I know my domain.
It’s something I say. It wasn’t anything I said before.
I understand my domain because I know it. That’s because I know the domain, what it is or why.
The spacious nexus is what defines my domain, essentially. I won’t be knowledgeable of everything, and that’s okay. There is no feeling of being inadequate because I don’t understand it, it’s just a matter of life. I believe it’s better to use my expertise in the domains I know, rather than try to go into domains I don’t or are yet to understand.
Generative AI
Generative AI or LLMs in particular are a very tempting thing to use in software development. LLMs made me realize how my own skills are still worth anything. I am not demotivated, rather, quite the opposite. It is the control entity for understanding.
- Do I not understand it?
- Then no, I haven’t understood the domain.
- Do I understand it?
- Potentially. But it seems more of a sign that I am yet to fully understanding my domain.
- Did it get something wrong?
- If I knew what was wrong, I’d know what else is and isn’t.
I think that programming is a creative work. In practical terms though, I’m more worried about security. It also gets rather expensive to extensively use coding agents for programming. so no. I don’t tend to get those new age problems by doing stuff the same way I did years prior.
I mildly like the idea of local LLMs. Though, hardware to run those is hard to come by. Regardless of the ethics and whatnot, I am likely to not use generative AI for software development, given its inaccessibility and potential of deskilling.