I’ve spent some time exploring GPT-4.1, OpenAI’s follow-up to GPT-4 and 4o, and it feels like a more focused refinement rather than a broad shift in how the model interacts. Where GPT-4o leaned heavily into multimodality and new interaction styles, GPT-4.1 feels more about tightening the core experience, particularly around reasoning, instruction-following, and coding tasks. GPT-4.1 was initially released on 14 April 2025 through the OpenAI API and the OpenAI Developer Playground. At that stage, it felt clearly aimed at developers and technical users who wanted early access. A month later, on 14 May 2025, OpenAI made GPT-4.1 available in ChatGPT for Plus and Pro subscribers, bringing those refinements into day-to-day conversational use.

I’ve been keeping an eye on OpenAI’s reasoning models since o1, which launched on 12 September 2024 as the first reasoning model released by OpenAI. It signalled a shift towards models that can spend more time thinking through complex problems, and providing thorough analysis, rather than optimising purely for speed and conversational flow. With o3 and o4-mini, OpenAI is pushing that direction further. These are positioned less as general chat upgrades and more as models designed for deliberate, structured reasoning and tool-aware problem solving.

Managing configuration files across multiple systems can quickly become a mess. From terminal preferences to shell settings, we all have a finely-tuned setup that we rely on day-to-day. But how do you keep those settings consistent across machines? And how do you avoid the chaos of losing them after a reinstall? That’s where a tool like Chezmoi comes in. What is Chezmoi? Chezmoi is a powerful dotfile manager designed to track, version, and securely deploy your configuration files across multiple operating systems. It helps you manage configuration files such as ~/.zshrc, ~/.gitconfig, ~/.config/ and more — all with a simple Git workflow.

Tailscale has transformed the way I manage and access my homelab, providing seamless, secure connectivity between all my devices—wherever I am. In my previous post, “How Tailscale Transformed Secure Access to My Homelab”, I shared how Tailscale has become an essential part of my daily workflow, enabling remote access to servers, VMs, and services without exposing anything to the public internet. Today, I want to highlight another standout feature: Taildrop. What is Taildrop? Taildrop is a built-in feature of Tailscale that allows you to securely send files between any devices connected to your Tailnet. Taildrop transfers files over encrypted peer-to-peer connections, using the fastest available path. This makes it a great solution for sending sensitive or large files without third-party servers in the middle.

Connecting to services in a homelab can be a headache, especially when you want strong security without the hassle of port forwarding, firewall rules, or wrestling with certificates. Last year, I discovered Tailscale and introduced it to my homelab. Tailscale has become my solution for secure, reliable, and zero-fuss access to certain services in my homelab—which I do not want to expose via the public internet. In this post, I’ll walk through what makes Tailscale so effective, how to set it up, and the best practices I’ve picked up along the way.