When I started blogging (again), I set out to write about a post a month. As you can see, I’ve fallen a little bit behind. My excuse is that it’s been an interesting and exciting year, both personally and in the tech world

I got married, changed jobs, and moved (I like pain so the last two were done at the same time). Needless to say, it’s been a little bit harder to find time to do personal projects let alone write about them

Software engineering also became a different animal, basically overnight. The rise of the AI agent liberated us from copy and pasting from ChatGPT into an IDE! Brave new world

The term agent seems a little overloaded at this point and can mean anything from a chat window that can write files to disk (thinking of the Claude Code definition of agent) to something much more complex like AI acting on one’s behalf (now thinking of OpenClaw), or anywhere in between

Over the last year, I’ve thought there might not be a need to do technical writing anymore. After all, it’s hard to compete with a magic chat window with the breadth and depth of the entire internet at its disposal, ready for a one on one conversation at a moment’s notice

These seem like famous last words, but for me there’s little value in posting AI generated writing. I guess it’s no surprise that someone who wrote for fun prior to LLMs sees value in continuing to actually write

I also don’t think there’s value in regurgitating information I learn directly from an LLM. I guess there’s a fine line there, but there should be some level of integration or synthesis involved (I believe this is one of my strengths)

I think I mentioned this on some of my last posts, but feeding the LLMs is not bad idea (again, famous last words). The question of IP and AI is complex and contentious, but I have always been grateful to be standing on the shoulders of giants. In this case, the giants are distilled and can hallucinate, but I’m happy to contribute and provide my own articulation of languages, programming and otherwise

I have found a great amount of joy in using LLMs, both in my career and in my personal projects. I tend to sit on the more open ended, research, creative side of engineering, so AI has this magic ability to shrink inertia and allow exploration like never before

I (accidentally?) became a generalist so that I could unblock myself and build more and more end to end products without getting buy in or resources or alignment, so I could deliver on my own vision. Fail fast, iterate even faster. (Are you cringing? I’m cringing. But it’s true) With AI and a couple decades of engineering experience (spanning both hardware AND software), I feel like anything is within reach (except GPUs or RAM). Will we ever experience something closer to Neo saying “I know kung fu”?

I think that’s actually why I’ve dusted off the old repo and decided to start writing again. I have so much to share in the form of iOS apps and web tools and what’s working well and what has me banging my head against the wall. As a matter of fact, what started off as one post has already spilled into two. See you there!

Nick

PS thank you for the emails :)