz

It’s mid-August.

Recently, the European Union started requiring generative AI providers to watermark things generated by their models.

Many have been arguing that almost every year since 2016 or so might become “The year of the Linux desktop.”

The biggest threat to open source is not the availability of LLMs to quickly generate large amounts of mostly working code.

Last week, I stumbled upon a Reddit post in which a user asked the community “What are LLMs made of?

Update July 21st, 2026: Arthur Perret has responded to this article.

It’s been a week since the INAS 2026 Conference for the International Network of Analytical Sociologists concluded at Nuffield College, Oxford University.

About three and a half years ago, OpenAI released the first generally available chatbot, ChatGPT, into the wild.

Last week, I wrote about uncomfortable truths concerning the imminent crash of the AI bubble.

The cracks are starting to form, and they’re visible.

The fact that education is publicly accessible, (mostly) free, and generally available to all members of society is younger than you probably think.

Technology is amazing.

A few weeks ago, a relative called me up and asked me if I had a use-case for an old computer of hers.

Today, I’d like to talk about something that is a bit off for this place: the endless discussions and uncertainty connected to how much computer memory one needs.

A few weeks ago, I have written about some issues I have experienced with software in recent time, prompted by issues with Zotero after their 8.0 update.

Humans operate by heuristics.

A few hours before starting to write these lines, I experienced something I hadn’t seen in almost a decade.

It has been a few days since I successfully defended my dissertation “On the Record: Understanding a Century of Congressional Lawmaking Through Speech and Vote Behavior” at the Institute for Analytical Sociology in Sweden.

Welcome to the very last article of this series on WebGL.

Last week was all about preparing the code to run post-processing, but aside from some simple tone-mapping, we haven’t done anything that actually requires this elaborate setup.

This week, I’ll be walking you through post-processing.

Last week, I walked you through animating the render so that it can convey a sense of movement.

This week, I’m going to take the rays we’ve set up last week, and start to animate the entire thing.

Welcome back to the third installment of my weirdly complicated journey through WebGL.

This is part two of a series on WebGL and creating an iris indicator from scratch.

I have this tradition.

It finally happened to me.

Academic dissertations are a very peculiar genre of text.

So, hear me out.

Thirteen years.