Some notes on what's new

Hey guys, you've probably noticed we've changed a thing or two about this website. If you don't like it yet, give it a chance, it might grow on you.

Two things are important: First, a stronger focus to let you just scroll through the main feed and still grasp the most relevant information. Second, with that comes a return to more blog-style publishing overall.

We've also added a system we call "Context on Demand". Right now it's a pretty standard RAG system that runs against the context of the article you're reading, all our published articles from the last five years, and also asks the model to add information if we can't serve the full answer. The model tries to make transparent when it does that.

This feature is still experimental, and as a frequent reader of this site, you know these things are prone to errors. But we've been using it ourselves for the last few years to research within our own articles and found it genuinely useful. We'll try to improve it. My hunch is that there's a market for human curation combined with AI contextualization. Quite frankly, it saves us time typing while still providing a solid source of information, to keep up with the speed of this market, since we're a tiny team.

That said, the page is much more beautiful without all these ads. We'll put some of our articles behind a paywall, as well as our archive from the last few years. For that, we're offering a new subscription. Subscribers get the site ad-free, plus an exclusive newsletter every two months or so where we summarize the most frontier, high-signal intel we could gather on one specific topic. Not the big studies, the smaller hints that often get lost in the noise. The first one in 2026 will be about the possible paths forward for agentic AI.

So if you're interested in that, it would be great if you support our site. Everything helps. And if you're not yet convinced, maybe we'll win you over next year. Enjoy the holidays!

Matthias, Max & Jonathan