I built a tiny AI that reads the security news for me and texts me the 3 things that matter.

I used to start every morning drowning. Twenty tabs of security news, five newsletters, a dozen feeds, all of it screaming that everything was urgent and none of it telling me what actually was. By the time I'd triaged the noise, I'd lost the sharpest hour of my day to other people's headlines.

So one weekend I built a small thing to do it for me. It reads the firehose, decides what genuinely matters, and sends me three sentences. That's it. It has quietly become one of the most useful tools I own, it took an afternoon, and by the end of this you'll have built your own, which, incidentally, is exactly the kind of project that makes an interviewer lean forward. Let me walk you through it.

What it actually does

Strip away the glamour and it's four simple moves, chained together. It gathers the day's security stories from a handful of sources. It hands them to an AI with instructions to summarise and rank them by what actually matters to someone like me. It sends the top few somewhere I'll see them, Slack, Discord, email, whatever you live in. And it runs itself every morning so I never think about it. Four moves.

Step one: gather the noise

Most security sites and newsletters publish an RSS feed, a plain, machine-readable list of their latest posts, whether or not they advertise it. In Python, a small library called feedparser turns any of those feeds into a tidy list of titles and summaries with a couple of lines. Pick five or six sources you trust, grab their feed URLs, and loop over them. At the end of this step you have a pile of the day's headlines in a variable.

Step two: let the AI be the editor

This is the part that would have been impossible a few years ago and is now almost trivial. You take that pile of headlines and hand it to a language model with a clear instruction, something like: here are today's security stories; pick the three that most matter to a working security engineer, and for each, give me one sentence on what happened and one on why I should care.

That's the whole trick. The AI isn't doing anything mystical. It's doing the triage you used to do with your coffee going cold, the fuzzy human judgment of "what here is actually important," which is exactly the kind of ambiguous, language-shaped task these models are genuinely good at. A single API call, a few lines of code, and the firehose becomes a short, ranked brief.

Step three: send it somewhere you'll actually see it

A brief you have to go and fetch is a brief you'll ignore. So push it to where you already are. Slack and Discord both let you create an incoming webhook, a URL you can POST a message to, in about two clicks, no bot-building required. One more short function, and your morning brief lands in a channel on its own. Prefer email? A simple send works just as well.

Step four: make it run itself

The final step is what turns a script into a tool. You schedule it. On a Mac or Linux box, a cron job runs it every morning at seven; on any small cloud server, the same. Set it once and forget it. Now, every day before you're even awake, the thing has read the internet and left three sentences waiting for you. The toil is gone, permanently, and you did that.

Why this is worth more than a morning

Beyond getting your hour back, this little project quietly demonstrates something real, which is why it belongs on a CV and in an interview. It shows you can wire together APIs, that you understand where AI genuinely helps (judgment and summarisation) versus where it doesn't (anything that must be exact), and that your instinct is to automate toil instead of suffering it. That combination, practical automation plus good judgment about AI, is close to the actual job description of a modern security engineer.

Put it on GitHub. Write a short README. Mention it the next time someone asks what you've built lately. A working tool that solves a real problem beats another certificate every single time.

Build it this weekend and tell me what sources you chose, I'm always looking to steal a good feed, and I want to see what your AI decides to put at the top on day one.