Articles
Worth reading
A short take on pieces worth your time — architecture, AI, system design, and engineering careers — each one linked back to its original source.
4 articles
Securing the future of AI agents
Rohin Shah and Four Flynn lay out Google DeepMind's roadmap for securing their own internal systems against AI agents that are capable enough to be genuinely useful but not reliably aligned: access control, anomaly detection on agent trajectories, and human escalation for suspicious behavior, treated as security engineering problems rather than alignment philosophy. What struck me is that this isn't speculative; they describe an internal prototype already monitoring coding-agent trajectories in production. A useful, concrete counterpoint to AI safety discussions that stay abstract, this is what "securing an agent deployment" looks like as actual infrastructure.
Practices for Governing Agentic AI Systems
Yonadav Shavit, Sandhini Agarwal, and their OpenAI co-authors were writing about agentic AI governance back in December 2023, before most of the industry had settled on what an "agent" even meant, and the practices they propose, clear task specification, human-in-the-loop checkpoints for consequential actions, and traceability for what an agent actually did, read as more prescient than dated two years on. It's a policy paper rather than an engineering blog post, but the practices map directly onto real design decisions if you're building anything that lets a model take actions with side effects. Good context for why "just let the agent decide" is a bigger decision than it sounds.
OWASP Top 10 for LLM Applications
The LLM-specific sibling to the general OWASP Top 10, and the risks it names, prompt injection sitting at number one, plus excessive agency, insecure output handling, and supply chain risk for models and training data, are the ones I now walk through on every project that wires an LLM into something with real permissions. It's a useful corrective to how much AI security discussion still focuses on model behavior in the abstract rather than the concrete ways an LLM-powered feature gets exploited in production. Pair this with the standard OWASP Top 10 rather than treating it as a replacement; most LLM apps are still full of ordinary web vulnerabilities too.
Introduction - OWASP Top 10:2025
The 2025 refresh of the industry's baseline security checklist, and the shift in framing is the interesting part: less a list of individual bugs to patch, more a set of root causes — insecure design, supply chain failures, mishandling of exceptional conditions — that produce those bugs in the first place. Worth reading as a gut-check on process, not just as a scanner ruleset, especially the expanded supply-chain category given how much of a modern stack you don't actually write yourself. The baseline every engineer building anything internet-facing should know cold.