F# and AI agents
3 November 2026
Installing Codex and Claude Code on my VPS took ten minutes. Deciding where they were allowed to write took me much longer. A field report on pnpm, a trusted ~/projects zone, Bubblewrap, and a sandbox I ended up leaving off.
F# and AI agents
27 October 2026
Six articles in, a local/cloud split is starting to take shape on the VPS. This isn't the series' final architecture, just the first version solid enough to be put to the test. Seventh article in the series.
F# and AI agents
23 October 2026
My local agent kept telling me “Build succeeded” and “Tests passed”, and I couldn't find a single file in the project. I was quick to blame the model: the first culprit turned out to be my Docker mount. This article opens the part of the series on the Hermes sandbox trap.
F# and AI agents
20 October 2026
After the local CPU cold shower, I needed a cloud path for interactive work. Two attempts, a clean failure and an immediate success: the question that matters isn't which cloud is best, but which one wires up simply without blowing up the bill. Sixth article in the series.
F# and AI agents
13 October 2026
Hermes' /bg mode delivers a real service: launch a task and get control back. But three separate experiences taught me that a result that comes back on its own has to be tested before it's trusted. Fifth article in the series.
F# and AI agents
6 October 2026
Numbers, not a feeling: what a local LLM actually delivers on a CPU-only VPS, compared to the cloud. Fourth article in the series.
F# and AI agents
29 September 2026
Assembling Docker, Ollama, and Hermes on the hardened VPS, and discovering that a model 'compatible with Ollama' isn't necessarily usable as-is by an agentic orchestrator. Third article in the series.
F# and AI agents
22 September 2026
Before installing the first AI agent on my VPS, I hardened it: key-only SSH, a firewall, Tailscale, and a safety net in case something goes wrong. Second article in the series on my agentic environment.
F# and AI agents
15 September 2026
I wanted to build my own agentic environment on a VPS, at a controlled cost. First article in a series documenting what actually works, what breaks, and the architecture that emerges from it.
F# and AI agents
3 September 2026
AI agents aren't just a tool to write code faster anymore. They're also changing how we need to design the systems they operate on. Two sides of the same coin, seen through the same Tech Lead lens.
F# and AI agents
2 December 2025
The project barely exists: a simple Next.js home page and a few empty F# projects. And yet, the most important part is already done. Before writing any feature, I focused on building solid foundations. Here’s how I prepare a SaaS so it stands tall from day one.
F# and AI agents
13 April 2025
A pragmatic comparison between C# and F# through the lens of Software Craftsmanship: expressiveness, testability, robustness, and domain modelling. No proselytising - just the facts.
F# and AI agents
31 January 2025
Discover why F#, the functional language of the .NET ecosystem, deserves your attention. Learn about its advantages, use cases, and how it enriches .NET development.