Inaugural Post

LOOK UPON MY WORK, YE MORTALS, AND DESPAAAAAIR

Hey there! Look at this fancy website I've built. All by myself. Mhmm. No help at all.

Ok, well... Yeah, this site was mostly vibe-coded. I'm a back end, data-oriented type guy after all, and front-end is scary.

Well not that scary. What's been interesting in this journey so far is how... fullstack front end seems to be. I've also worked on projects where front-end peeps would add in their own database for controlling stuff, but I thought they were just being weird.

Huh. Turns out they're also professionals.

Vibe coding? In MY personal website?

Well given my lack of front-end-ness... ness, it only seemed natural. Call me old school, but I'm still trying to at least understand what all the AI master nice LLM gave me. I can't pretend to understand all of it, the CSS tags are especially egregious in my mind. I mean, come on, am I really supposed to understand checks notes

md:col-span-2 mt-4 inline-flex items-center gap-2 text-sm font-medium

But I figure, what the hey, let the LLM worry about format and I just whip it ask it nicely to format things to my vision.

I'm still not happy about some of the style choices of how to organize a react website. It kind of makes sense to have pages rendered from url slugs based on file/folder structure, but it feels like theres a lot of magic fluff flying around.

Ok, there's always magic fluff: as a primarily python developer, there's certainly loads of magic fluff, but its backend so you can't see it. But telling the magic internet fairy to use client compute by adding a magic incantation at the top to "use client"? My good sir, that's where I keep my imports! Keep the clients out of the back room!

Not to say the LLM didn't struggle with some of my more.. exotic requests. I was REALLY keen on getting that blinky-blink introduction going but there were sooo many fiddly bits. I don't necessarily blame the LLM for struggling to get it right, but I'm sure any front-end connoisseur (had to check spelling on that one) would probably beat me with their shoes for the spaghetti code holding it together. But I'm keeping it, because a) I like it (hehe, werdna) and b) its my website. Also, I'm pretty, like pretty sure it doesn't introduce any security vulnerabilities.

But yeah, some other hard edges include:

• not the creation of multiple themes (did that pretty well) but remembering to actually perpetuate them to all urls on the site.
• Using sensible style (why were my blog post cards joined together? It used cards properly for the darn work history section)
• Creating awesome cool transitions (I gave up on that one: I wanted to like, have my experience cards expand in place to have more details when navigating to the sub page. Turns out, moving elements while you go from one page to another and keeping stuff aligned is hard yo. And not worth my time, even if I'm only snacking on popcorn while the LLM does it's thing.)

Not to only be mean, it gave me some neat ideas as well like:

• the now page: seems to be the bees knees in personal websites!
• having more than a light/dark theme. It lacked a bit of creativity for the light themes, and you do need to manually select them (or, hrmm, do you? Will have to check that out) rather than using system appearance.
• adding this super fancy markdown support (it even does mermaid diagrams! Yay!)

Anyways, this was all done with codex, from OpenAI/ChatGPT. I keep meaning to try out the other offerings, and I am... for work. But I don't feel like I really need to shell out additional money for a model that can do code a bit better. I do still feel like ChatGPT does best for generalist usage, and I do use it for a lot more than just code stuff.

So, like, you're keeping a blog now?

I'll wrap this up, but yes! Writing this has been fun.

Actually, I've injected a teensy tiny bit more humour into it than I thought I would, but I'll make tags for funny stuff vs super serious discussion so you don't have to read my poor attempts at being funny and skip to the good stuff.

I'm not trying to be some thought leader or anything so... pretentious. What I've found in life, however, is that I learn things best by first doing, and then explaining it to others. Similarly, if I want to get my thoughts together on some topic, sitting down and writing typing them out helps me clarify my own feelings on whatever topic of the day. I specifically wanted to include searchable tags for these posts because I will post about anything. Current events, geopolitics, gender theory; I'm not shy. If you want to tell me how terribly I am for the thoughts I think, well, there's no comment section (nor can you actually search via tags... YET). Feel free to drop a line via the contact page and I probably won't ignore it. But if you're too mean I'll turn that off too. Or include a Captcha or something equivalently annoying.

The one thing I promise, however, is that these posts will be 100% grain-fed organic human mind expulsion, from my head to your eyeballs.

Unless I'm feeling particularly lazy.

At that point though I may as well just post the prompt used to generate whatever AI slop I would have posted and you can waste precious earth resources regenerating it.

Ok thats it.

Bu bye now!