The last post on this blog went up in April 2023. A lot has happened since then, and most of it falls into the two areas I spend my working days in: artificial intelligence and cybersecurity. So the blog is coming back with a narrower focus and a steadier rhythm.
What changes
From today there will be one post every day, alternating between two topics. Artificial Intelligence posts cover what new models and tools change in practice, how to use AI safely at work, and the bigger questions about frontier AI, AI safety and superintelligence. Cybersecurity posts cover vulnerabilities that are being exploited right now, step-by-step defensive guides, and prioritised checklists for small businesses that don’t have a security team.
Why these two topics together
I find it hard to write about one without the other. Attackers use AI to write better phishing emails and to find vulnerabilities faster, and defenders use it to triage alerts and review code. AI systems themselves (chatbots, copilots and autonomous agents) have also opened a new attack surface, from prompt injection to company data leaking through a well-meaning employee’s chat window.
How posts are written
Each post ends with what to do about the topic, since a summary of what happened is rarely enough to act on. Facts are checked against primary sources such as vendor advisories, CISA, NIST, OWASP and official announcements, and those sources are linked at the end. Rumours, unconfirmed breach claims and leaderboard speculation stay out, or are labelled as unconfirmed when they matter. Most posts should take about five minutes to read, roughly the length of a coffee.
Coming up this week
- Tuesday: the vulnerabilities CISA confirmed as actively exploited in early November 2026, and who needs to patch what.
- Wednesday: prompt injection, why AI agents can be tricked and how to limit the damage.
- Thursday: ransomware defence for small businesses.
- Friday: what “superintelligence” means, and why serious people worry about it.
- Saturday: a practical plan for moving from passwords to passkeys.
- Sunday: using AI chatbots at work without leaking company data.
The older posts stay where they are. Some have aged better than others, and I’ll correct the ones that need it, starting with my 2023 ransomware post.
If there’s a topic you want covered, leave a comment. Helping others still helps me.