Well, it’s been over a year since I updated last. That’s certainly not what I’d intended when I started this blog back in early 2024. I am resolving to get better about this.

Jason and I originally made the decision to move to the UK back in the summer of 2024, after the Labour Party won the election. He had just retired and we thought it would be great for him to move back after 35 years to be close to his family once again. We hoped (ultimately naively) that the new government would undo the Conservative party’s increases to the spouse visa financial fence, and so we started making plans. Only a few weeks later, Jason received the diagnosis that his abdominal aortic aneurysm had grown too large and would need surgery to repair. The surgery happened later that year and then half-way through 2025, when he had just recovered from that, Jason was diagnosed with prostate cancer which would need months of treatment and recovery. It’s true what they say: nothing makes the universe laugh like people making plans.

We finally moved in March of this year – or, rather, technically and legally only he did. I came over on a visitor visa to help him with the process of re-integration into the country. We moved in with his mum in Cornwall and got to work. Coming back to a country after three decades is a lot like becoming an adult for the first time, as we had to newly get Jason a bank account, get him registered to vote, get him his diving license (thankfully, this was able to be simply re-issued), and get him a credit card so he could begin the process of building his UK credit (which is separate from his US credit). After a couple months in Cornwall, Jason was finally and truly “back”, though finding his own flat in Cornwall turned out to be impossible given his lack of UK rental history and the low availability of rental properties in that area. After weeks of frustration, we decided to entertain the longshot of looking for rentals in an area that I would be able to find work in easily (given my skillset and work history) once I would have a spouse visa. Several searches on LinkedIn turned up the candidates: London, Cambridge, Oxford, Edinburgh, and Manchester. We dismissed the first three as too expensive, the fourth as too cold, and the last as our best bet. Thanks to a saturated rental market in Greater Manchester, Jason finally secured a flat just two weeks before my planned return to New Orleans. Those two weeks were a literal whirlwind of furnishing and outfitting a bare apartment, including finding a fridge/freezer and washer. We achieved our final basic comfort item, a sofa, just two days before my flight out, so I barely got to enjoy what we worked so hard to create. That didn’t matter to me, though, since my main concern was achieved: I was leaving Jason with his own flat in a nice area and I knew he would be comfortable and safe.

I spent six weeks visiting my family in New Orleans, the first of which was literally recovering from the craziness of the Greater Manchester move. After that, I helped my mom with a bunch of stuff around the house. During my downtime, it finally hit me that I was unemployed. I mean, I technically had been since the end of February, but helping Jason in the UK had been a full-time job in itself. Now, however, I was starting to think about what I might want to do when I can finally immigrate, and that’s when my thoughts turned to artificial intelligence as a skill. I had been firmly anti-AI for years, but now I realized that I should learn how to use it to make my resume more attractive to potential employers, especially if a skilled worker visa might turn out to be a more achievable route than Jason alone trying to achieve the financial requirement for sponsoring me for a spouse visa. I settled on the decision to try Claude and began to experiment with it here and there.

My return to the UK on a fresh visitor visa occurred in late September, just in time to celebrate Jason’s 66th birthday with him. I then began pondering ways to get him back to his pre-surgery and pre-cancer-treatment self in terms of fitness (not to mention that I could also certainly benefit). Fully aware that the coming UK winter would make the outdoors less and less attractive to us both, I started trying to figure out what I’d like us to go see so we could have some goals of landmarks and green spaces to visit that might help push our motivation over the hump of “it’s cold/windy/rainy outside”. In the meantime, I had gotten back into using Claude and finally decided to apply it to actual programming on real hardware. My first project was a simple one: I was trying to reduce smartphone screentime, and so I wanted an offline Pokemon type checker as a companion for retro gaming. I settled on using my Flipper Zero for this, mainly to see if it could be done on that device. After a few coding sessions, I was shocked to have the finished app, installed and working perfectly on my device. That’s when I truly realized just how powerful AI had become… and how important of a tool it was for me to learn for my future prospects.

I moved on from this by thinking about ways that I could build something that would be of use to me in my daily life, especially towards my various ambitions of desired ways to improve my habits. I had been trying to read more books, so my next project would be a book tracker. As for “rewards”, I immediately thought of my former work at UNM interning at the Bosque Ecosystem Monitoring Program and my former hobby of dabbling in entomology. Many of the insects I saw over the summer in Cornwall had been completely new to me, and so thought a fun way to learn about the insects of the UK would be to make collecting them virtually a reward for finishing books. Thus, Bookworm was born, which lived completely in HTML on a computer or smartphone browser. This project also revealed to me something important which I’d already suspected: all of the ways that AI can be wrong when it works from its “own” knowledge. The amount of fact-checking, both by Claude and by myself, that caught real errors was both significant and astonishing. Despite that, I enjoyed using Bookworm for a while and even shared it with a couple of friends who I knew were avid readers. However, my mind eventually turned to wondering what else I could do with the other hardware I already owned, such as my Miyoo Mini Plus, my Cardputer ADV, and my Raspberry Pis.

I recalled my old Pwnagotchi virtual pet, which I had purchased about a year ago from a website that sells them pre-made (I don’t know how to solder yet). It had been fun to use for a while to learn about how wi-fi networks and basic cybersecurity work (I never “cracked” any passwords nor do I fancy myself a “hacker”), but after a couple of months it was tossed into a drawer and forgotten. I brought it over with me on this UK visit, along with all of the other tech I just mentioned, and I realized the little Raspberry Pi Zero 2 W board with the e-ink screen could be the base for my own virtual pet project. I brainstormed with Claude about the board’s capabilities and settled on a Bluetooth-eating pet. After a solid week of steadily vibe-coding the project through flashes of inspiration and loads of “what-if” ideas, I had something that amazed me: Harald. The project has a general name of BT Dragon on its GitHub, but it will always be Harald to me. I was able to combine the book tracker and insect raising into Harald as well as loads of other features that used nothing more than the old Pwnagotchi hardware and my Android smartphone. Through developing Harald, I have learned so much about everything he can detect: Bluetooth, wi-fi, and navigational satellites (“stars”). Harald can use GPS to traverse and “claim” squares on a real-world map, as well as nearby historical landmarks and green spaces. Here, at last, I had a tool for motivation to get outside beyond “because I should for my physical and mental health”. My ADHD brain needs concrete goals and little boosts of dopamine all along the way, and Harald gives me that through the educational content of his program and its frequent rewards. It’s so wild to me that I used AI to create my own custom toy that suits my needs perfectly. I look forward to seeing what else I can achieve in his future development.