An AI keeps up with the war in Ukraine, daily and in a weekly recap, and with the tech industry weekly. It drops short briefings in my inbox, so I'm up to speed in five minutes.
There are two things I want to keep a close eye on: the war in Ukraine and the AI and tech world. Both move too fast to skim, and too slow to justify spending hours on them every day.
So I built a system that does the groundwork. An AI reads the news stream every day, weighs what matters, and leaves a short briefing in my inbox. Not a wall of headlines, but a readable overview: what's going on, how the balance sits, and where it comes from.
Following a war the way it actually unfolds means wading through dozens of videos, analyses and reports. Most days I don't have that time, and then it stays stuck at scattered fragments without any coherence. I wanted the essence without losing my whole morning to it.
The same goes for the tech industry: more appears than anyone can follow, and the difference between noise and a real shift only shows once you put it side by side. I wanted a fixed, reliable summary that's simply there, every day again, without me having to go and fetch it.
I wasn't short on news. I was short on someone who lines it up for me every morning.
The system runs in three steps that follow each other on their own. It gathers the sources, has an AI read and weigh them, and delivers the result to a fixed place. I don't have to do anything: the briefing is there before I'm awake.
From a fixed set of channels that closely follow the war and the tech world, the new material is pulled in and staged automatically every day.
An AI model reads the whole stack, weighs what really counts, and writes it up as a structured briefing: summary, balance, and the key developments.
The briefing arrives as an email and is kept in a searchable archive, so I can also read back how something developed over the weeks.
Along the way I learned that the setup matters more than the model's cleverness. By giving the briefing a fixed shape - summary, balance, developments, sources - it stayed usable every single day, instead of a slab of text that looked different each time. And because the source is video, it only works well once the speech has been turned into clean text before the model gets to weigh in on it.
This is a rebuilt example of the daily Ukraine briefing: a short summary, the balance of the day, the developments that stood out, and the sources underneath. The layout is exactly what's ready every morning.
The day is marked by a shift in Ukrainian operations: from direct frontal fighting to systematically knocking out Russian supply lines and fuel infrastructure. Major territorial gains at the front remain absent, but the pressure on logistics forces Russian units into withdrawals.
Operationally the balance tips toward Ukraine: failing resupply and fuel shortages weigh heavier than the lack of breakthroughs on the map. Russia compensates with scaled-up production, but runs into structural shortages.
Rebuilt example, based on a real edition. The structure is exactly what the system delivers each day.
The briefings are now live: daily Ukraine, weekly Ukraine and weekly tech - refreshed automatically. The briefings themselves are written in Dutch.
The example above is the daily Ukraine briefing. Below are the three tracks the system delivers.
The same mechanism delivers three different briefings, each with its own rhythm and vantage point.
Every morning the fresh state of play: what happened that day, short and lined up.
A step back. This briefing pulls in sources with a longer horizon and bundles all the daily reports from that week into one overview, so the bigger picture shows.
What genuinely shifted in the AI and tech industry, apart from the daily noise. Once a week, weighed on what sticks.
The briefing is a tool, not a replacement for my own judgment. When I have the time, I watch the underlying videos back myself and go through the sources alongside them. The summary helps me choose where to dig deeper, but the final word stays with me.
And I'm open about the fact that the daily Ukraine briefing is written from a Ukrainian perspective. That's a deliberate angle, not a neutral account. Anyone reading along knows which lens it was written through.
A summary is always a choice. The AI weighs what stands out, but it can miss a detail that turns out to matter later, or carry over a claim that was only a hunch in the source too. That's why the sources aren't there for show: they're there to be checked.
And a single perspective is coloured by definition. It works for me because I know what I'm reading and top it up with my own viewing. But it's a briefing I steer and check, not an oracle that owns the truth.
You don't need the automation to start. The heart of this system is the instruction you give your AI, and you can use that today. Paste a handful of articles or transcripts about a topic you follow, and have an AI turn it into a fixed briefing with the prompt below.
You are my news analyst. I'll give you sources (articles or transcripts) about a topic I follow. Topic: [for example: the war in Ukraine] Perspective: [for example: from a Ukrainian perspective - state this explicitly in the briefing] Turn it into a short, structured daily briefing with fixed sections: 1. Summary - the core in a few sentences 2. Balance of the day - how it adds up, not just what happened 3. Key developments - at most 4 points that genuinely stand out 4. Sources - numbered, so I can check each point Be sober and concrete. If a claim in the source is uncertain, say so instead of presenting it as fact.
These briefings are one of the streams that feed my AI assistant Kai, and they land in the same knowledge base as my second brain. Where the Knowledge Router decides what I read, these briefings do the groundwork of summarizing. How these and the other flows came to be, from the first tinkering to a server that runs day and night, is in this build story. Want something like this yourself, even without the tech? In this step-by-step guide I show how to build your own AI news curator with the same approach, from pen and paper to fully automated.
The briefings are based on a fixed set of YouTube channels that follow the war and the tech industry. A public overview of those channels and the rest of my sources is on my sources page.
The example on this page is rebuilt; the daily Ukraine briefing is written from a Ukrainian perspective.