on the use of AI in my work at techtravels

Like many others, I’ve discovered the awesome power that is Generative AI. While it’s undoubtedly helped revolutionize how I work, and what I work on in the software engineering field — it’s also had a crazy, positive, impact on the rate in which I’m completing personal projects. If you’ve been here long, you know what types of projects I’m talking about. Mostly Amiga-related reverse-engineering projects.

Claude Code, and others including ChatGPT/Codex latest Astra model, have accelerated the speed at which I’m able to research new topics, analyze datasets, like logic analyzer traces, assemble sample hypothesis about what is happening, produce technical papers, reverse engineer PAL devices, and the list goes on and on. Maybe you’ve seen my latest posts?

The technology is crazy powerful, and I’m certainly taking advantage of the backend-compute-subsidizing that is going on right now. You’d be a fool not to!

Some people have associated the use of AI with low-quality, low-effort, fake, or meaningless work. Because of the ease of use, it’s not surprising many people play around with some prompts, twist the available knobs, and then post the results with minimal input and effort on their part. AI is a tool, perhaps, similar to a chainsaw. Sure, I can squeeze the trigger and maybe(dangerously!!), fall a tree but that’s miles away from what these guys do, for example. Same tool, but used with considerably different levels of expertise, finesse, and skill. The output is different, and it shows.

I take a special pride in what I produce, whether it’s a blog post, technical paper, or code stored in my github repo. It’s probably far from perfect in all cases, but it’s not for lack of trying, or lack of effort given the limited free time I have to spend on it. I often obsess on making sure I get everything right. GenAI helps me get this part right, too. It has, for example, helped me collect a list of claims I’ve made, and validate each of them for accuracy. This doesn’t mean I don’t get things wrong, because I do, but it’s a bit different than the apathetic teenager at the drive-through forgetting the hash brown that you ordered. They perhaps simply don’t care. I do, and put great effort to help prevent mistakes.

I’m really happy to share my, recently accelerated, efforts to help us collectively understand how these awesome machines of the 1980s and 1990s work.

Humanly yours, Keith M

(really needed to resist the urge to run this through an agent for conciseness, which I sometimes struggle with)