I was on a call with a colleague, walking through some Wardley Maps I'd drawn for various AI tools. Partway through my overview, he stopped me and asked a question I've heard a couple of times before: "How do you find out about these things?" I paused, realizing that the answer has less to do with any magic trick and more to do with habits— essentially, what I choose to read, watch, and listen to on a daily basis. In other words, it comes down to my information diet.

Or: why "where" your MCP server runs matters just as much as "what" it does.

Enterprise Search has always been a hard problem – one that lacks any "silver bullet" solution.

I vividly remember one of the first tasks my engineering manager handed me: figure out how to automate rotating the client_secret for an App Registration in Entra ID (the product formerly known as Azure AD).

I have observed several avoidable missteps when people try to build a brand-new engineering team from scratch, and I believe there are a few rules of thumb that can help.

Since I am not a scalable replacement for ChatGPT or Google, and because I felt this would be helpful beyond my own company, I felt compelled to write this post due to the sheer volume of questions I get on an almost daily basis about how to get a client_id and client_secret for an app registration from Entra ID (formerly known as Azure AD).

I was trying to run SpinRite 6.1 on some drives from an old Drobo 5N NAS that bit the dust a few years back, but I didn't have an x86 machine handy, but after a bit of research I found a very elegant way to do this with QEMU.

All has been quiet on the blogging front, since life has been busy: on top of my wife and I moving houses in August, as part of the COO Leadership Development Program (LDP) at Autodesk I had the opportunity to travel to to Tokyo in May, met with my business challenge team in London in July, and experienced the culmination of our program with a presentation in beautiful San Diego before AU, the Design and Make conference.

I would be remiss if I wrote my next blog post without reflecting on a recent amazing and profound experience I had in Tokyo, as part of Autodesk's COO Leadership Development Program.

Python has become the lingua franca for developing various AI/ML solutions (and more), but getting started with setting up a proper Python environment can be tricky and involve several hours of research or asking colleagues how they go about it, and many "getting started with Python" guides tend to gloss over these aspects, which is why I hope in this post to give a no-nonsense, guide to setting up Python environments.